Search results for: kernel methods

Authors: H. Ben Salah, A. Gannoun, C. de Peretti, A. Trabelsi

Abstract:

The Downside Risk (DSR) model for portfolio optimisation allows to overcome the drawbacks of the classical mean-variance model concerning the asymetry of returns and the risk perception of investors. This model optimization deals with a positive definite matrix that is endogenous with respect to portfolio weights. This aspect makes the problem far more difficult to handle. For this purpose, Athayde (2001) developped a new recurcive minimization procedure that ensures the convergence to the solution. However, when a finite number of observations is available, the portfolio frontier presents an appearance which is not very smooth. In order to overcome that, Athayde (2003) proposed a mean kernel estimation of the returns, so as to create a smoother portfolio frontier. This technique provides an effect similar to the case in which we had continuous observations. In this paper, taking advantage on the the robustness of the median, we replace the mean estimator in Athayde's model by a nonparametric median estimator of the returns. Then, we give a new version of the former algorithm (of Athayde (2001, 2003)). We eventually analyse the properties of this improved portfolio frontier and apply this new method on real examples.

Keywords: Downside Risk, Kernel Method, Median, Nonparametric Estimation, Semivariance

Procedia PDF Downloads 455

Authors: Joseph Thomas Eghwerido, Julian I. Mbegbu

Abstract:

In modelling environment processes, we apply multidisciplinary knowledge to explain, explore and predict the Earth's response to natural human-induced environmental changes. Thus, the analysis of spatial-time ecological and environmental studies, the spatial parameters of interest are always heterogeneous. This often negates the assumption of stationarity. Hence, the dispersion of the transportation of atmospheric pollutants, landscape or topographic effect, weather patterns depends on a good estimate of spatial covariance. The generalized linear mixed model, although linear in the expected value parameters, its likelihood varies nonlinearly as a function of the covariance parameters. As a consequence, computing estimates for a linear mixed model requires the iterative solution of a system of simultaneous nonlinear equations. In other to predict the variables at unsampled locations, we need to know the estimate of the present sampled variables. The geostatistical methods for solving this spatial problem assume covariance stationarity (locally defined covariance) and uniform in space; which is not apparently valid because spatial processes often exhibit nonstationary covariance. Hence, they have globally defined covariance. We shall consider different existing methods of solutions of spatial covariance of a space-time processes at unsampled locations. This stationary covariance changes with locations for multiple time set with some asymptotic properties.

Keywords: parametric, nonstationary, Kernel, Kriging

Procedia PDF Downloads 227

Authors: Daniel James Pitchforth, Timothy James Rogers, Ulf Tyge Tygesen, Elizabeth Jane Cross

Abstract:

Wave loading is a primary cause of fatigue within offshore structures and its quantification presents a challenging and important subtask within the SHM framework. The accurate representation of physics in such environments is difficult, however, driving the development of data-driven techniques in recent years. Within many industrial applications, empirical laws remain the preferred method of wave loading prediction due to their low computational cost and ease of implementation. This paper aims to develop an approach that combines data-driven Gaussian process models with physical empirical solutions for wave loading, including Morison’s Equation. The aim here is to incorporate physics directly into the covariance function (kernel) of the Gaussian process, enforcing derived behaviors whilst still allowing enough flexibility to account for phenomena such as vortex shedding, which may not be represented within the empirical laws. The combined approach has a number of advantages, including improved performance over either component used independently and interpretable hyperparameters.

Keywords: offshore structures, Gaussian processes, Physics informed machine learning, Kernel design

Procedia PDF Downloads 152

Authors: Malahat Rezaee, Mahiran Basri, Raja Noor Zaliha Raja Abdul Rahman, Abu Bakar Salleh

Abstract:

Sodium diclofenac is one of the most commonly used drugs of nonsteroidal anti-inflammatory drugs (NSAIDs). It is especially effective in the controlling the severe conditions of inflammation and pain, musculoskeletal disorders, arthritis, and dysmenorrhea. Formulation as nanoemulsions is one of the nanoscience approaches that have been progressively considered in pharmaceutical science for transdermal delivery of drug. Nanoemulsions are a type of emulsion with particle sizes ranging from 20 nm to 200 nm. An emulsion is formed by the dispersion of one liquid, usually the oil phase in another immiscible liquid, water phase that is stabilized using surfactant. Palm kernel oil esters (PKOEs), in comparison to other oils; contain higher amounts of shorter chain esters, which suitable to be applied in micro and nanoemulsion systems as a carrier for actives, with excellent wetting behavior without the oily feeling. This research was aimed to study the effect of O/S ratio on stability and rheological behavior of sodium diclofenac loaded and unloaded palm kernel oil esters nanoemulsion systems. The effect of different O/S ratio of 0.25, 0.50, 0.75, 1.00 and 1.25 on stability of the drug-loaded and unloaded nanoemulsion formulations was evaluated by centrifugation, freeze-thaw cycle and storage stability tests. Lecithin and cremophor EL were used as surfactant. The stability of the prepared nanoemulsion formulations was assessed based on the change in zeta potential and droplet size as a function of time. Instability mechanisms including coalescence and Ostwald ripening for the nanoemulsion system were discussed. In comparison between drug-loaded and unloaded nanoemulsion formulations, drug-loaded formulations represented smaller particle size and higher stability. In addition, the O/S ratio of 0.5 was found to be the best ratio of oil and surfactant for production of a nanoemulsion with the highest stability. The effect of O/S ratio on rheological properties of drug-loaded and unloaded nanoemulsion systems was studied by plotting the flow curves of shear stress (τ) and viscosity (η) as a function of shear rate (γ). The data were fitted to the Power Law model. The results showed that all nanoemulsion formulations exhibited non-Newtonian flow behaviour by displaying shear thinning behaviour. Viscosity and yield stress were also evaluated. The nanoemulsion formulation with the O/S ratio of 0.5 represented higher viscosity and K values. In addition, the sodium diclofenac loaded formulations had more viscosity and higher yield stress than drug-unloaded formulations.

Keywords: nanoemulsions, palm kernel oil esters, sodium diclofenac, rheoligy, stability

Procedia PDF Downloads 387

Authors: P. S. Jagadeesh Kumar, Mingmin Pan, Yang Yung, Tracy Lin Huan

Abstract:

Glaucoma is a group of visual maladies represented by the scheduled optic nerve neuropathy; means to the increasing dwindling in vision ground, resulting in loss of sight. In this paper, a novel support vector machine based retinal therapeutic for glaucoma using machine learning algorithm is conservative. The algorithm has fitting pragmatism; subsequently sustained on correlation clustering mode, it visualizes perfect computations in the multi-dimensional space. Support vector clustering turns out to be comparable to the scale-space advance that investigates the cluster organization by means of a kernel density estimation of the likelihood distribution, where cluster midpoints are idiosyncratic by the neighborhood maxima of the concreteness. The predicted planning has 91% attainment rate on data set deterrent on a consolidation of 500 realistic images of resolute and glaucoma retina; therefore, the computational benefit of depending on the cluster overlapping system pedestal on machine learning algorithm has complete performance in glaucoma therapeutic.

Keywords: machine learning algorithm, correlation clustering mode, cluster overlapping system, glaucoma, kernel density estimation, retinal therapeutic

Procedia PDF Downloads 211

Authors: Asraful Islam

Abstract:

Extended dependency graph (EDG) is a state-of-the-art isomorphic graph to represent normal logic programs (NLPs) that can characterize the consistency of NLPs by graph analysis. To construct the vertices and arcs of an EDG, additional renaming atoms and rules besides those the given program provides are used, resulting in higher space complexity compared to the corresponding traditional dependency graph (TDG). In this article, we propose an extended labeled dependency graph (ELDG) to represent an NLP that shares an equal number of nodes and arcs with TDG and prove that it is isomorphic to the domain program. The number of nodes and arcs used in the underlying dependency graphs are formulated to compare the space complexity. Results show that ELDG uses less memory to store nodes, arcs, and cycles compared to EDG. To exhibit the desirability of ELDG, firstly, the stable models of the kernel form of NLP are characterized by the admissible coloring of ELDG; secondly, a relation of the stable models of a kernel program with the handles of the minimal, odd cycles appearing in the corresponding ELDG has been established; thirdly, to our best knowledge, for the first time an inverse transformation from a dependency graph to the representing NLP w.r.t. ELDG has been defined that enables transferring analytical results from the graph to the program straightforwardly.

Keywords: normal logic program, isomorphism of graph, extended labelled dependency graph, inverse graph transforma-tion, graph colouring

Procedia PDF Downloads 179

Authors: Pei-Jyun Hsieh, Cheng-Hsuan Li, Bor-Chen Kuo

Abstract:

For hyperspectral image classification, feature reduction is an important pre-processing for avoiding the Hughes phenomena due to the difficulty for collecting training samples. Hence, lots of researches developed feature selection methods such as F-score, HSIC (Hilbert-Schmidt Independence Criterion), and etc., to improve hyperspectral image classification. However, most of them only consider the class separability in the original space, i.e., a linear class separability. In this study, we proposed a nonlinear class separability measure based on kernel trick for selecting an appropriate feature subset. The proposed nonlinear class separability was formed by a generalized RBF kernel with different bandwidths with respect to different features. Moreover, it considered the within-class separability and the between-class separability. A genetic algorithm was applied to tune these bandwidths such that the smallest with-class separability and the largest between-class separability simultaneously. This indicates the corresponding feature space is more suitable for classification. In addition, the corresponding nonlinear classification boundary can separate classes very well. These optimal bandwidths also show the importance of bands for hyperspectral image classification. The reciprocals of these bandwidths can be viewed as weights of bands. The smaller bandwidth, the larger weight of the band, and the more importance for classification. Hence, the descending order of the reciprocals of the bands gives an order for selecting the appropriate feature subsets. In the experiments, three hyperspectral image data sets, the Indian Pine Site data set, the PAVIA data set, and the Salinas A data set, were used to demonstrate the selected feature subsets by the proposed nonlinear feature selection method are more appropriate for hyperspectral image classification. Only ten percent of samples were randomly selected to form the training dataset. All non-background samples were used to form the testing dataset. The support vector machine was applied to classify these testing samples based on selected feature subsets. According to the experiments on the Indian Pine Site data set with 220 bands, the highest accuracies by applying the proposed method, F-score, and HSIC are 0.8795, 0.8795, and 0.87404, respectively. However, the proposed method selects 158 features. F-score and HSIC select 168 features and 217 features, respectively. Moreover, the classification accuracies increase dramatically only using first few features. The classification accuracies with respect to feature subsets of 10 features, 20 features, 50 features, and 110 features are 0.69587, 0.7348, 0.79217, and 0.84164, respectively. Furthermore, only using half selected features (110 features) of the proposed method, the corresponding classification accuracy (0.84168) is approximate to the highest classification accuracy, 0.8795. For other two hyperspectral image data sets, the PAVIA data set and Salinas A data set, we can obtain the similar results. These results illustrate our proposed method can efficiently find feature subsets to improve hyperspectral image classification. One can apply the proposed method to determine the suitable feature subset first according to specific purposes. Then researchers can only use the corresponding sensors to obtain the hyperspectral image and classify the samples. This can not only improve the classification performance but also reduce the cost for obtaining hyperspectral images.

Keywords: hyperspectral image classification, nonlinear feature selection, kernel trick, support vector machine

Procedia PDF Downloads 239

Authors: Mohammad Izadkhah, Mojtaba Hoseini, Alireza Khalili Tehrani

Abstract:

In this paper we presented a new method for tracking flying targets in color video sequences based on contour and kernel. The aim of this work is to overcome the problem of losing target in changing light, large displacement, changing speed, and occlusion. The proposed method is made in three steps, estimate the target location by particle filter, segmentation target region using neural network and find the exact contours by greedy snake algorithm. In the proposed method we have used both region and contour information to create target candidate model and this model is dynamically updated during tracking. To avoid the accumulation of errors when updating, target region given to a perceptron neural network to separate the target from background. Then its output used for exact calculation of size and center of the target. Also it is used as the initial contour for the greedy snake algorithm to find the exact target's edge. The proposed algorithm has been tested on a database which contains a lot of challenges such as high speed and agility of aircrafts, background clutter, occlusions, camera movement, and so on. The experimental results show that the use of neural network increases the accuracy of tracking and segmentation.

Keywords: video tracking, particle filter, greedy snake, neural network

Procedia PDF Downloads 307

Authors: A. Bahrani

Abstract:

In order to study the effect of restricted irrigation systems and different potassium fertilizer on water use efficiency and yield of rapeseed (Brassica napus L.), an experiment was conducted in an arid area in Khuzestan, Iran in 2013. The main plots consisted of three irrigation methods: FI (full irrigation), alternate furrow irrigation (AFI) and fixed furrow irrigation (FFI). Each subplot received three rates of K fertiliser application: 0, 150 or 300 kg ha-1. The results showed that the plots receiving the full irrigation resulted in significantly higher grain yields, 1000-kernel weight and grain number per pod than both alternate treatments. However, the highest WUE were obtained in alternate furrow irrigation and 300 kg K ha-1 and the lowest one was found in the FI treatment and 0 kg K ha-1. Potassium application increased RWC in alternate furrow irrigation and fixed furrow irrigation than FI treatment. Maximum oil content was observed in those treatments where full irrigation was applied while minimum oil content was produced in FFI irrigated treatments. Potassium fertilizer also increased grain oil by 15 % than control. Deficit irrigation reduced oleic acid and erucic acid. However, oleic acid and linoleic acid increased with increasing of potassium.

Keywords: erucic acid, irrigation methods, linoleic acid, oil percent, oleic acid

Procedia PDF Downloads 250

Authors: Haifeng Wang, Haili Zhang

Abstract:

Most movie recommendation systems have been developed for customers to find items of interest. This work introduces a predictive model usable by small and medium-sized enterprises (SMEs) who are in need of a data-based and analytical approach to stock proper movies for local audiences and retain more customers. We used classification models to extract features from thousands of customers’ demographic, behavioral and social information to predict their movie genre preference. In the implementation, a Gaussian kernel support vector machine (SVM) classification model and a logistic regression model were established to extract features from sample data and their test error-in-sample were compared. Comparison of error-out-sample was also made under different Vapnik–Chervonenkis (VC) dimensions in the machine learning algorithm to find and prevent overfitting. Gaussian kernel SVM prediction model can correctly predict movie genre preferences in 85% of positive cases. The accuracy of the algorithm increased to 93% with a smaller VC dimension and less overfitting. These findings advance our understanding of how to use machine learning approach to predict customers’ preferences with a small data set and design prediction tools for these enterprises.

Keywords: computational social science, movie preference, machine learning, SVM

Procedia PDF Downloads 232

Authors: Mahtab Davari, Charles Edward Okon, Somayeh Aghanavesi

Abstract:

Every institute of learning is usually interested in the performance of enrolled students. The level of these performances determines the approach an institute of study may adopt in rendering academic services. The focus of this paper is to evaluate students' academic performance in given courses of study using machine learning methods. This study evaluated various supervised machine learning classification algorithms such as Logistic Regression (LR), Support Vector Machine, Random Forest, Decision Tree, K-Nearest Neighbors, Linear Discriminant Analysis, and Quadratic Discriminant Analysis, using selected features to predict study performance. The accuracy, precision, recall, and F1 score obtained from a 5-Fold Cross-Validation were used to determine the best classification algorithm to predict students’ performances. SVM (using a linear kernel), LDA, and LR were identified as the best-performing machine learning methods. Also, using the LR model, this study identified students' educational habits such as reading and paying attention in class as strong determinants for a student to have an above-average performance. Other important features include the academic history of the student and work. Demographic factors such as age, gender, high school graduation, etc., had no significant effect on a student's performance.

Keywords: student performance, supervised machine learning, classification, cross-validation, prediction

Procedia PDF Downloads 85

Authors: Parashram Jakappa Patil

Abstract:

India is the global leader in world cashew business and cashew-nut industry is one of the important food processing industries in world. However India is the largest producer, processor, exporter and importer eschew in the world. India is providing cashew to the rest of the world. India is meeting world demand of cashew. India has a tremendous potential of cashew production and export to other countries. Every year India earns more than 2000 cores rupees through cashew trade. Cashew industry is one of the important small scale industries in the country which is playing significant role in rural development. It is generating more than 400000 jobs at remote area and 95% cashew worker are women, it is giving income to poor cashew farmers, majority cashew processing units are small and cottage, it is helping to stop migration from young farmers for employment opportunities, it is motivation rural entrepreneurship development and it is also helping to environment protection etc. Hence India cashew business is very important agribusiness in India which has potential make inclusive development. World Bank and IMF recognized cashew-nut industry is one the important tool for poverty eradication at global level. It shows important of cashew business and its strong existence in India. In spite of such huge potential cashew processing industry is facing different problems such as lack of infrastructure ability, lack of supply of raw cashew, lack of availability of finance, collection of raw cashew, unavailability of warehouse, marketing of cashew kernels, lack of technical knowledge and especially processing technology and packaging of finished products. This industry has great prospects such as scope for more cashew cultivation and cashew production, employment generation, formation of cashew processing units, alcohols production from cashew apple, shield oil production, rural development, poverty elimination, development of social and economic backward class and environment protection etc. This industry has domestic as well as foreign market; India has tremendous potential in this regard. The cashew is a poor men’s crop but rich men’s food. The cashew is a source of income and livelihood for poor farmers. Cashew-nut industry may play very important role in the development of hilly region. The objectives of this paper are to identify problems of cashew processing and use of processing technology, problems of cashew kernel packaging, evolving of cashew processing technology over the year and its impact on final product and impact of good processing by adopting appropriate technology packaging on international trade of cashew-nut. The most important problem of cashew processing industry is that is processing and packaging. Bad processing reduce the quality of cashew kernel at large extent especially broken of cashew kernel which has very less price in market compare to whole cashew kernel and not eligible for export. On the other hand if there is no good packaging of cashew kernel will get moisture which destroy test of it. International trade of cashew-nut is depend of two things one is cashew processing and other is packaging. This study has strong relevance because cashew-nut industry is the labour oriented, where processing technology is not playing important role because 95% processing work is manual. Hence processing work was depending on physical performance of worker which makes presence of large workforce inevitable. There are many cashew processing units closed because they are not getting sufficient work force. However due to advancement in technology slowly this picture is changing and processing work get improve. Therefore it is interesting to explore all the aspects in context of cashew processing and packaging of cashew business.

Keywords: cashew, processing technology, packaging, international trade, change

Procedia PDF Downloads 391

Authors: Shafiullah Soomro

Abstract:

Active contour is one of the image segmentation techniques and its goal is to capture required object boundaries within an image. In this paper, we propose a novel image segmentation method by using an active contour method based on anisotropic diffusion feature enhancement technique. The traditional active contour methods use only pixel information to perform segmentation, which produces inaccurate results when an image has some noise or complex background. We use Perona and Malik diffusion scheme for feature enhancement, which sharpens the object boundaries and blurs the background variations. Our main contribution is the formulation of a new SPF (signed pressure force) function, which uses global intensity information across the regions. By minimizing an energy function using partial differential framework the proposed method captures semantically meaningful boundaries instead of catching uninterested regions. Finally, we use a Gaussian kernel which eliminates the problem of reinitialization in level set function. We use several synthetic and real images from different modalities to validate the performance of the proposed method. In the experimental section, we have found the proposed method performance is better qualitatively and quantitatively and yield results with higher accuracy compared to other state-of-the-art methods.

Keywords: active contours, anisotropic diffusion, level-set, partial differential equations

Procedia PDF Downloads 138

Authors: Nelson Bii, Christopher Ouma, John Odhiambo

Abstract:

Non-response is a potential source of errors in sample surveys. It introduces bias and large variance in the estimation of finite population parameters. Regression models have been recognized as one of the techniques of reducing bias and variance due to random non-response using auxiliary data. In this study, it is assumed that random non-response occurs in the survey variable in the second stage of cluster sampling, assuming full auxiliary information is available throughout. Auxiliary information is used at the estimation stage via a regression model to address the problem of random non-response. In particular, the auxiliary information is used via an improved Nadaraya-Watson kernel regression technique to compensate for random non-response. The asymptotic bias and mean squared error of the estimator proposed are derived. Besides, a simulation study conducted indicates that the proposed estimator has smaller values of the bias and smaller mean squared error values compared to existing estimators of finite population mean. The proposed estimator is also shown to have tighter confidence interval lengths at a 95% coverage rate. The results obtained in this study are useful, for instance, in choosing efficient estimators of the finite population mean in demographic sample surveys.

Keywords: mean squared error, random non-response, two-stage cluster sampling, confidence interval lengths

Procedia PDF Downloads 106

Authors: Pratibha Srivastava, Chithra V. J., Sudhakar S., Nitin K. D.

Abstract:

A closed-loop system comprises of a controller, a response system, and an actuating system. The controller, which is the system under test for us, excites the actuators based on feedback from the sensors in a periodic manner. The sensors should provide the feedback to the System Under Test (SUT) within a deterministic time post excitation of the actuators. Any delay or miss in the generation of response or acquisition of excitation pulses may lead to control loop controller computation errors, which can be catastrophic in certain cases. Such systems categorised as hard real-time systems that need special strategies. The real-time operating systems available in the market may be the best solutions for such kind of simulations, but they pose limitations like the availability of the X Windows system, graphical interfaces, other user tools. In this paper, we present strategies that can be used on a general purpose operating system (Bare Linux Kernel) to achieve a deterministic deadline and hence have the added advantages of a GPOS with real-time features. Techniques shall be discussed how to make the time-critical application run with the highest priority in an uninterrupted manner, reduced network latency for distributed architecture, real-time data acquisition, data storage, and retrieval, user interactions, etc.

Keywords: real time data acquisition, real time kernel preemption, scheduling, network latency

Procedia PDF Downloads 104

Authors: Nurul Aini Abdul Wahab, Nor Syamim Halidin, Sayidatina Aisah Masnan, Nur Izzati Romli

Abstract:

There are many applications of principle component method for reducing the large set of variables in various fields. Fisher’s Discriminant function is also a popular tool for classification. In this research, the researcher focuses on studying the performance of Principle Component-Fisher’s Discriminant function in helping to classify rice kernels to their defined classes. The data were collected on the smells or odour of the rice kernel using odour-detection sensor, Cyranose. 32 variables were captured by this electronic nose (e-nose). The objective of this research is to measure how well a combination model, between principle component and linear discriminant, to be as a classification model. Principle component method was used to reduce all 32 variables to a smaller and manageable set of components. Then, the reduced components were used to develop the Fisher’s Discriminant function. In this research, there are 4 defined classes of rice kernel which are Aromatic, Brown, Ordinary and Others. Based on the output from principle component method, the 32 variables were reduced to only 2 components. Based on the output of classification table from the discriminant analysis, 40.76% from the total observations were correctly classified into their classes by the PC-Discriminant function. Indirectly, it gives an idea that the classification model developed has committed to more than 50% of misclassifying the observations. As a conclusion, the Fisher’s Discriminant function that was built on a 2-component from PCA (2-PC-D) is not satisfying to classify the rice kernels into its defined classes.

Keywords: classification model, discriminant function, principle component analysis, variable reduction

Procedia PDF Downloads 304

Authors: Bowei Yuan, Shi Li, Liuyang Song, Huaqing Wang, Lingli Cui

Abstract:

Convolutional neural networks (CNN) have received extensive attention in the field of fault diagnosis. Many fault diagnosis methods use CNN for fault type identification. However, when the amount of raw data collected by sensors is massive, the neural network needs to perform a time-consuming classification task. In this paper, a mechanical fault diagnosis method based on vibration signal down-sampling and the improved one-dimensional convolutional neural network is proposed. Through the robust principal component analysis, the low-rank feature matrix of a large amount of raw data can be separated, and then down-sampling is realized to reduce the subsequent calculation amount. In the improved one-dimensional CNN, a smaller convolution kernel is used to reduce the number of parameters and computational complexity, and regularization is introduced before the fully connected layer to prevent overfitting. In addition, the multi-connected layers can better generalize classification results without cumbersome parameter adjustments. The effectiveness of the method is verified by monitoring the signal of the centrifugal pump test bench, and the average test accuracy is above 98%. When compared with the traditional deep belief network (DBN) and support vector machine (SVM) methods, this method has better performance.

Keywords: fault diagnosis, vibration signal down-sampling, 1D-CNN

Procedia PDF Downloads 101

Authors: Kaustav Mukherjee

Abstract:

In this paper we propose a Neural Style Transfer solution whereby we have created a Lightweight Separable Convolution Kernel Based GAN Architecture (SC-GAN) which will very useful for designing filter for Mobile Phone Cameras and also Edge Devices which will convert any image to its 2D ANIMATED COMIC STYLE Movies like HEMAN, SUPERMAN, JUNGLE-BOOK. This will help the 2D animation artist by relieving to create new characters from real life person's images without having to go for endless hours of manual labour drawing each and every pose of a cartoon. It can even be used to create scenes from real life images.This will reduce a huge amount of turn around time to make 2D animated movies and decrease cost in terms of manpower and time. In addition to that being extreme light-weight it can be used as camera filters capable of taking Comic Style Shots using mobile phone camera or edge device cameras like Raspberry Pi 4,NVIDIA Jetson NANO etc. Existing Methods like CartoonGAN with the model size close to 170 MB is too heavy weight for mobile phones and edge devices due to their scarcity in resources. Compared to the current state of the art our proposed method which has a total model size of 31 MB which clearly makes it ideal and ultra-efficient for designing of camera filters on low resource devices like mobile phones, tablets and edge devices running OS or RTOS. .Owing to use of high resolution input and usage of bigger convolution kernel size it produces richer resolution Comic-Style Pictures implementation with 6 times lesser number of parameters and with just 25 extra epoch trained on a dataset of less than 1000 which breaks the myth that all GAN need mammoth amount of data. Our network reduces the density of the Gan architecture by using Depthwise Separable Convolution which does the convolution operation on each of the RGB channels separately then we use a Point-Wise Convolution to bring back the network into required channel number using 1 by 1 kernel.This reduces the number of parameters substantially and makes it extreme light-weight and suitable for mobile phones and edge devices. The architecture mentioned in the present paper make use of Parameterised Batch Normalization Goodfellow etc al. (Deep Learning OPTIMIZATION FOR TRAINING DEEP MODELS page 320) which makes the network to use the advantage of Batch Norm for easier training while maintaining the non-linear feature capture by inducing the learnable parameters

Keywords: comic stylisation from camera image using GAN, creating 2D animated movie style custom stickers from images, depth-wise separable convolutional neural network for light-weight GAN architecture for EDGE devices, GAN architecture for 2D animated cartoonizing neural style, neural style transfer for edge, model distilation, perceptual loss

Procedia PDF Downloads 96

Authors: Afshin Keshvadi, Johari Endan, Haniff Harun, Desa Ahmad, Farah Saleena

Abstract:

Experiments were conducted to develop a model for analyzing the ripening process of oil palm fresh fruits in relation to oil yield and oil quality of palm oil produced. This research was carried out on 8-year-old Tenera (Dura × Pisifera) palms planted in 2003 at the Malaysian Palm Oil Board Research Station. Fresh fruit bunches were harvested from designated palms during January till May of 2010. The bunches were divided into three regions (top, middle and bottom), and fruits from the outer and inner layers were randomly sampled for analysis at 8, 12, 16 and 20 weeks after anthesis to establish relationships between maturity and oil development in the mesocarp and kernel. Computations on data related to ripening time, oil content and oil quality were performed using several computer software programs (MSTAT-C, SAS and Microsoft Excel). Nine nonlinear mathematical models were utilized using MATLAB software to fit the data collected. The results showed mean mesocarp oil percent increased from 1.24 % at 8 weeks after anthesis to 29.6 % at 20 weeks after anthesis. Fruits from the top part of the bunch had the highest mesocarp oil content of 10.09 %. The lowest kernel oil percent of 0.03 % was recorded at 12 weeks after anthesis. Palmitic acid and oleic acid comprised of more than 73 % of total mesocarp fatty acids at 8 weeks after anthesis, and increased to more than 80 % at fruit maturity at 20 weeks. The Logistic model with the highest R2 and the lowest root mean square error was found to be the best fit model.

Keywords: oil palm, oil yield, ripening process, anthesis, fatty acids, modeling

Procedia PDF Downloads 279

Authors: Mohammad A. Alia, Abdelfatah Aref Tamimi, Omaima N. A. Al-Allaf

Abstract:

This paper reviews a comparison study on the most common used authentication methods. Some of these methods are actually based on cryptography. In this study, we show the main cryptographic services. Also, this study presents a specific discussion about authentication service, since the authentication service is classified into several categorizes according to their methods. However, this study gives more about the real life example for each of the authentication methods. It talks about the simplest authentication methods as well about the available biometric authentication methods such as voice, iris, fingerprint, and face authentication.

Keywords: information security, cryptography, system access control, authentication, network security

Procedia PDF Downloads 431

Authors: Angga Pratama Herman, Muhammad Shahbaz, Suzana Yusup

Abstract:

Bottom ash is a solid waste from thermal power plant and it is usually disposed of into landfills and ash ponds. These disposal methods are not sustainable since new lands need to be acquired as the landfills and ash ponds are fill to its capacity. Bottom ash also classified as hazardous material that makes the disposal methods may have contributed to the environmental effect to the area. Hence, more research needs to be done to explore the potential of recycling the bottom ash as more useful product. The objective of this research is to explore the potential of utilizing bottom ash as catalyst in biomass steam gasification. In this research, bottom ash was used as catalyst in gasification of Palm Kernel Shell (PKS) using Thermo Gravimetric Analyzer coupled with mass spectrometry (TGA/MS). The effects of temperature (650 – 750 °C), particle size (0.5 – 1.0 mm) and bottom ash percentage (2 % - 10 %) were studied with and without steam. The experimental arrays were designed using expert method of Central Composite Design (CCD). Results show maximum yield of hydrogen gas was 34.3 mole % for gasification without steam and 61.4 Mole % with steam. Similar trend was observed for syngas production. The maximum syngas yield was 59.5 mole % for without steam and it reached up to 81.5 mole% with the use of steam. The optimal condition for both product gases was temperature 700 °C, particle size 0.75 mm and cool bottom ash % 0.06. In conclusion, the use of bottom ash as catalyst is possible for biomass steam gasification and the product gases composition are comparable with previous researches, however the results need to be validated for bench or pilot scale study.

Keywords: bottom ash, biomass steam gasification, catalyst, lab scale

Procedia PDF Downloads 258

Authors: Jiangxia Han, Liang Xue, Keda Chen

Abstract:

Despite the significant progress over the last decades in reservoir simulation using numerical discretization, meshing is complex. Moreover, the high degree of freedom of the space-time flow field makes the solution process very time-consuming. Therefore, we present Physics-Informed Convolutional Neural Networks(PICNN) as a hybrid scientific theory and data method for reservoir modeling. Besides labeled data, the model is driven by the scientific theories of the underlying problem, such as governing equations, boundary conditions, and initial conditions. PICNN integrates governing equations and boundary conditions into the network architecture in the form of a customized convolution kernel. The loss function is composed of data matching, initial conditions, and other measurable prior knowledge. By customizing the convolution kernel and minimizing the loss function, the neural network parameters not only fit the data but also honor the governing equation. The PICNN provides a methodology to model and history-match flow and transport problems in porous media. Numerical results demonstrate that the proposed PICNN can provide an accurate physical solution from a limited dataset. We show how this method can be applied in the context of a forward simulation for continuous problems. Furthermore, several complex scenarios are tested, including the existence of data noise, different work schedules, and different good patterns.

Keywords: convolutional neural networks, deep learning, flow and transport in porous media, physics-informed neural networks, reservoir simulation

Procedia PDF Downloads 100

Authors: Jorge Gonzalez-Camus

Abstract:

In this work is proved the existence of at least one minimal and maximal mild solutions to the Cauchy problem, for fractional evolution equation of neutral type, involving a general kernel. An operator A generating a resolvent family and integral resolvent family on a Banach space X and a kernel belonging to a large class appears in the equation, which covers many relevant cases from physics applications, in particular, the important case of time - fractional evolution equations of neutral type. The main tool used in this work was the Kuratowski measure of noncompactness and fixed point theorems, specifically Darbo-type, and an iterative method of lower and upper solutions, based in an order in X induced by a normal cone P. Initially, the equation is a Cauchy problem, involving a fractional derivate in Caputo sense. Then, is formulated the equivalent integral version, and defining a convenient functional, using the theory of resolvent families, and verifying the hypothesis of the fixed point theorem of Darbo type, give us the existence of mild solution for the initial problem. Furthermore, the existence of minimal and maximal mild solutions was proved through in an iterative method of lower and upper solutions, using the Azcoli-Arzela Theorem, and the Gronwall’s inequality. Finally, we recovered the case derivate in Caputo sense.

Keywords: fractional evolution equations, Volterra integral equations, minimal and maximal mild solutions, neutral type equations, non-local in time equations

Procedia PDF Downloads 142

Authors: Chaouki Ghenai

Abstract:

Computational fluid dynamics analysis of the burning of syngas fuels derived from biomass and plastic solid waste mixture through gasification process is presented in this paper. The syngas fuel is burned in gas turbine can combustor. Gas turbine can combustor with swirl is designed to burn the fuel efficiently and reduce the emissions. The main objective is to test the impact of the alternative syngas fuel compositions and lower heating value on the combustion performance and emissions. The syngas fuel is produced by blending Palm Kernel Shell (PKS) with Polyethylene (PE) waste via catalytic steam gasification (fluidized bed reactor). High hydrogen content syngas fuel was obtained by mixing 30% PE waste with PKS. The syngas composition obtained through the gasification process is 76.2% H2, 8.53% CO, 4.39% CO2 and 10.90% CH4. The lower heating value of the syngas fuel is LHV = 15.98 MJ/m3. Three fuels were tested in this study natural gas (100%CH4), syngas fuel and pure hydrogen (100% H2). The power from the combustor was kept constant for all the fuels tested in this study. The effect of syngas fuel composition and lower heating value on the flame shape, gas temperature, mass of carbon dioxide (CO2) and nitrogen oxides (NOX) per unit of energy generation is presented in this paper. The results show an increase of the peak flame temperature and NO mass fractions for the syngas and hydrogen fuels compared to natural gas fuel combustion. Lower average CO2 emissions at the exit of the combustor are obtained for the syngas compared to the natural gas fuel.

Keywords: CFD, combustion, emissions, gas turbine combustor, gasification, solid waste, syngas, waste to energy

Procedia PDF Downloads 562

Authors: Acacia A. Salomão, Willian L. Gomes da Silva, Gustavo G. Shimamoto, Matthieu Tubino

Abstract:

Biodiesel physical and chemical properties depend on the raw material composition used in its synthesis. Saturated fatty acid esters confer high oxidative stability, while unsaturated fatty acid esters improve the cold flow properties. In this study, an alternative vegetal source - the macauba kernel oil - was used in the biodiesel synthesis instead of conventional sources. Macauba can be collected from native palm trees and is found in several regions in Brazil. Its oil is a promising source when compared to several other oils commonly obtained from food products, such as soybean, corn or canola oil, due to its specific characteristics. However, the usage of biodiesel made from macauba oil alone is not recommended due to the difficulty of producing macauba in large quantities. For this reason, this project proposes the usage of blends of the macauba oil with conventional oils. These blends were prepared by mixing the macauba biodiesel with biodiesels obtained from soybean, corn, and from residual frying oil, in the following proportions: 20:80, 50:50 e 80:20 (w/w). Three parameters were evaluated, using the standard methods, in order to check the quality of the produced biofuel and its blends: oxidative stability, cold filter plugging point (CFPP), and iodine value. The induction period (IP) expresses the oxidative stability of the biodiesel, the CFPP expresses the lowest temperature in which the biodiesel flows through a filter without plugging the system and the iodine value is a measure of the number of double bonds in a sample. The biodiesels obtained from soybean, residual frying oil and corn presented iodine values higher than 110 g/100 g, low oxidative stability and low CFPP. The IP values obtained from these biodiesels were lower than 8 h, which is below the recommended standard value. On the other hand, the CFPP value was found within the allowed limit (5 ºC is the maximum). Regarding the macauba biodiesel, a low iodine value was observed (31.6 g/100 g), which indicates the presence of high content of saturated fatty acid esters. The presence of saturated fatty acid esters should imply in a high oxidative stability (which was found accordingly, with IP = 64 h), and high CFPP, but curiously the latter was not observed (-3 ºC). This behavior can be explained by looking at the size of the carbon chains, as 65% of this biodiesel is composed by short chain saturated fatty acid esters (less than 14 carbons). The high oxidative stability and the low CFPP of macauba biodiesel are what make this biofuel a promising source. The soybean, corn and residual frying oil biodiesels also have low CFPP, but low oxidative stability. Therefore the blends proposed in this work, if compared to the common biodiesels, maintain the flow properties but present enhanced oxidative stability.

Keywords: biodiesel, blends, macauba kernel oil, stability oxidative

Procedia PDF Downloads 502

Authors: Luh Eka Suryani, Purhadi

Abstract:

Poisson regression is a non-linear regression model with response variable in the form of count data that follows Poisson distribution. Modeling for a pair of count data that show high correlation can be analyzed by Poisson Bivariate Regression. Data, the number of infant mortality and maternal mortality, are count data that can be analyzed by Poisson Bivariate Regression. The Poisson regression assumption is an equidispersion where the mean and variance values are equal. However, the actual count data has a variance value which can be greater or less than the mean value (overdispersion and underdispersion). Violations of this assumption can be overcome by applying Generalized Poisson Regression. Characteristics of each regency can affect the number of cases occurred. This issue can be overcome by spatial analysis called geographically weighted regression. This study analyzes the number of infant mortality and maternal mortality based on conditions in East Java in 2016 using Geographically Weighted Bivariate Generalized Poisson Regression (GWBGPR) method. Modeling is done with adaptive bisquare Kernel weighting which produces 3 regency groups based on infant mortality rate and 5 regency groups based on maternal mortality rate. Variables that significantly influence the number of infant and maternal mortality are the percentages of pregnant women visit health workers at least 4 times during pregnancy, pregnant women get Fe3 tablets, obstetric complication handled, clean household and healthy behavior, and married women with the first marriage age under 18 years.

Keywords: adaptive bisquare kernel, GWBGPR, infant mortality, maternal mortality, overdispersion

Procedia PDF Downloads 130

Authors: Yifan Fan, Xudong Luo, Pingping Lin

Abstract:

An essential task in the field of artificial intelligence is to allow computers to interact with people through natural language. Therefore, researches such as virtual assistants and dialogue systems have received widespread attention from industry and academia. The response generation plays a crucial role in dialogue systems, so to push forward the research on this topic, this paper surveys various methods for response generation. We sort out these methods into three categories. First one includes finite state machine methods, framework methods, and instance methods. The second contains full-text indexing methods, ontology methods, vast knowledge base method, and some other methods. The third covers retrieval methods and generative methods. We also discuss some hybrid methods based knowledge and deep learning. We compare their disadvantages and advantages and point out in which ways these studies can be improved further. Our discussion covers some studies published in leading conferences such as IJCAI and AAAI in recent years.

Keywords: deep learning, generative, knowledge, response generation, retrieval

Procedia PDF Downloads 103

Authors: Xiongxiong You, Zhanwen Niu

Abstract:

Choosing an appropriate surrogate model plays an important role in surrogates-assisted evolutionary algorithms (SAEAs) since there are many types and different kernel functions in the surrogate model. In this paper, an adaptive selection of the best suitable surrogate model method is proposed to solve different kinds of expensive optimization problems. Firstly, according to the prediction residual error sum of square (PRESS) and different model selection strategies, the excellent individual surrogate models are integrated into multiple ensemble models in each generation. Then, based on the minimum root of mean square error (RMSE), the best suitable surrogate model is selected dynamically. Secondly, two methods with dynamic number of models and selection strategies are designed, which are used to show the influence of the number of individual models and selection strategy. Finally, some compared studies are made to deal with several commonly used benchmark problems, as well as a rotor system optimization problem. The results demonstrate the accuracy and robustness of the proposed method.

Keywords: adaptive selection, expensive optimization, rotor system, surrogates assisted evolutionary algorithms

Procedia PDF Downloads 114

Authors: M. Zare, S. Biau, M. Corq, Y. Roquelaure

Abstract:

A wide variety of observational methods have been developed to evaluate the ergonomic workloads in manufacturing. However, the precision and accuracy of these methods remain a subject of debate. The aims of this study were to develop biomechanical methods to evaluate ergonomic workloads and to compare them with observational methods. Two observational methods, i.e. SCANIA Ergonomic Standard (SES) and Rapid Upper Limb Assessment (RULA), were used to assess ergonomic workloads at two simulated workstations. They included four tasks such as tightening & loosening, attachment of tubes and strapping as well as other actions. Sensors were also used to measure biomechanical data (Inclinometers, Accelerometers, and Goniometers). Our findings showed that in assessment of some risk factors both RULA & SES were in agreement with the results of biomechanical methods. However, there was disagreement on neck and wrist postures. In conclusion, the biomechanical approach was more precise than observational methods, but some risk factors evaluated with observational methods were not measurable with the biomechanical techniques developed.

Keywords: ergonomic, observational method, biomechanical methods, workload

Procedia PDF Downloads 353

Authors: B. Sellami, M. Belloufi

Abstract:

Conjugate gradient methods are an important class of methods for unconstrained optimization, especially for large-scale problems. Recently, they have been much studied. In this paper, we propose a new two-parameter family of conjugate gradient methods for unconstrained optimization. The two-parameter family of methods not only includes the already existing three practical nonlinear conjugate gradient methods, but also has other family of conjugate gradient methods as subfamily. The two-parameter family of methods with the Wolfe line search is shown to ensure the descent property of each search direction. Some general convergence results are also established for the two-parameter family of methods. The numerical results show that this method is efficient for the given test problems. In addition, the methods related to this family are uniformly discussed.

Keywords: unconstrained optimization, conjugate gradient method, line search, global convergence

Procedia PDF Downloads 414