Search results for: Bayesian filtering

Authors: Tzu-Ting Hu, Jung-Der Wang, Ming-Yeng Lin, Jin-Luh Chen, Perng-Jy Tsai

Abstract:

The vinyl chloride monomer (VCM) has been classified as group 1 (human) carcinogen by the IARC. Workers exposed to VCM are known associated with the development of the liver cancer and hence might cause economical and health losses. Particularly, for those work for the petrochemical industry have been seriously concerned in the environmental and occupational health field. Considering assessing workers’ health risks and their resultant economical and health losses requires the establishment of long-term VCM exposure data for any similar exposure group (SEG) of interest, the development of suitable technologies has become an urgent and important issue. In the present study, VCM exposures for petrochemical industry workers were determined firstly based on the database of the 'Workplace Environmental Monitoring Information Systems (WEMIS)' provided by Taiwan OSHA. Considering the existence of miss data, the reconstruction of historical exposure techniques were then used for completing the long-term exposure data for SEGs with routine operations. For SEGs with non-routine operations, exposure modeling techniques, together with their time/activity records, were adopted for determining their long-term exposure concentrations. The Bayesian decision analysis (BDA) was adopted for conducting exposure and health risk assessments for any given SEG in the petrochemical industry. The resultant excessive cancer risk was then used to determine the corresponding loss of quality-adjusted life expectancy (QALE). Results show that low average concentrations can be found for SEGs with routine operations (e.g., VCM rectification 0.0973 ppm, polymerization 0.306 ppm, reaction tank 0.33 ppm, VCM recovery 1.4 ppm, control room 0.14 ppm, VCM storage tanks 0.095 ppm and wastewater treatment 0.390 ppm), and the above values were much lower than that of the permissible exposure limit (PEL; 3 ppm) of VCM promulgated in Taiwan. For non-routine workers, though their high exposure concentrations, their low exposure time and frequencies result in low corresponding health risks. Through the consideration of exposure assessment results, health risk assessment results, and QALE results simultaneously, it is concluded that the proposed method was useful for prioritizing SEGs for conducting exposure abatement measurements. Particularly, the obtained QALE results further indicate the importance of reducing workers’ VCM exposures, though their exposures were low as in comparison with the PEL and the acceptable health risk.

Keywords: exposure assessment, health risk assessment, petrochemical industry, quality-adjusted life years, vinyl chloride monomer

Procedia PDF Downloads 166

Authors: Yelin Zhao, Xinxiu Li, Martin Smelik, Oleg Sysoev, Firoj Mahmud, Dina Mansour Aly, Mikael Benson

Abstract:

Background: Early diagnosis of pancreatic cancer is clinically challenging due to vague, or no symptoms, and lack of biomarkers. Polygenic risk score (PRS) scores may provide a valuable tool to assess increased or decreased risk of PC. This study aimed to develop such PRS by filtering genetic variants identified by GWAS using transcriptional programs identified by single-cell RNA sequencing (scRNA-seq). Methods: ScRNA-seq data from 24 pancreatic ductal adenocarcinoma (PDAC) tumor samples and 11 normal pancreases were analyzed to identify differentially expressed genes (DEGs) in in tumor and microenvironment cell types compared to healthy tissues. Pathway analysis showed that the DEGs were enriched for hundreds of significant pathways. These were clustered into 40 “programs” based on gene similarity, using the Jaccard index. Published genetic variants associated with PDAC were mapped to each program to generate program PRSs (pPRSs). These pPRSs, along with five previously published PRSs (PGS000083, PGS000725, PGS000663, PGS000159, and PGS002264), were evaluated in a European-origin population from the UK Biobank, consisting of 1,310 PDAC participants and 407,473 non-pancreatic cancer participants. Stepwise Cox regression analysis was performed to determine associations between pPRSs with the development of PC, with adjustments of sex and principal components of genetic ancestry. Results: The PDAC genetic variants were mapped to 23 programs and were used to generate pPRSs for these programs. Four distinct pPRSs (P1, P6, P11, and P16) and two published PRSs (PGS000663 and PGS002264) were significantly associated with an increased risk of developing PC. Among these, P6 exhibited the greatest hazard ratio (adjusted HR[95% CI] = 1.67[1.14-2.45], p = 0.008). In contrast, P10 and P4 were associated with lower risk of developing PC (adjusted HR[95% CI] = 0.58[0.42-0.81], p = 0.001, and adjusted HR[95% CI] = 0.75[0.59-0.96], p = 0.019). By comparison, two of the five published PRS exhibited an association with PDAC onset with HR (PGS000663: adjusted HR[95% CI] = 1.24[1.14-1.35], p < 0.001 and PGS002264: adjusted HR[95% CI] = 1.14[1.07-1.22], p < 0.001). Conclusion: Compared to published PRSs, scRNA-seq-based pPRSs may be used not only to assess increased but also decreased risk of PDAC.

Keywords: cox regression, pancreatic cancer, polygenic risk score, scRNA-seq, UK biobank

Procedia PDF Downloads 59

Authors: Hesam Gholami

Abstract:

Elevated suspended sediment loads in riverine systems resulting from accelerated erosion due to human activities are a serious threat to the sustainable management of watersheds and ecosystem services therein worldwide. Therefore, mitigation of deleterious sediment effects as a distributed or non-point pollution source in the catchments requires reliable provenance information. Sediment tracing or sediment fingerprinting, as a combined process consisting of sampling, laboratory measurements, different statistical tests, and the application of mixing or unmixing models, is a useful technique for discriminating the sources of sediments. From 1996 to the present, different aspects of this technique, such as grouping the sources (spatial and individual sources), discriminating the potential sources by different statistical techniques, and modification of mixing and unmixing models, have been introduced and modified by many researchers worldwide, and have been applied to identify the provenance of fine materials in agricultural, rural, mountainous, and coastal catchments, and in large catchments with numerous lakes and reservoirs. In the last two decades, efforts exploring the uncertainties associated with sediment fingerprinting results have attracted increasing attention. The frameworks used to quantify the uncertainty associated with fingerprinting estimates can be divided into three groups comprising Monte Carlo simulation, Bayesian approaches and generalized likelihood uncertainty estimation (GLUE). Given the above background, the primary goal of this study was to apply geochemical fingerprinting within the GLUE framework in the estimation of sub-basin spatial sediment source contributions in the arid Mehran River catchment in southern Iran, which drains into the Persian Gulf. The accuracy of GLUE predictions generated using four different sets of statistical tests for discriminating three sub-basin spatial sources was evaluated using 10 virtual sediments (VS) samples with known source contributions using the root mean square error (RMSE) and mean absolute error (MAE). Based on the results, the contributions modeled by GLUE for the western, central and eastern sub-basins are 1-42% (overall mean 20%), 0.5-30% (overall mean 12%) and 55-84% (overall mean 68%), respectively. According to the mean absolute fit (MAF; ≥ 95% for all target sediment samples) and goodness-of-fit (GOF; ≥ 99% for all samples), our suggested modeling approach is an accurate technique to quantify the source of sediments in the catchments. Overall, the estimated source proportions can help watershed engineers plan the targeting of conservation programs for soil and water resources.

Keywords: sediment source tracing, generalized likelihood uncertainty estimation, virtual sediment mixtures, Iran

Procedia PDF Downloads 43

Authors: Maxence Queyrel, Edi Prifti, Alexandre Templier, Jean-Daniel Zucker

Abstract:

Analysis of the human microbiome using metagenomic sequencing data has demonstrated high ability in discriminating various human diseases. Raw metagenomic sequencing data require multiple complex and computationally heavy bioinformatics steps prior to data analysis. Such data contain millions of short sequences read from the fragmented DNA sequences and stored as fastq files. Conventional processing pipelines consist in multiple steps including quality control, filtering, alignment of sequences against genomic catalogs (genes, species, taxonomic levels, functional pathways, etc.). These pipelines are complex to use, time consuming and rely on a large number of parameters that often provide variability and impact the estimation of the microbiome elements. Training Deep Neural Networks directly from raw sequencing data is a promising approach to bypass some of the challenges associated with mainstream bioinformatics pipelines. Most of these methods use the concept of word and sentence embeddings that create a meaningful and numerical representation of DNA sequences, while extracting features and reducing the dimensionality of the data. In this paper we present an end-to-end approach that classifies patients into disease groups directly from raw metagenomic reads: metagenome2vec. This approach is composed of four steps (i) generating a vocabulary of k-mers and learning their numerical embeddings; (ii) learning DNA sequence (read) embeddings; (iii) identifying the genome from which the sequence is most likely to come and (iv) training a multiple instance learning classifier which predicts the phenotype based on the vector representation of the raw data. An attention mechanism is applied in the network so that the model can be interpreted, assigning a weight to the influence of the prediction for each genome. Using two public real-life data-sets as well a simulated one, we demonstrated that this original approach reaches high performance, comparable with the state-of-the-art methods applied directly on processed data though mainstream bioinformatics workflows. These results are encouraging for this proof of concept work. We believe that with further dedication, the DNN models have the potential to surpass mainstream bioinformatics workflows in disease classification tasks.

Keywords: deep learning, disease prediction, end-to-end machine learning, metagenomics, multiple instance learning, precision medicine

Procedia PDF Downloads 97

Authors: Alicia Yanadira Nava Fuentes, Ilse Cervantes Camacho, Amadeo José Argüelles Cruz, Ana María Balboa Verduzco

Abstract:

Augmented reality promises being present in future driving, with its immersive technology let to show directions and maps to identify important places indicating with graphic elements when the car driver requires the information. On the other side, driving is considered a multitasking activity and, for some people, a complex activity where different situations commonly occur that require the immediate attention of the car driver to make decisions that contribute to avoid accidents; therefore, the main aim of the project is the instrumentation of a platform with biometric sensors that allows evaluating the performance in driving vehicles with the influence of augmented reality devices to detect the level of attention in drivers, since it is important to know the effect that it produces. In this study, the physiological sensors EPOC X (EEG), ECG06 PRO and EMG Myoware are joined in the driving test platform with a Logitech G29 steering wheel and the simulation software City Car Driving in which the level of traffic can be controlled, as well as the number of pedestrians that exist within the simulation obtaining a driver interaction in real mode and through a MSP430 microcontroller achieves the acquisition of data for storage. The sensors bring a continuous analog signal in time that needs signal conditioning, at this point, a signal amplifier is incorporated due to the acquired signals having a sensitive range of 1.25 mm/mV, also filtering that consists in eliminating the frequency bands of the signal in order to be interpretative and without noise to convert it from an analog signal into a digital signal to analyze the physiological signals of the drivers, these values are stored in a database. Based on this compilation, we work on the extraction of signal features and implement K-NN (k-nearest neighbor) classification methods and decision trees (unsupervised learning) that enable the study of data for the identification of patterns and determine by classification methods different effects of augmented reality on drivers. The expected results of this project include are a test platform instrumented with biometric sensors for data acquisition during driving and a database with the required variables to determine the effect caused by augmented reality on people in simulated driving.

Keywords: augmented reality, driving, physiological signals, test platform

Procedia PDF Downloads 106

Authors: Kevin De Castro Cogle, Mirian Kubo, Maria Anastasiadi, Fady Mohareb, Claire Rossi

Abstract:

Background: The bioaccessibility of bioactive compounds is a critical determinant of the nutritional quality of various food products. Despite its importance, there is a limited number of comprehensive studies aimed at assessing how the composition of a food matrix influences the bioaccessibility of a compound of interest. This knowledge gap has prompted a growing need to investigate the intricate relationship between food matrix formulations and the bioaccessibility of bioactive compounds. One such class of bioactive compounds that has attracted considerable attention is curcuminoids. These naturally occurring phytochemicals, extracted from the roots of Curcuma longa, have gained popularity owing to their purported health benefits and also well known for their poor bioaccessibility Project aim: The primary objective of this research project is to systematically assess the influence of matrix composition on the bioaccessibility of curcuminoids. Additionally, this study aimed to develop a series of predictive models for bioaccessibility, providing valuable insights for optimising the formula for functional foods and provide more descriptive nutritional information to potential consumers. Methods: Food formulations enriched with curcuminoids were subjected to in vitro digestion simulation, and their bioaccessibility was characterized with chromatographic and spectrophotometric techniques. The resulting data served as the foundation for the development of predictive models capable of estimating bioaccessibility based on specific physicochemical properties of the food matrices. Results: One striking finding of this study was the strong correlation observed between the concentration of macronutrients within the food formulations and the bioaccessibility of curcuminoids. In fact, macronutrient content emerged as a very informative explanatory variable of bioaccessibility and was used, alongside other variables, as predictors in a Bayesian hierarchical model that predicted curcuminoid bioaccessibility accurately (optimisation performance of 0.97 R2) for the majority of cross-validated test formulations (LOOCV of 0.92 R2). These preliminary results open the door to further exploration, enabling researchers to investigate a broader spectrum of food matrix types and additional properties that may influence bioaccessibility. Conclusions: This research sheds light on the intricate interplay between food matrix composition and the bioaccessibility of curcuminoids. This study lays a foundation for future investigations, offering a promising avenue for advancing our understanding of bioactive compound bioaccessibility and its implications for the food industry and informed consumer choices.

Keywords: bioactive bioaccessibility, food formulation, food matrix, machine learning, probabilistic modelling

Procedia PDF Downloads 38

Authors: Folorunsho J. O., Umar M. A.

Abstract:

Access to potable water supply in both the rural and urban regions of the world has been neglected, and this has severely affected man and the aesthetics of the natural environment of man. This has further worsened the issue of diseases prevalence. This study considered the effects of inadequate domestic water supply on human health in selected neighbourhoods of Lokoja. The study used descriptive statistics such as relative frequencies, percentages and inferential statistics to analyse the data obtained through the use of structured questionnaire. The results revealed that the females and male constituted 56% and 44% of the respondents respectively; 62% of the respondents married and 32% are unmarried; respondents between ages 31 and 40 years constitute majority of the study population, while respondents with tertiary education constituted 35%, and those with secondary education were 32% of the total respondents. Furthermore, civil servants constituted 40% and unemployed 16% of the total respondents. In terms of monthly income, 40% of the respondents was found to earn between ₦31,000 - 40,000 monthly. On the perception of households on the availability and adequacy of domestic water supply, the study revealed that 64.7% of the respondents have pipe-borne water as their main source of water supply, with only 28.5% out of the 64.7% have pipe-borne water supply daily. On the relationship between water supply characteristics and health status among households, the result shows that 76% of the respondents perceived a strong relationship between water supply and health status. Cumulatively, 67% of the respondents confirm that both the quality and quantity of water supplied play a critical role in determining health status of residents of the study area. The respondents also reported skin diseases (96%), diarrhoea (96%), malaria (91%), cholera (67%), dysentery (67%), and respiratory diseases (67%) as the most perceived and experienced in the area, the disease rate in the prevalence order of malaria (81%), diarrhoea (61%), skin diseases (58%), cholera (34%), dysentery (31%) and respiratory disease (14%) respectively. Finally, the results further showed how households cope with inadequate water supply with 52% of the respondents confirm that they regularly treat their water before it was deployed for domestic uses, while 35%, 26%, 25%, 10% and 4% of the 52% respectively, adopted boiling, addition of alums, filtering with fabrics, chlorination and bleaching as the preferred treatment methods. The study thus recommended policy options that will aggressively launch adequate potable water supply infrastructure in the study area.Keywords: Potable Water, Supply, Human Health, Perception, Chlorination

Keywords: potable water, human health, perception, chlorination

Procedia PDF Downloads 32

Authors: Masoud Sheidai, Fahimeh Koohdar, Hashem Sharifi

Abstract:

Juglans regia (L.) commonly known as Persian walnut of the genus Juglans L. (Juglandaceae) is one of the most important cultivated plant species due to its high-quality wood and edible nuts. The genetic diversity analysis is essential for conservation and management of tree species. Persian walnut is native from South-Eastern Europe to North-Western China through Tibet, Nepal, Northern India, Pakistan, and Iran. The species like Persian walnut, which has a wide range of geographical distribution, should harbor extensive genetic variability to adapt to environmental fluctuations they face. We aimed to study the population genetic structure of seven Persian walnut populations including three wild and four cultivated populations by using ISSR (Inter simple sequence repeats) and SRAP (Sequence related amplified polymorphism) molecular markers. We also aimed to compare the genetic variability revealed by ISSR neutral multilocus marker and rDNA ITS sequences. The studied populations differed in morphological features as the samples in each population were clustered together and were separate from the other populations. Three wild populations studied were placed close to each other. The mantel test after 5000 times permutation performed between geographical distance and morphological distance in Persian walnut populations produced significant correlation (r = 0.48, P = 0.002). Therefore, as the populations become farther apart, they become more divergent in morphological features. ISSR analysis produced 47 bands/ loci, while we obtained 15 SRAP bands. Gst and other differentiation statistics determined for these loci revealed that most of the ISSR and SRAP loci have very good discrimination power and can differentiate the studied populations. AMOVA performed for these loci produced a significant difference (< 0.05) supporting the above-said result. AMOVA produced significant genetic difference based on ISSR data among the studied populations (PhiPT = 0.52, P = 0.001). AMOVA revealed that 53% of the total variability is due to among population genetic difference, while 47% is due to within population genetic variability. The results showed that both multilocus molecular markers and ITS sequences can differentiate Persian walnut populations. The studied populations differed genetically and showed isolation by distance (IBD). ITS sequence based MP and Bayesian phylogenetic trees revealed that Iranian walnut cultivars form a distinct clade separated from the cultivars studied from elsewhere. Almost all clades obtained have high bootstrap value. The results indicated that a combination of multilpcus and sequencing molecular markers can be used in genetic differentiation of Persian walnut.

Keywords: genetic diversity, population, molecular markers, genetic difference

Procedia PDF Downloads 132

Authors: Huseyin Guler, Cigdem Kosar

Abstract:

The aim of this study is to propose Bridge estimator for seasonal unit root tests. Seasonality is an important factor for many economic time series. Some variables may contain seasonal patterns and forecasts that ignore important seasonal patterns have a high variance. Therefore, it is very important to eliminate seasonality for seasonal macroeconomic data. There are some methods to eliminate the impacts of seasonality in time series. One of them is filtering the data. However, this method leads to undesired consequences in unit root tests, especially if the data is generated by a stochastic seasonal process. Another method to eliminate seasonality is using seasonal dummy variables. Some seasonal patterns may result from stationary seasonal processes, which are modelled using seasonal dummies but if there is a varying and changing seasonal pattern over time, so the seasonal process is non-stationary, deterministic seasonal dummies are inadequate to capture the seasonal process. It is not suitable to use seasonal dummies for modeling such seasonally nonstationary series. Instead of that, it is necessary to take seasonal difference if there are seasonal unit roots in the series. Different alternative methods are proposed in the literature to test seasonal unit roots, such as Dickey, Hazsa, Fuller (DHF) and Hylleberg, Engle, Granger, Yoo (HEGY) tests. HEGY test can be also used to test the seasonal unit root in different frequencies (monthly, quarterly, and semiannual). Another issue in unit root tests is the lag selection. Lagged dependent variables are added to the model in seasonal unit root tests as in the unit root tests to overcome the autocorrelation problem. In this case, it is necessary to choose the lag length and determine any deterministic components (i.e., a constant and trend) first, and then use the proper model to test for seasonal unit roots. However, this two-step procedure might lead size distortions and lack of power in seasonal unit root tests. Recent studies show that Bridge estimators are good in selecting optimal lag length while differentiating nonstationary versus stationary models for nonseasonal data. The advantage of this estimator is the elimination of the two-step nature of conventional unit root tests and this leads a gain in size and power. In this paper, the Bridge estimator is proposed to test seasonal unit roots in a HEGY model. A Monte-Carlo experiment is done to determine the efficiency of this approach and compare the size and power of this method with HEGY test. Since Bridge estimator performs well in model selection, our approach may lead to some gain in terms of size and power over HEGY test.

Keywords: bridge estimators, HEGY test, model selection, seasonal unit root

Procedia PDF Downloads 299

Authors: Yonas Shuke Kitawa

Abstract:

Background: Although malaria incidence has substantially fallen sharply over the past few years, the rate of decline varies by district, time, and malaria type. Despite this turn-down, malaria remains a major public health threat in various districts of Ethiopia. Consequently, the present study is aimed at developing a predictive model that helps to identify the spatio-temporal variation in malaria risk by multiple plasmodium species. Methods: We propose a multivariate spatio-temporal Bayesian model to obtain a more coherent picture of the temporally varying spatial variation in disease risk. The spatial autocorrelation in such a data set is typically modeled by a set of random effects that assign a conditional autoregressive prior distribution. However, the autocorrelation considered in such cases depends on a binary neighborhood matrix specified through the border-sharing rule. Over here, we propose a graph-based optimization algorithm for estimating the neighborhood matrix that merely represents the spatial correlation by exploring the areal units as the vertices of a graph and the neighbor relations as the series of edges. Furthermore, we used aggregated malaria count in southern Ethiopia from August 2013 to May 2019. Results: We recognized that precipitation, temperature, and humidity are positively associated with the malaria threat in the area. On the other hand, enhanced vegetation index, nighttime light (NTL), and distance from coastal areas are negatively associated. Moreover, nonlinear relationships were observed between malaria incidence and precipitation, temperature, and NTL. Additionally, lagged effects of temperature and humidity have a significant effect on malaria risk by either species. More elevated risk of P. falciparum was observed following the rainy season, and unstable transmission of P. vivax was observed in the area. Finally, P. vivax risks are less sensitive to environmental factors than those of P. falciparum. Conclusion: The improved inference was gained by employing the proposed approach in comparison to the commonly used border-sharing rule. Additionally, different covariates are identified, including delayed effects, and elevated risks of either of the cases were observed in districts found in the central and western regions. As malaria transmission operates in a spatially continuous manner, a spatially continuous model should be employed when it is computationally feasible.

Keywords: disease mapping, MSTCAR, graph-based optimization algorithm, P. falciparum, P. vivax, waiting matrix

Procedia PDF Downloads 42

Authors: Aime P. Lampad-Dela Pena, Rhodora V. Azanza, Cesar L. Villanoy, Ephrime B. Metillo, Aletta T. Yniguez

Abstract:

Phytoplankton community changes in relation to environmental parameters were compared between and within, the three interconnected basins. Phytoplankton samples were collected from thirteen stations of Baler Bay and Casiguran Sound, Aurora last May 2013 by filtering 10 L buckets of surface water and 5 L Niskin samples at 20 meters and at 30 to 40 meters depths through a 20um sieve. Duplicate samples per station were preserved, counted, and identified up to genus level, in order to determine the horizontal and vertical spatial variation of different phytoplankton functional groups during the summer ebb and flood flow. Baler Bay, Outer and Inner Casiguran Sound had a total of 89 genera from four phytoplankton groups: Diatom (62), Dinoflagellate (25), Silicoflagellate (1) and Cyanobacteria (1). Non-toxic diatom Chaetoceros spp. bloom (averaged 2.0 x 105 to 2.73 x 106 cells L⁻¹) co-existed with Bacteriastrum spp. at surface waters in Inner and Outer Casiguran. Pseudonitzschia spp. (1.73 x 106 cells L⁻¹) bloomed at bottom waters of the innermost embayment near Casiguran mangrove estuary. Cyanobacteria Trichodesmium spp. significantly increased during ebb tide at the mid-water layers (20 meters depth) in the three basins (ranged from 6, 900 to 15, 125 filaments L⁻¹), forming another bloom. Gonyaulax spp. - dominated dinoflagellate did not significantly change with depth across the three basins. Overall, diatoms and dinoflagellates community assemblages significantly changed between sites (p < 0.001) while diatoms and cyanobacteria varied within Casiguran outer and inner sites (p < 0.001) only. Tidal fluctuations significantly affected dinoflagellates and diatom groups (p < 0.001) in inner and baler sites. Chlorophyll significantly varied between (KW, p < 0.001) and within each basins (KW, p < 0.05), no tidal influence, with the highest value at inner Casiguran and at deeper waters indicating deep chlorophyll maxima. Aurora’s distinct shelf morphology favoring counterclockwise circulation pattern, advective transport, and continuous stratification of the water column could basically affect the phytoplankton assemblages and water quality of Baler Bay and Casiguran inner and outer basins. Observed spatial phytoplankton community changes with multi-species diatom and cyanobacteria bloom at different water layers of the three inter-connected embayments would be vital for any environmental management initiatives in Aurora.

Keywords: aurora fronting Philippines Rise, intermonsoon, multi-species diatom bloom, spatial variability

Procedia PDF Downloads 113

Authors: Shakila N. Senarath, Dinesh Asanka, Janaka Wijayanayake

Abstract:

The necessity of cosmetic products is arising to fulfill consumer needs of personality appearance and hygiene. A cosmetic product consists of various chemical ingredients which may help to keep the skin healthy or may lead to damages. Every chemical ingredient in a cosmetic product does not perform on every human. The most appropriate way to select a healthy cosmetic product is to identify the texture of the body first and select the most suitable product with safe ingredients. Therefore, the selection process of cosmetic products is complicated. Consumer surveys have shown most of the time, the selection process of cosmetic products is done in an improper way by consumers. From this study, a content-based system is suggested that recommends cosmetic products for the human factors. To such an extent, the skin type, gender and price range will be considered as human factors. The proposed system will be implemented by using Machine Learning. Consumer skin type, gender and price range will be taken as inputs to the system. The skin type of consumer will be derived by using the Baumann Skin Type Questionnaire, which is a value-based approach that includes several numbers of questions to derive the user’s skin type to one of the 16 skin types according to the Bauman Skin Type indicator (BSTI). Two datasets are collected for further research proceedings. The user data set was collected using a questionnaire given to the public. Those are the user dataset and the cosmetic dataset. Product details are included in the cosmetic dataset, which belongs to 5 different kinds of product categories (Moisturizer, Cleanser, Sun protector, Face Mask, Eye Cream). An alternate approach of TF-IDF (Term Frequency – Inverse Document Frequency) is applied to vectorize cosmetic ingredients in the generic cosmetic products dataset and user-preferred dataset. Using the IF-IPF vectors, each user-preferred products dataset and generic cosmetic products dataset can be represented as sparse vectors. The similarity between each user-preferred product and generic cosmetic product will be calculated using the cosine similarity method. For the recommendation process, a similarity matrix can be used. Higher the similarity, higher the match for consumer. Sorting a user column from similarity matrix in a descending order, the recommended products can be retrieved in ascending order. Even though results return a list of similar products, and since the user information has been gathered, such as gender and the price ranges for product purchasing, further optimization can be done by considering and giving weights for those parameters once after a set of recommended products for a user has been retrieved.

Keywords: content-based filtering, cosmetics, machine learning, recommendation system

Procedia PDF Downloads 108

Authors: Ole-Christian Galbo Engstrøm, Erik Schou Dreier, Birthe Møller Jespersen, Kim Steenstrup Pedersen

Abstract:

Grain quality analysis is a multi-parameterized problem that includes a variety of qualitative and quantitative parameters such as grain type classification, damage type classification, and nutrient regression. Currently, these parameters require human inspection, a multitude of instruments employing a variety of sensor technologies, and predictive model types or destructive and slow chemical analysis. This paper investigates the feasibility of applying near-infrared hyperspectral imaging (NIR-HSI) to grain quality analysis. For this study two datasets of NIR hyperspectral images in the wavelength range of 900 nm - 1700 nm have been used. Both datasets contain images of sparsely and densely packed grain kernels. The first dataset contains ~87,000 image crops of bulk wheat samples from 63 harvests where protein value has been determined by the FOSS Infratec NOVA which is the golden industry standard for protein content estimation in bulk samples of cereal grain. The second dataset consists of ~28,000 image crops of bulk grain kernels from seven different wheat varieties and a single rye variety. In the first dataset, protein regression analysis is the problem to solve while variety classification analysis is the problem to solve in the second dataset. Deep convolutional neural networks (CNNs) have the potential to utilize spatio-spectral correlations within a hyperspectral image to simultaneously estimate the qualitative and quantitative parameters. CNNs can autonomously derive meaningful representations of the input data reducing the need for advanced preprocessing techniques required for classical chemometric model types such as artificial neural networks (ANNs) and partial least-squares regression (PLS-R). A comparison between different CNN architectures utilizing 2D and 3D convolution is conducted. These results are compared to the performance of ANNs and PLS-R. Additionally, a variety of preprocessing techniques from image analysis and chemometrics are tested. These include centering, scaling, standard normal variate (SNV), Savitzky-Golay (SG) filtering, and detrending. The results indicate that the combination of NIR-HSI and CNNs has the potential to be the foundation for an automatic system unifying qualitative and quantitative grain quality analysis within a single sensor technology and predictive model type.

Keywords: deep learning, grain analysis, hyperspectral imaging, preprocessing techniques

Procedia PDF Downloads 70

Authors: Alberto Mínguez-Martínez, Jesús de Vicente y Oliva

Abstract:

Modern manufacturing processes have led to the miniaturization of systems and, as a result, parts at the micro-and nanoscale are produced. This trend seems to become increasingly important in the near future. Besides, as a requirement of Industry 4.0, the digitalization of the models of production and processes makes it very important to ensure that the dimensions of newly manufactured parts meet the specifications of the models. Therefore, it is possible to reduce the scrap and the cost of non-conformities, ensuring the stability of the production at the same time. To ensure the quality of manufactured parts, it becomes necessary to carry out traceable measurements at scales lower than one millimeter. Providing adequate traceability to the SI unit of length (the meter) to 2D and 3D measurements at this scale is a problem that does not have a unique solution in industrial environments. Researchers in the field of dimensional metrology all around the world are working on this issue. A solution for industrial environments, even if it is not complete, will enable working with some traceability. At this point, we believe that the study of the surfaces could provide us with a first approximation to a solution. Among the different options proposed in the literature, the areal topography methods may be the most relevant because they could be compared to those measurements performed using Coordinate Measuring Machines (CMM’s). These measuring methods give (x, y, z) coordinates for each point, expressing it in two different ways, either expressing the z coordinate as a function of x, denoting it as z(x), for each Y-axis coordinate, or as a function of the x and y coordinates, denoting it as z (x, y). Between others, optical measuring instruments, mainly microscopes, are extensively used to carry out measurements at scales lower than one millimeter because it is a non-destructive measuring method. In this paper, the authors propose a calibration procedure for the scales of optical measuring instruments, particularizing for a confocal microscope, using material standards easy to find and calibrate in metrology and quality laboratories in industrial environments. Confocal microscopes are measuring instruments capable of filtering the out-of-focus reflected light so that when it reaches the detector, it is possible to take pictures of the part of the surface that is focused. Varying and taking pictures at different Z levels of the focus, a specialized software interpolates between the different planes, and it could reconstruct the surface geometry into a 3D model. As it is easy to deduce, it is necessary to give traceability to each axis. As a complementary result, the roughness Ra parameter will be traced to the reference. Although the solution is designed for a confocal microscope, it may be used for the calibration of other optical measuring instruments by applying minor changes.

Keywords: industrial environment, confocal microscope, optical measuring instrument, traceability

Procedia PDF Downloads 117

Authors: M. Kharrat, G. Moreau, Z. Aboura

Abstract:

The latest advances in the weaving industry, combined with increasingly sophisticated means of materials processing, have made it possible to produce complex 3D composite structures. Mainly used in aeronautics, composite materials with 3D architecture offer better mechanical properties than 2D reinforced composites. Nevertheless, these materials require a good understanding of their behavior. Because of the complexity of such materials, the damage mechanisms are multiple, and the scenario of their appearance and evolution depends on the nature of the exerted solicitations. The AE technique is a well-established tool for discriminating between the damage mechanisms. Suitable sensors are used during the mechanical test to monitor the structural health of the material. Relevant AE-features are then extracted from the recorded signals, followed by a data analysis using pattern recognition techniques. In order to better understand the damage scenarios of interlock composite materials, a multi-instrumentation was set-up in this work for tracking damage initiation and development, especially in the vicinity of the first significant damage, called macro-damage. The deployed instrumentation includes video-microscopy, Digital Image Correlation, Acoustic Emission (AE) and micro-tomography. In this study, a multi-variable AE data analysis approach was developed for the discrimination between the different signal classes representing the different emission sources during testing. An unsupervised classification technique was adopted to perform AE data clustering without a priori knowledge. The multi-instrumentation and the clustered data served to label the different signal families and to build a learning database. This latter is useful to construct a supervised classifier that can be used for automatic recognition of the AE signals. Several materials with different ingredients were tested under various solicitations in order to feed and enrich the learning database. The methodology presented in this work was useful to refine the damage threshold for the new generation materials. The damage mechanisms around this threshold were highlighted. The obtained signal classes were assigned to the different mechanisms. The isolation of a 'noise' class makes it possible to discriminate between the signals emitted by damages without resorting to spatial filtering or increasing the AE detection threshold. The approach was validated on different material configurations. For the same material and the same type of solicitation, the identified classes are reproducible and little disturbed. The supervised classifier constructed based on the learning database was able to predict the labels of the classified signals.

Keywords: acoustic emission, classifier, damage mechanisms, first damage threshold, interlock composite materials, pattern recognition

Procedia PDF Downloads 131

Authors: Gertrude I Hewapathirana, Mohamed M. Moustafa, Michel G. Zaitouni

Abstract:

The purposes of this research are: (a) to provide an epistemological and ontological understanding of the WLB theory, practice, and research to illuminate how the WLB evolved between 2000 to 2020 and (b) to analyze peer-reviewed research to identify the gaps, hotspots, underlying dynamics, theoretical and thematic trends, influential authors, research collaborations, geographic networks, and the multidisciplinary nature of the WLB theory to guide future researchers. The research used four-step bibliometric network analysis to explore five research questions. Using keywords such as WLB and associated variants, 1190 peer-reviewed articles were extracted from the Scopus database and transformed to a plain text format for filtering. The analysis was conducted using the R version 4.1 software (R Development Core Team, 2021) and several libraries such as bibliometrics, word cloud, and ggplot2. We used the VOSviewer software (van Eck & Waltman, 2019) for network visualization. The WLB theory has grown into a multifaceted, multidisciplinary field of research. There is a paucity of research between 2000 to 2005 and an exponential growth from 2006 to 2015. The rapid increase of WLB research in the USA, UK, and Australia reflects the increasing workplace stresses due to hyper competitive workplaces, inflexible work systems, and increasing diversity and the emergence of WLB support mechanisms, legal and constitutional mandates to enhance employee and family wellbeing at multilevel social systems. A severe knowledge gap exists due to inadequate publications disseminating the "core" WLB research. "Locally-centralized-globally-discrete" collaboration among researchers indicates a "North-South" divide between developed and developing nations. A shortage in WLB research in developing nations and a lack of research collaboration hinder a global understanding of the WLB as a universal phenomenon. Policymakers and practitioners can use the findings to initiate supporting policies, and innovative work systems. The boundary expansion of the WLB concepts, categories, relations, and properties would facilitate researchers/theoreticians to test a variety of new dimensions. This is the most comprehensive WLB landscape analysis that reveals emerging trends, concepts, networks, underlying dynamics, gaps, and growing theoretical and disciplinary boundaries. It portrays the WLB as a universal theory.

Keywords: work-life balance, co-citation networks; keyword co-occurrence network, bibliometric analysis

Procedia PDF Downloads 169

Authors: Gilberto Goracci, Fabio Curti

Abstract:

This work presents a robust, light, and inexpensive algorithm to perform autonomous orbit determination using onboard magnetometer data in real-time. Magnetometers are low-cost and reliable sensors typically available on a spacecraft for attitude determination purposes, thus representing an interesting choice to perform real-time orbit determination without the need to add additional sensors to the spacecraft itself. Magnetic field measurements can be exploited by Extended/Unscented Kalman Filters (EKF/UKF) for orbit determination purposes to make up for GPS outages, yielding errors of a few kilometers and tens of meters per second in the position and velocity of a spacecraft, respectively. While this level of accuracy shows that Kalman filtering represents a solid baseline for autonomous orbit determination, it is not enough to provide a reliable state estimation in the absence of GPS signals. This work combines the solidity and reliability of the EKF with the versatility of a Recurrent Neural Network (RNN) architecture to further increase the precision of the state estimation. Deep learning models, in fact, can grasp nonlinear relations between the inputs, in this case, the magnetometer data and the EKF state estimations, and the targets, namely the true position, and velocity of the spacecraft. The model has been pre-trained on Sun-Synchronous orbits (SSO) up to 2126 kilometers of altitude with different initial conditions and levels of noise to cover a wide range of possible real-case scenarios. The orbits have been propagated considering J2-level dynamics, and the geomagnetic field has been modeled using the International Geomagnetic Reference Field (IGRF) coefficients up to the 13th order. The training of the module can be completed offline using the expected orbit of the spacecraft to heavily reduce the onboard computational burden. Once the spacecraft is launched, the model can use the GPS signal, if available, to fine-tune the parameters on the actual orbit onboard in real-time and work autonomously during GPS outages. In this way, the provided module shows versatility, as it can be applied to any mission operating in SSO, but at the same time, the training is completed and eventually fine-tuned, on the specific orbit, increasing performances and reliability. The results provided by this study show an increase of one order of magnitude in the precision of state estimate with respect to the use of the EKF alone. Tests on simulated and real data will be shown.

Keywords: artificial intelligence, extended Kalman filter, orbit determination, magnetic field

Procedia PDF Downloads 67

Authors: Jianan Sun, Ziwen Ye

Abstract:

Multidimensional computerized adaptive testing (MCAT) is a popular research topic in psychometrics. It is important for practitioners to clearly know the item-trait patterns of administered items when a test like MCAT is operated. Item-trait pattern recognition refers to detecting which latent traits in a psychological test are measured by each of the specified items. If the item-trait patterns of the replenished items in MCAT item pool are well detected, the interpretability of the items can be improved, which can further promote the abilities of the examinees who attending the MCAT to be accurately estimated. This research explores to solve the item-trait pattern recognition problem of the replenished items in MCAT item pool from the perspective of statistical variable selection. The popular multidimensional item response theory model, multidimensional two-parameter logistic model, is assumed to fit the response data of MCAT. The proposed method uses the least absolute shrinkage and selection operator (LASSO) to detect item-trait patterns of replenished items based on the essential information of item responses and ability estimates of examinees collected from a designed MCAT procedure. Several advantages of the proposed method are outlined. First, the proposed method does not strictly depend on the relative order between the replenished items and the selected operational items, so it allows the replenished items to be mixed into the operational items in reasonable order such as considering content constraints or other test requirements. Second, the LASSO used in this research improves the interpretability of the multidimensional replenished items in MCAT. Third, the proposed method can exert the advantage of shrinkage method idea for variable selection, so it can help to check item quality and key dimension features of replenished items and saves more costs of time and labors in response data collection than traditional factor analysis method. Moreover, the proposed method makes sure the dimensions of replenished items are recognized to be consistent with the dimensions of operational items in MCAT item pool. Simulation studies are conducted to investigate the performance of the proposed method under different conditions for varying dimensionality of item pool, latent trait correlation, item discrimination, test lengths and item selection criteria in MCAT. Results show that the proposed method can accurately detect the item-trait patterns of the replenished items in the two-dimensional and the three-dimensional item pool. Selecting enough operational items from the item pool consisting of high discriminating items by Bayesian A-optimality in MCAT can improve the recognition accuracy of item-trait patterns of replenished items for the proposed method. The pattern recognition accuracy for the conditions with correlated traits is better than those with independent traits especially for the item pool consisting of comparatively low discriminating items. To sum up, the proposed data-driven method based on the LASSO can accurately and efficiently detect the item-trait patterns of replenished items in MCAT.

Keywords: item-trait pattern recognition, least absolute shrinkage and selection operator, multidimensional computerized adaptive testing, variable selection

Procedia PDF Downloads 96

Authors: Dahia Chibouti, Benoit Trouette, Eric Chenier

Abstract:

With the development of micro-electro-mechanical systems (MEMS), understanding fluid flow and heat transfer at the micrometer scale is crucial. In the case where the flow characteristic length scale is narrowed to around ten times the mean free path of gas molecules, the classical fluid mechanics and energy equations are still valid in the bulk flow, but particular attention must be paid to the gas/solid interface boundary conditions. Indeed, in the vicinity of the wall, on a thickness of about the mean free path of the molecules, called the Knudsen layer, the gas molecules are no longer in local thermodynamic equilibrium. Therefore, macroscopic models based on the continuity of velocity, temperature and heat flux jump conditions must be applied at the fluid/solid interface to take this non-equilibrium into account. Although these macroscopic models are widely used, the assumptions on which they depend are not necessarily verified in realistic cases. In order to get rid of these assumptions, simulations at the molecular scale are carried out to study how molecule interaction with walls can change the fluid flow and heat transfers at the vicinity of the walls. The developed approach is based on a kind of heterogeneous multi-scale method: micro-domains overlap the continuous domain, and coupling is carried out through exchanges of information between both the molecular and the continuum approaches. In practice, molecular dynamics describes the fluid flow and heat transfers in micro-domains while the Navier-Stokes and energy equations are used at larger scales. In this framework, two kinds of micro-simulation are performed: i) in bulk, to obtain the thermo-physical properties (viscosity, conductivity, ...) as well as the equation of state of the fluid, ii) close to the walls to identify the relationships between the slip velocity and the shear stress or between the temperature jump and the normal temperature gradient. The coupling strategy relies on an implicit formulation of the quantities extracted from micro-domains. Indeed, using the results of the molecular simulations, a Bayesian regression is performed in order to build continuous laws giving both the behavior of the physical properties, the equation of state and the slip relationships, as well as their uncertainties. These latter allow to set up a learning strategy to optimize the number of micro simulations. In the present contribution, the first results regarding this coupling associated with the learning strategy are illustrated through parametric studies of convergence criteria, choice of basis functions and noise of input data. Anisothermic flows of a Lennard Jones fluid in micro-channels are finally presented.

Keywords: multi-scale, microfluidics, micro-channel, hybrid approach, coupling

Procedia PDF Downloads 144

Authors: Alexandru-Ion Marinescu

Abstract:

There exist a plethora of methods in the scientific literature which tackle the well-established task of credit score evaluation. In its most abstract form, a credit scoring algorithm takes as input several credit applicant properties, such as age, marital status, employment status, loan duration, etc. and must output a binary response variable (i.e. “GOOD” or “BAD”) stating whether the client is susceptible to payment return delays. Data imbalance is a common occurrence among financial institution databases, with the majority being classified as “GOOD” clients (clients that respect the loan return calendar) alongside a small percentage of “BAD” clients. But it is the “BAD” clients we are interested in since accurately predicting their behavior is crucial in preventing unwanted loss for loan providers. We add to this whole context the constraint that the algorithm must yield an actual, tractable mathematical formula, which is friendlier towards financial analysts. To this end, we have turned to genetic algorithms and genetic programming, aiming to evolve actual mathematical expressions using specially tailored mutation and crossover operators. As far as data representation is concerned, we employ a very flexible mechanism – LINQ expression trees, readily available in the C# programming language, enabling us to construct executable pieces of code at runtime. As the title implies, they model trees, with intermediate nodes being operators (addition, subtraction, multiplication, division) or mathematical functions (sin, cos, abs, round, etc.) and leaf nodes storing either constants or variables. There is a one-to-one correspondence between the client properties and the formula variables. The mutation and crossover operators work on a flattened version of the tree, obtained via a pre-order traversal. A consequence of our chosen technique is that we can identify and discard client properties which do not take part in the final score evaluation, effectively acting as a dimensionality reduction scheme. We compare ourselves with state of the art approaches, such as support vector machines, Bayesian networks, and extreme learning machines, to name a few. The data sets we benchmark against amount to a total of 8, of which we mention the well-known Australian credit and German credit data sets, and the performance indicators are the following: percentage correctly classified, area under curve, partial Gini index, H-measure, Brier score and Kolmogorov-Smirnov statistic, respectively. Finally, we obtain encouraging results, which, although placing us in the lower half of the hierarchy, drive us to further refine the algorithm.

Keywords: expression trees, financial credit scoring, genetic algorithm, genetic programming, symbolic evolution

Procedia PDF Downloads 93

Authors: Ryan Chalk, G. M. Shafiullah

Abstract:

Climate change is a pertinent issue facing governments and societies around the world. The industrial revolution has resulted in a steady increase in the average global temperature. The mining and energy production industries have been significant contributors to this change prompting government to intervene by promoting low emission technology within these sectors. This paper initially reviews the energy problem in Australia and the mining sector with a focus on the energy requirements and production methods utilised in Western Australia (WA). Renewable energy in the form of utility-scale solar photovoltaics (PV) provides a solution to these problems by providing emission-free energy which can be used to supplement the existing natural gas turbines in operation at the proposed site. This research presents a custom renewable solution for the mining site considering the specific township network, local weather conditions, and seasonal load profiles. A summary of the required PV output is presented to supply slightly over 50% of the towns power requirements during the peak (summer) period, resulting in close to full coverage in the trench (winter) period. Dig Silent Power Factory Software has been used to simulate the characteristics of the existing infrastructure and produces results of integrating PV. Large scale PV penetration in the network introduce technical challenges, that includes; voltage deviation, increased harmonic distortion, increased available fault current and power factor. Results also show that cloud cover has a dramatic and unpredictable effect on the output of a PV system. The preliminary analyses conclude that mitigation strategies are needed to overcome voltage deviations, unacceptable levels of harmonics, excessive fault current and low power factor. Mitigation strategies are proposed to control these issues predominantly through the use of high quality, made for purpose inverters. Results show that use of inverters with harmonic filtering reduces the level of harmonic injections to an acceptable level according to Australian standards. Furthermore, the configuration of inverters to supply active and reactive power assist in mitigating low power factor problems. Use of FACTS devices; SVC and STATCOM also reduces the harmonics and improve the power factor of the network, and finally, energy storage helps to smooth the power supply.

Keywords: climate change, mitigation strategies, photovoltaic (PV), power quality

Procedia PDF Downloads 144

Authors: H. Jamili, S. Shakiba

Abstract:

Today, one of the main concerns of human beings is facing the unpleasant changes of the environment. Buildings are responsible for a significant amount of natural resources consumption and carbon emissions production. In such a situation, this thought comes to mind that changing each building into a phenomenon of benefit to the environment. A change in a way that each building functions as an element that supports the environment, and construction, in addition to answering the need of humans, is encouraged, the way planting a tree is, and it is no longer seen as a threat to alive beings and the planet. Prospect: Today, different ideas of developing materials that can smartly function are realizing. For instance, Programmable Materials, which in different conditions, can respond appropriately to the situation and have features of modification in shape, size, physical properties and restoration, and repair quality. Studies are to progress having this purpose to plan for these materials in a way that they are easily available, and to meet this aim, there is no need to use expensive materials and high technologies. In these cases, physical attributes of materials undertake the role of sensors, wires and actuators then materials will become into robots itself. In fact, we experience robotics without robots. In recent decades, AI and technology advances have dramatically improving the performance of materials. These achievements are a combination of software optimizations and physical productions such as multi-materials 3D printing. These capabilities enable us to program materials in order to change shape, appearance, and physical properties to interact with different situations. nIt is expected that further achievements like Memory Materials and Self-learning Materials are also added to the Smart Materials family, which are affordable, available, and of use for a variety of applications and industries. From the architectural standpoint, the building skin is significantly considered in this research, concerning the noticeable surface area the buildings skin have in urban space. The purpose of this research would be finding a way that the programmable materials be used in building skin with the aim of having an effective and positive interaction. A Pixel Façade would be a solution for programming a building skin. The Pixel Facadeincludes components that contain a series of attributes that help buildings for their needs upon their environmental criteria. A PIXEL contains series of smart materials and digital controllers together. It not only benefits its physical properties, such as control the amount of sunlight and heat, but it enhances building performance by providing a list of features, depending on situation criteria. The features will vary depending on locations and have a different function during the daytime and different seasons. The primary role of a PIXEL FAÇADE can be defined as filtering pollutions (for inside and outside of the buildings) and providing clean energy as well as interacting with other PIXEL FACADES to estimate better reactions.

Keywords: building skin, environmental crisis, pixel facade, programmable materials, smart materials

Procedia PDF Downloads 66

Authors: Elia Favarelli, Enrico Testi, Andrea Giorgetti

Abstract:

Autonomous structural health monitoring (SHM) of many structures and bridges became a topic of paramount importance for maintenance purposes and safety reasons. This paper proposes a set of machine learning (ML) tools to perform automatic feature selection and detection of anomalies in a bridge from vibrational data and compare different feature extraction schemes to increase the accuracy and reduce the amount of data collected. As a case study, the Z-24 bridge is considered because of the extensive database of accelerometric data in both standard and damaged conditions. The proposed framework starts from the first four fundamental frequencies extracted through operational modal analysis (OMA) and clustering, followed by density-based time-domain filtering (tracking). The fundamental frequencies extracted are then fed to a dimensionality reduction block implemented through two different approaches: feature selection (intelligent multiplexer) that tries to estimate the most reliable frequencies based on the evaluation of some statistical features (i.e., mean value, variance, kurtosis), and feature extraction (auto-associative neural network (ANN)) that combine the fundamental frequencies to extract new damage sensitive features in a low dimensional feature space. Finally, one class classifier (OCC) algorithms perform anomaly detection, trained with standard condition points, and tested with normal and anomaly ones. In particular, a new anomaly detector strategy is proposed, namely one class classifier neural network two (OCCNN2), which exploit the classification capability of standard classifiers in an anomaly detection problem, finding the standard class (the boundary of the features space in normal operating conditions) through a two-step approach: coarse and fine boundary estimation. The coarse estimation uses classics OCC techniques, while the fine estimation is performed through a feedforward neural network (NN) trained that exploits the boundaries estimated in the coarse step. The detection algorithms vare then compared with known methods based on principal component analysis (PCA), kernel principal component analysis (KPCA), and auto-associative neural network (ANN). In many cases, the proposed solution increases the performance with respect to the standard OCC algorithms in terms of F1 score and accuracy. In particular, by evaluating the correct features, the anomaly can be detected with accuracy and an F1 score greater than 96% with the proposed method.

Keywords: anomaly detection, frequencies selection, modal analysis, neural network, sensor network, structural health monitoring, vibration measurement

Procedia PDF Downloads 92

Authors: M. Qamarul Islam, Ergun Dogan, Mehmet Yazici

Abstract:

There is a growing body of evidence that non-normal data is more prevalent in nature than the normal one. Examples can be quoted from, but not restricted to, the areas of Economics, Finance and Actuarial Science. The non-normality considered here is expressed in terms of fat-tailedness and asymmetry of the relevant distribution. In this study a skew t distribution that can be used to model a data that exhibit inherent non-normal behavior is considered. This distribution has tails fatter than a normal distribution and it also exhibits skewness. Although maximum likelihood estimates can be obtained by solving iteratively the likelihood equations that are non-linear in form, this can be problematic in terms of convergence and in many other respects as well. Therefore, it is preferred to use the method of modified maximum likelihood in which the likelihood estimates are derived by expressing the intractable non-linear likelihood equations in terms of standardized ordered variates and replacing the intractable terms by their linear approximations obtained from the first two terms of a Taylor series expansion about the quantiles of the distribution. These estimates, called modified maximum likelihood estimates, are obtained in closed form. Hence, they are easy to compute and to manipulate analytically. In fact the modified maximum likelihood estimates are equivalent to maximum likelihood estimates, asymptotically. Even in small samples the modified maximum likelihood estimates are found to be approximately the same as maximum likelihood estimates that are obtained iteratively. It is shown in this study that the modified maximum likelihood estimates are not only unbiased but substantially more efficient than the commonly used moment estimates or the least square estimates that are known to be biased and inefficient in such cases. Furthermore, in conventional regression analysis, it is assumed that the error terms are distributed normally and, hence, the well-known least square method is considered to be a suitable and preferred method for making the relevant statistical inferences. However, a number of empirical researches have shown that non-normal errors are more prevalent. Even transforming and/or filtering techniques may not produce normally distributed residuals. Here, a study is done for multiple linear regression models with random error having non-normal pattern. Through an extensive simulation it is shown that the modified maximum likelihood estimates of regression parameters are plausibly robust to the distributional assumptions and to various data anomalies as compared to the widely used least square estimates. Relevant tests of hypothesis are developed and are explored for desirable properties in terms of their size and power. The tests based upon modified maximum likelihood estimates are found to be substantially more powerful than the tests based upon least square estimates. Several examples are provided from the areas of Economics and Finance where such distributions are interpretable in terms of efficient market hypothesis with respect to asset pricing, portfolio selection, risk measurement and capital allocation, etc.

Keywords: least square estimates, linear regression, maximum likelihood estimates, modified maximum likelihood method, non-normality, robustness

Procedia PDF Downloads 381

Authors: Ning Chang, Zelong Yuan, Yunpeng Wang, Jianchun Wang

Abstract:

The utilization of Large Eddy Simulation (LES) has been extensive in turbulence research. LES concentrates on resolving the significant grid-scale motions while representing smaller scales through subfilter-scale (SFS) models. The deconvolution model, among the available SFS models, has proven successful in LES of engineering and geophysical flows. Nevertheless, the thorough investigation of how sub-filter scale dynamics and filter anisotropy affect SFS modeling accuracy remains lacking. The outcomes of LES are significantly influenced by filter selection and grid anisotropy, factors that have not been adequately addressed in earlier studies. This study examines two crucial aspects of LES: Firstly, the accuracy of direct deconvolution models (DDM) is evaluated concerning sub-filter scale (SFS) dynamics across varying filter-to-grid ratios (FGR) in isotropic turbulence. Various invertible filters are employed, including Gaussian, Helmholtz I and II, Butterworth, Chebyshev I and II, Cauchy, Pao, and rapidly decaying filters. The importance of FGR becomes evident as it plays a critical role in controlling errors for precise SFS stress prediction. When FGR is set to 1, the DDM models struggle to faithfully reconstruct SFS stress due to inadequate resolution of SFS dynamics. Notably, prediction accuracy improves when FGR is set to 2, leading to accurate reconstruction of SFS stress, except for cases involving Helmholtz I and II filters. Remarkably high precision, nearly 100%, is achieved at an FGR of 4 for all DDM models. Furthermore, the study extends to filter anisotropy and its impact on SFS dynamics and LES accuracy. By utilizing the dynamic Smagorinsky model (DSM), dynamic mixed model (DMM), and direct deconvolution model (DDM) with anisotropic filters, aspect ratios (AR) ranging from 1 to 16 are examined in LES filters. The results emphasize the DDM’s proficiency in accurately predicting SFS stresses under highly anisotropic filtering conditions. Notably high correlation coefficients exceeding 90% are observed in the a priori study for the DDM’s reconstructed SFS stresses, surpassing those of the DSM and DMM models. However, these correlations tend to decrease as filter anisotropy increases. In the a posteriori analysis, the DDM model consistently outperforms the DSM and DMM models across various turbulence statistics, including velocity spectra, probability density functions related to vorticity, SFS energy flux, velocity increments, strainrate tensors, and SFS stress. It is evident that as filter anisotropy intensifies, the results of DSM and DMM deteriorate, while the DDM consistently delivers satisfactory outcomes across all filter-anisotropy scenarios. These findings underscore the potential of the DDM framework as a valuable tool for advancing the development of sophisticated SFS models for LES in turbulence research.

Keywords: deconvolution model, large eddy simulation, subfilter scale modeling, turbulence

Procedia PDF Downloads 42

Authors: Lydia Tan, Philippe Lemey, Lieselot Houspie, Marco Viveen, Darren Martin, Frank Coenjaerts

Abstract:

Introduction: Human respiratory syncytial virus (hRSV) is the leading cause of severe respiratory tract infections in infants under the age of two. Genomic substitutions and related evolutionary dynamics of hRSV are of great influence on virus transmission behavior. The evolutionary patterns formed are due to a precarious interplay between the host immune response and RSV, thereby selecting the most viable and less immunogenic strains. Studying genomic profiles can teach us which genes and consequent proteins play an important role in RSV survival and transmission dynamics. Study design: In this study, genetic diversity and evolutionary rate analysis were conducted on 36 RSV subgroup B whole genome sequences and 37 subgroup A genome sequences. Clinical RSV isolates were obtained from nasopharyngeal aspirates and swabs of children between 2 weeks and 5 years old of age. These strains, collected during epidemic seasons from 2001 to 2011 in the Netherlands and Belgium by either conventional or 454-sequencing. Sequences were analyzed for genetic diversity, recombination events, synonymous/non-synonymous substitution ratios, epistasis, and translational consequences of mutations were mapped to known 3D protein structures. We used Bayesian statistical inference to estimate the rate of RSV genome evolution and the rate of variability across the genome. Results: The A and B profiles were described in detail and compared to each other. Overall, the majority of the whole RSV genome is highly conserved among all strains. The attachment protein G was the most variable protein and its gene had, similar to the non-coding regions in RSV, more elevated (two-fold) substitution rates than other genes. In addition, the G gene has been identified as the major target for diversifying selection. Overall, less gene and protein variability was found within RSV-B compared to RSV-A and most protein variation between the subgroups was found in the F, G, SH and M2-2 proteins. For the F protein mutations and correlated amino acid changes are largely located in the F2 ligand-binding domain. The small hydrophobic phosphoprotein and nucleoprotein are the most conserved proteins. The evolutionary rates were similar in both subgroups (A: 6.47E-04, B: 7.76E-04 substitution/site/yr), but estimates of the time to the most recent common ancestor were much lower for RSV-B (B: 19, A: 46.8 yrs), indicating that there is more turnover in this subgroup. Conclusion: This study provides a detailed description of whole RSV genome mutations, the effect on translation products and the first estimate of the RSV genome evolution tempo. The immunogenic G protein seems to require high substitution rates in order to select less immunogenic strains and other conserved proteins are most likely essential to preserve RSV viability. The resulting G gene variability makes its protein a less interesting target for RSV intervention methods. The more conserved RSV F protein with less antigenic epitope shedding is, therefore, more suitable for developing therapeutic strategies or vaccines.

Keywords: drug target selection, epidemiology, respiratory syncytial virus, RSV

Procedia PDF Downloads 383

Authors: Jurriaan Gillissen

Abstract:

This work is concerned with stabilizing the numerical integration of the Navier-Stokes equation (NSE), backwards in time. Applications involve the detection of sources of, e.g., sound, heat, and pollutants. Stable reverse numerical integration of parabolic differential equations is also relevant for image de-blurring. While the literature addresses the reverse integration problem of the advection-diffusion equation, the problem of numerical reverse integration of the NSE has, to our knowledge, not yet been addressed. Owing to the presence of viscosity, the NSE is irreversible, i.e., when going backwards in time, the fluid behaves, as if it had a negative viscosity. As an effect, perturbations from the perfect solution, due to round off errors or discretization errors, grow exponentially in time, and reverse integration of the NSE is inherently unstable, regardless of using an implicit time integration scheme. Consequently, some sort of filtering is required, in order to achieve a stable, numerical, reversed integration. The challenge is to find a filter with a minimal adverse affect on the accuracy of the reversed integration. In the present work, we explore an adjoint gradient method (AGM) to achieve this goal, and we apply this technique to two-dimensional (2D), decaying turbulence. The AGM solves for the initial velocity field u0 at t = 0, that, when integrated forward in time, produces a final velocity field u1 at t = 1, that is as close as is feasibly possible to some specified target field v1. The initial field u0 defines a minimum of a cost-functional J, that measures the distance between u1 and v1. In the minimization procedure, the u0 is updated iteratively along the gradient of J w.r.t. u0, where the gradient is obtained by transporting J backwards in time from t = 1 to t = 0, using the adjoint NSE. The AGM thus effectively replaces the backward integration by multiple forward and backward adjoint integrations. Since the viscosity is negative in the adjoint NSE, each step of the AGM is numerically stable. Nevertheless, when applied to turbulence, the AGM develops instabilities, which limit the backward integration to small times. This is due to the exponential divergence of phase space trajectories in turbulent flow, which produces a multitude of local minima in J, when the integration time is large. As an effect, the AGM may select unphysical, noisy initial conditions. In order to improve this situation, we propose two remedies. First, we replace the integration by a sequence of smaller integrations, i.e., we divide the integration time into segments, where in each segment the target field v1 is taken as the initial field u0 from the previous segment. Second, we add an additional term (regularizer) to J, which is proportional to a high-order Laplacian of u0, and which dampens the gradients of u0. We show that suitable values for the segment size and for the regularizer, allow a stable reverse integration of 2D decaying turbulence, with accurate results for more then O(10) turbulent, integral time scales.

Keywords: time reversed integration, parabolic differential equations, adjoint gradient method, two dimensional turbulence

Procedia PDF Downloads 199

Authors: Bander Alzahrani, Mohammed Alreshoodi

Abstract:

Information-centric networking (ICN) using architectures such as the Publish-Subscribe Internet Technology (PURSUIT) has been proposed as a new networking model that aims at replacing the current used end-centric networking model of the Internet. This emerged model focuses on what is being exchanged rather than which network entities are exchanging information, which gives the control plane functions such as routing and host location the ability to be specified according to the content items. The forwarding plane of the PURSUIT ICN architecture uses a simple and light mechanism based on Bloom filter technologies to forward the packets. Although this forwarding scheme solve many problems of the today’s Internet such as the growth of the routing table and the scalability issues, it is vulnerable to brute force attacks which are starting point to distributed- denial-of-service (DDoS) attacks. In this work, we design and analyze a novel source-routing and information delivery technique that keeps the simplicity of using Bloom filter-based forwarding while being able to deter different attacks such as denial of service attacks at the ingress of the network. To achieve this, special forwarding nodes called Edge-FW are directly attached to end user nodes and used to perform a security test for malicious injected random packets at the ingress of the path to prevent any possible attack brute force attacks at early stage. In this technique, a core entity of the PURSUIT ICN architecture called topology manager, that is responsible for finding shortest path and creating a forwarding identifiers (FId), uses a cryptographically secure hash function to create a 64-bit hash, h, over the formed FId for authentication purpose to be included in the packet. Our proposal restricts the attacker from injecting packets carrying random FIds with a high amount of filling factor ρ, by optimizing and reducing the maximum allowed filling factor ρm in the network. We optimize the FId to the minimum possible filling factor where ρ ≤ ρm, while it supports longer delivery trees, so the network scalability is not affected by the chosen ρm. With this scheme, the filling factor of any legitimate FId never exceeds the ρm while the filling factor of illegitimate FIds cannot exceed the chosen small value of ρm. Therefore, injecting a packet containing an FId with a large value of filling factor, to achieve higher attack probability, is not possible anymore. The preliminary analysis of this proposal indicates that with the designed scheme, the forwarding function can detect and prevent malicious activities such DDoS attacks at early stage and with very high probability.

Keywords: forwarding identifier, filling factor, information centric network, topology manager

Procedia PDF Downloads 131

Authors: Jean-Philippe Amat, Timothée Vergne, Aymeric Hans, Bénédicte Ferry, Pascal Hendrikx, Jackie Tapprest, Barbara Dufour, Agnès Leblond

Abstract:

The surveillance of infectious diseases is necessary to describe their occurrence and help the planning, implementation and evaluation of risk mitigation activities. However, the exact number of detected cases may remain unknown whether surveillance is based on serological tests because identifying seroconversion may be difficult. Moreover, incomplete detection of cases or outbreaks is a recurrent issue in the field of disease surveillance. This study addresses these two issues. Using a viral animal disease as an example (equine viral arteritis), the goals were to establish suitable rules for identifying seroconversion in order to estimate the number of cases and outbreaks detected by a surveillance system in France between 2006 and 2013, and to assess the sensitivity of this system by estimating the total number of outbreaks that occurred during this period (including unreported outbreaks) using a capture-recapture model. Data from horses which exhibited at least one positive result in serology using viral neutralization test between 2006 and 2013 were used for analysis (n=1,645). Data consisted of the annual antibody titers and the location of the subjects (towns). A consensus among multidisciplinary experts (specialists in the disease and its laboratory diagnosis, epidemiologists) was reached to consider seroconversion as a change in antibody titer from negative to at least 32 or as a three-fold or greater increase. The number of seroconversions was counted for each town and modeled using a unilist zero-truncated binomial (ZTB) capture-recapture model with R software. The binomial denominator was the number of horses tested in each infected town. Using the defined rules, 239 cases located in 177 towns (outbreaks) were identified from 2006 to 2013. Subsequently, the sensitivity of the surveillance system was estimated as the ratio of the number of detected outbreaks to the total number of outbreaks that occurred (including unreported outbreaks) estimated using the ZTB model. The total number of outbreaks was estimated at 215 (95% credible interval CrI95%: 195-249) and the surveillance sensitivity at 82% (CrI95%: 71-91). The rules proposed for identifying seroconversion may serve future research. Such rules, adjusted to the local environment, could conceivably be applied in other countries with surveillance programs dedicated to this disease. More generally, defining ad hoc algorithms for interpreting the antibody titer could be useful regarding other human and animal diseases and zoonosis when there is a lack of accurate information in the literature about the serological response in naturally infected subjects. This study shows how capture-recapture methods may help to estimate the sensitivity of an imperfect surveillance system and to valorize surveillance data. The sensitivity of the surveillance system of equine viral arteritis is relatively high and supports its relevance to prevent the disease spreading.

Keywords: Bayesian inference, capture-recapture, epidemiology, equine viral arteritis, infectious disease, seroconversion, surveillance

Procedia PDF Downloads 265

Authors: Akshay B. Pawar, Rohit Y. Parasnis

Abstract:

Heart Rate Variability (HRV) is a widely used indicator of the regulation between the autonomic nervous system (ANS) and the cardiovascular system. Besides being non-invasive, it also has the potential to predict mortality in cases involving critical injuries. The gold standard method for determining HRV is based on the analysis of RR interval time series extracted from ECG signals. However, because it is much more convenient to obtain photoplethysmogramic (PPG) signals as compared to ECG signals (which require the attachment of several electrodes to the body), many researchers have used pulse cycle intervals instead of RR intervals to estimate HRV. They have also compared this method with the gold standard technique. Though most of their observations indicate a strong correlation between the two methods, recent studies show that in healthy subjects, except for a few parameters, the pulse-based method cannot be a surrogate for the standard RR interval- based method. Moreover, the former tends to overestimate short-term variability in heart rate. This calls for improvements in or alternatives to the pulse-cycle interval method. In this study, besides the systolic peak-peak interval method (PP method) that has been studied several times, four recent PPG-based techniques, namely the first derivative peak-peak interval method (P1D method), the second derivative peak-peak interval method (P2D method), the valley-valley interval method (VV method) and the tangent-intersection interval method (TI method) were compared with the gold standard technique. ECG and PPG signals were obtained from 10 young and healthy adults (consisting of both males and females) seated in the armchair position. In order to de-noise these signals and eliminate baseline drift, they were passed through certain digital filters. After filtering, the following HRV parameters were computed from PPG using each of the five methods and also from ECG using the gold standard method: time domain parameters (SDNN, pNN50 and RMSSD), frequency domain parameters (Very low-frequency power (VLF), Low-frequency power (LF), High-frequency power (HF) and Total power or “TP”). Besides, Poincaré plots were also plotted and their SD1/SD2 ratios determined. The resulting sets of parameters were compared with those yielded by the standard method using measures of statistical correlation (correlation coefficient) as well as statistical agreement (Bland-Altman plots). From the viewpoint of correlation, our results show that the best PPG-based methods for the determination of most parameters and Poincaré plots are the P2D method (shows more than 93% correlation with the standard method) and the PP method (mean correlation: 88%) whereas the TI, VV and P1D methods perform poorly (<70% correlation in most cases). However, our evaluation of statistical agreement using Bland-Altman plots shows that none of the five techniques agrees satisfactorily well with the gold standard method as far as time-domain parameters are concerned. In conclusion, excellent statistical correlation implies that certain PPG-based methods provide a good amount of information on the pattern of heart rate variation, whereas poor statistical agreement implies that PPG cannot completely replace ECG in the determination of HRV.

Keywords: photoplethysmography, heart rate variability, correlation coefficient, Bland-Altman plot

Procedia PDF Downloads 288