Search results for: clustering algorithms

Authors: Isabel Moscol, Carlos J. Díaz, Ciro Rodríguez

Abstract:

Hip replacement becomes necessary when a person suffers severe pain or considerable functional limitations and the best option to enhance their quality of life is through the replacement of the damaged joint. One of the main components in femoral prostheses is the stem which distributes the loads from the joint to the proximal femur. To preserve more bone stock and avoid weakening of the diaphysis, a short starting stem was selected, generated from the intramedullary morphology of the patient's femur. It ensures the implantability of the design and leads to geometric delimitation for personalized optimization with machine learning (ML) and metaheuristic algorithms. The present study attempts to design a cementless short stem to make the strain deviation before and after implantation close to zero, promoting its fixation and durability. Regression models developed to estimate the percentage change of maximum principal stresses were used as objective optimization functions by the metaheuristic algorithm. The latter evaluated different geometries of the short stem with the modification of certain parameters in oblique sections from the osteotomy plane. The optimized geometry reached a global stress shielding (SS) of 18.37% with a determination factor (R²) of 0.667. The predicted results favour implantability integration in the short stem optimization to effectively reduce SS in the proximal femur.

Keywords: machine learning techniques, metaheuristic algorithms, short-stem design, stress shielding, hip replacement

Procedia PDF Downloads 196

Authors: Keerti Kembhavi, M. R. Archana, V. Anjaneyappa

Abstract:

Most asphalt pavement condition evaluation use rating frameworks in which asphalt pavement distress is estimated by type, extent, and severity. Rating is carried out by the pavement condition rating (PCR), which is tedious and expensive. This paper presents the development of a low-cost technique for image pavement distress analysis that permits the identification of pothole and cracks. The paper explores the application of image processing tools for the detection of potholes and cracks. Longitudinal cracking and pothole are detected using Fuzzy-C- Means (FCM) and proceeded with the Spectral Theory algorithm. The framework comprises three phases, including image acquisition, processing, and extraction of features. A digital camera (Gopro) with the holder is used to capture pavement distress images on a moving vehicle. FCM classifier and Spectral Theory algorithms are used to compute features and classify the longitudinal cracking and pothole. The Matlab2016Ra Image preparing tool kit utilizes performance analysis to identify the viability of pavement distress on selected urban stretches of Bengaluru city, India. The outcomes of image evaluation with the utilization semi-computerized image handling framework represented the features of longitudinal crack and pothole with an accuracy of about 80%. Further, the detected images are validated with the actual dimensions, and it is seen that dimension variability is about 0.46. The linear regression model y=1.171x-0.155 is obtained using the existing and experimental / image processing area. The R2 correlation square obtained from the best fit line is 0.807, which is considered in the linear regression model to be ‘large positive linear association’.

Keywords: crack detection, pothole detection, spectral clustering, fuzzy-c-means

Procedia PDF Downloads 181

Authors: Chonlakorn Wiboonsiriruk, Ekachai Phaisangittisagul, Chadchai Srisurangkul, Itsuo Kumazawa

Abstract:

Public transportation is a crucial aspect of passenger transportation, with buses playing a vital role in the transportation service. Passenger counting is an essential tool for organizing and managing transportation services. However, manual counting is a tedious and time-consuming task, which is why computer vision algorithms are being utilized to make the process more efficient. In this study, different object detection algorithms combined with passenger tracking are investigated to compare passenger counting performance. The system employs the EfficientDet algorithm, which has demonstrated superior performance in terms of speed and accuracy. Our results show that the proposed system can accurately count passengers in varying conditions with an accuracy of 94%.

Keywords: computer vision, object detection, passenger counting, public transportation

Procedia PDF Downloads 155

Authors: Omar Bouhassoun, Cristian Garrido, César Hueso

Abstract:

The radiation shielding design is an optimization problem with multiple -constrained- objective functions (radiation dose, weight, price, etc.) that depend on several parameters (material, thickness, position, etc.). The classical approach for shielding design consists of a brute force trial-and-error process subject to previous designer experience. Therefore, the result is an empirical solution but not optimal, which can degrade the overall performance of the shielding. In order to automate the shielding design procedure, the IDOM Shielding Optimization Tool (ISOT) has been developed. This software combines optimization algorithms with the capabilities to read/write input files, run calculations, as well as parse output files for different radiation transport codes. In the first stage, the software was established to adjust the input files for two well-known Monte Carlo codes (MCNP and Serpent) and optimize the result (weight, volume, price, dose rate) using multi-objective genetic algorithms. Nevertheless, its modular implementation easily allows the inclusion of more radiation transport codes and optimization algorithms. The work related to the development of ISOT and its verification on a simple 3D multi-layer shielding problem using both MCNP and Serpent will be presented. ISOT looks very promising for achieving an optimal solution to complex shielding problems.

Keywords: optimization, shielding, nuclear, genetic algorithm

Procedia PDF Downloads 110

Authors: D. I. George Amalarethinam, A. Emima

Abstract:

Educational Data Mining (EDM) is one of the newest topics to emerge in recent years, and it is concerned with developing methods for analyzing various types of data gathered from the educational circle. EDM methods and techniques with machine learning algorithms are used to extract meaningful and usable information from huge databases. For scientists and researchers, realistic applications of Machine Learning in the EDM sectors offer new frontiers and present new problems. One of the most important research areas in EDM is predicting student success. The prediction algorithms and techniques must be developed to forecast students' performance, which aids the tutor, institution to boost the level of student’s performance. This paper examines various classification techniques in prediction methods and data mining tools used in EDM.

Keywords: classification technique, data mining, EDM methods, prediction methods

Procedia PDF Downloads 117

Authors: Bircan Demiral

Abstract:

Cognitive radio (CR) is the promising technology that addresses the spectrum scarcity problem for future wireless communications. Orthogonal Frequency Division Multiplexing (OFDM) technology provides more power band ratios for cognitive radio networks (CRNs). While CR is a solution to the spectrum scarcity, it also brings up the capacity problem. In this paper, a novel power allocation algorithm that aims at maximizing the sum capacity in the OFDM based cognitive radio networks is proposed. Proposed allocation algorithm is based on the previously developed water-filling algorithm. To reduce the computational complexity calculating in water filling algorithm, proposed algorithm allocates the total power according to each subcarrier. The power allocated to the subcarriers increases sum capacity. To see this increase, Matlab program was used, and the proposed power allocation was compared with average power allocation, water filling and general power allocation algorithms. The water filling algorithm performed worse than the proposed algorithm while it performed better than the other two algorithms. The proposed algorithm is better than other algorithms in terms of capacity increase. In addition the effect of the change in the number of subcarriers on capacity was discussed. Simulation results show that the increase in the number of subcarrier increases the capacity.

Keywords: cognitive radio network, OFDM, power allocation, water filling

Procedia PDF Downloads 137

Authors: Jeryes Danial, Yosi Ben Asher

Abstract:

Algorithms for controlling a swarm of robots is an active research field, out of which cattle herding is one of the most complex problems to solve. In this paper, we derive an independent herding algorithm that is specifically designed for a swarm of quadcopters. The algorithm works by devising flight trajectories that cause the cows to run-away in the desired direction and hence herd cows that are distributed in a given field towards a common gathering point. Unlike previously proposed swarm herding algorithms, this algorithm does not use a flocking model but rather stars each cow separately. The effectiveness of this algorithm is verified experimentally using a simulator. We use a special set of experiments attempting to demonstrate that the herding times of this algorithm correspond to field diameter small constant regardless of the number of cows in the field. This is an optimal result indicating that the algorithm groups the cows into intermediate groups and herd them as one forming ever closing bigger groups.

Keywords: swarm, independent, distributed, algorithm

Procedia PDF Downloads 177

Authors: Jing Wu, Wei Lv, Yibing Li, Yuanfan You

Abstract:

The separation of speech signals has become a research hotspot in the field of signal processing in recent years. It has many applications and influences in teleconferencing, hearing aids, speech recognition of machines and so on. The sounds received are usually noisy. The issue of identifying the sounds of interest and obtaining clear sounds in such an environment becomes a problem worth exploring, that is, the problem of blind source separation. This paper focuses on the under-determined blind source separation (UBSS). Sparse component analysis is generally used for the problem of under-determined blind source separation. The method is mainly divided into two parts. Firstly, the clustering algorithm is used to estimate the mixing matrix according to the observed signals. Then the signal is separated based on the known mixing matrix. In this paper, the problem of mixing matrix estimation is studied. This paper proposes an improved algorithm to estimate the mixing matrix for speech signals in the UBSS model. The traditional potential algorithm is not accurate for the mixing matrix estimation, especially for low signal-to noise ratio (SNR).In response to this problem, this paper considers the idea of an improved potential function method to estimate the mixing matrix. The algorithm not only avoids the inuence of insufficient prior information in traditional clustering algorithm, but also improves the estimation accuracy of mixing matrix. This paper takes the mixing of four speech signals into two channels as an example. The results of simulations show that the approach in this paper not only improves the accuracy of estimation, but also applies to any mixing matrix.

Keywords: DBSCAN, potential function, speech signal, the UBSS model

Procedia PDF Downloads 135

Authors: Sonveer Singh, Sanjay Agrawal, D. V. Avasthi, Jayant Shekhar

Abstract:

The aim of this paper is to improve the efficiency of photovoltaic thermal (PVT) system with the help of Genetic Algorithms with multi-objective function. There are some parameters that affect the efficiency of PVT system like depth and length of the channel, velocity of flowing fluid through the channel, thickness of the tedlar and glass, temperature of inlet fluid i.e. all above parameters are considered for optimization. An attempt has been made to the model and optimizes the parameters of glazed hybrid single channel PVT module when two objective functions have been considered separately. The two objective function for optimization of PVT module is overall electrical and thermal efficiency. All equations for PVT module have been derived. Using genetic algorithms (GAs), above two objective functions of the system has been optimized separately and analysis has been carried out for two cases. Two cases are: Case-I; Improvement in electrical and thermal efficiency when overall electrical efficiency is optimized, Case-II; Improvement in electrical and thermal efficiency when overall thermal efficiency is optimized. All the parameters that are used in genetic algorithms are the parameters that could be changed, and the non-changeable parameters, like solar radiation, ambient temperature cannot be used in the algorithm. It has been observed that electrical efficiency (14.08%) and thermal efficiency (19.48%) are obtained when overall thermal efficiency was an objective function for optimization. It is observed that GA is a very efficient technique to estimate the design parameters of hybrid single channel PVT module.

Keywords: genetic algorithm, energy, exergy, PVT module, optimization

Procedia PDF Downloads 605

Authors: Nicolae Bold, Daniel Nijloveanu

Abstract:

The idea of cropping-system is a method used by farmers. It is an environmentally-friendly method, protecting the natural resources (soil, water, air, nutritive substances) and increase the production at the same time, taking into account some crop particularities. The combination of this powerful method with the concepts of genetic algorithms results into a possibility of generating sequences of crops in order to form a rotation. The usage of this type of algorithms has been efficient in solving problems related to optimization and their polynomial complexity allows them to be used at solving more difficult and various problems. In our case, the optimization consists in finding the most profitable rotation of cultures. One of the expected results is to optimize the usage of the resources, in order to minimize the costs and maximize the profit. In order to achieve these goals, a genetic algorithm was designed. This algorithm ensures the finding of several optimized solutions of cropping-systems possibilities which have the highest profit and, thus, which minimize the costs. The algorithm uses genetic-based methods (mutation, crossover) and structures (genes, chromosomes). A cropping-system possibility will be considered a chromosome and a crop within the rotation is a gene within a chromosome. Results about the efficiency of this method will be presented in a special section. The implementation of this method would bring benefits into the activity of the farmers by giving them hints and helping them to use the resources efficiently.

Keywords: chromosomes, cropping, genetic algorithm, genes

Procedia PDF Downloads 428

Authors: Huijie Wang, Bin Zhang

Abstract:

In China, the middle reaches of the Yangtze River are well-developed, boasting a wealth of different types of watershed landscape. In this regard, landscape character assessment (LCA) can serve as a basis for protection, management and planning of trans-regional watershed landscape types. For this study, we chose the middle reaches of the Yangtze River in Wuhan metropolitan area as our study site, wherein the water system consists of rich variety in landscape types. We analyzed trans-regional data to cluster and identify types of landscape characteristics at two levels. 55 basins were analyzed as variables with topography, land cover and river system features in order to identify the watershed landscape character types. For watershed landscape, drainage density and degree of curvature were specified as special variables to directly reflect the regional differences of river system features. Then, we used the principal component analysis (PCA) method and hierarchical clustering algorithm based on the geographic information system (GIS) and statistical products and services solution (SPSS) to obtain results for clusters of watershed landscape which were divided into 8 characteristic groups. These groups highlighted watershed landscape characteristics of different river systems as well as key landscape characteristics that can serve as a basis for targeted protection of watershed landscape characteristics, thus helping to rationally develop multi-value landscape resources and promote coordinated development of trans-regions.

Keywords: GIS, hierarchical clustering, landscape character, landscape typology, principal component analysis, watershed

Procedia PDF Downloads 230

Authors: Xu LiYun, Briand Florent, Fan GuoLiang

Abstract:

The loop-layout design problem (LLDP) aims at optimizing the sequence of positioning of the machines around the cyclic production line. Traffic congestion is the usual criteria to minimize in this type of problem, i.e. the number of additional cycles spent by each part in the network until the completion of its required routing sequence of machines. This paper aims at applying several improvements mechanisms such as a positioned-based crossover operator for the Genetic Algorithm (GA) called a Two Points Crossover (TPC) and an offspring selection process. The performance of the improved GA is measured using well-known examples from literature and compared to other evolutionary algorithms. Good results show that GA can still be competitive for this type of problem against more recent evolutionary algorithms.

Keywords: crossover, genetic algorithm, layout design problem, loop-layout, manufacturing optimization

Procedia PDF Downloads 279

Authors: Melkamu A. Zeru, Yamral M. Warkaw, Aweke A. Mitku, Muluwerk Ayele

Abstract:

Introduction: Malaria is a severe health threat in the World, mainly in Africa. It is the major cause of health problems in which the risk of morbidity and mortality associated with malaria cases are characterized by spatial variations across the county. This study aimed to investigate the spatial patterns and predictors of malaria distribution in Ethiopia. Methods: A weighted sample of 15,239 individuals with rapid diagnosis tests was obtained from the Central Statistical Agency and Ethiopia malaria indicator survey of 2015. Global Moran's I and Moran scatter plots were used in determining the distribution of malaria cases, whereas the local Moran's I statistic was used in identifying exposed areas. In data manipulation, machine learning was used for variable reduction and statistical software R, Stata, and Python were used for data management and analysis. The auto logistics spatial binary regression model was used to investigate the predictors of malaria. Results: The final auto logistics regression model reported that male clients had a positive significant effect on malaria cases as compared to female clients [AOR=2.401, 95 % CI: (2.125 - 2.713)]. The distribution of malaria across the regions was different. The highest incidence of malaria was found in Gambela [AOR=52.55, 95%CI: (40.54-68.12)] followed by Beneshangul [AOR=34.95, 95%CI: (27.159 - 44.963)]. Similarly, individuals in Amhara [AOR=0.243, 95% CI:(0.1950.303],Oromiya[AOR=0.197,95%CI:(0.1580.244)],DireDawa[AOR=0.064,95%CI(0.049-0.082)],AddisAbaba[AOR=0.057,95%CI:(0.044-0.075)], Somali[AOR=0.077,95%CI:(0.059-0.097)], SNNPR[OR=0.329, 95%CI: (0.261- 0.413)] and Harari [AOR=0.256, 95%CI:(0.201 - 0.325)] were less likely to had low incidence of malaria as compared with Tigray. Furthermore, for a one-meter increase in altitude, the odds of a positive rapid diagnostic test (RDT) decrease by 1.6% [AOR = 0.984, 95% CI :( 0.984 - 0.984)]. The use of a shared toilet facility was found as a protective factor for malaria in Ethiopia [AOR=1.671, 95% CI: (1.504 - 1.854)]. The spatial autocorrelation variable changes the constant from AOR = 0.471 for logistic regression to AOR = 0.164 for auto logistics regression. Conclusions: This study found that the incidence of malaria in Ethiopia had a spatial pattern that is associated with socio-economic, demographic, and geographic risk factors. Spatial clustering of malaria cases had occurred in all regions, and the risk of clustering was different across the regions. The risk of malaria was found to be higher for those who live in soil floor-type houses as compared to those who live in cement or ceramics floor type. Similarly, households with thatched, metal and thin, and other roof-type houses have a higher risk of malaria than ceramic tiles roof houses. Moreover, using a protected anti-mosquito net reduced the risk of malaria incidence.

Keywords: malaria, Ethiopia, auto logistics, spatial model, spatial clustering

Procedia PDF Downloads 34

Authors: S. Ghanaraja, Subrata Ray, S. K. Nath

Abstract:

Particle reinforced Metal Matrix Composite (MMC) succeeds in synergizing the metallic matrix with ceramic particle reinforcements to result in improved strength, particularly at elevated temperatures, but adversely it affects the ductility of the matrix because of agglomeration and porosity. The present study investigates the outcome of tensile properties in a cast and hot forged composite reinforced simultaneously with coarse and fine particles. Nano-sized alumina particles have been generated by milling mixture of aluminum and manganese dioxide powders. Milled particles after drying are added to molten metal and the resulting slurry is cast. The microstructure of the composites shows good distribution of both the size categories of particles without significant clustering. The presence of nanoparticles along with coarser particles in a composite improves both strength and ductility considerably. Delay in debonding of coarser particles to higher stress is due to reduced mismatch in extension caused by increased strain hardening in presence of the nanoparticles. However, higher addition of powder mix beyond a limit results in deterioration of mechanical properties, possibly due to clustering of nanoparticles. The porosity in cast composite generally increases with the increasing addition of powder mix as observed during process and on forging it has got reduced. The base alloy and nanocomposites show improvement in flow stress which could be attributed to lowering of porosity and grain refinement as a consequence of forging.

Keywords: aluminium, alumina, nano-particle reinforced composites, porosity

Procedia PDF Downloads 249

Authors: Jihen Bennaceur, Wissem Zouaghi, Ali Mabrouk

Abstract:

The major aim of cyber threat intelligence (CTI) is to provide sophisticated knowledge about cybersecurity threats to ensure internal and external safeguards against modern cyberattacks. Inaccurate, incomplete, outdated, and invaluable threat intelligence is the main problem. Therefore, data analysis based on AI algorithms is one of the emergent solutions to overcome the threat of information-sharing issues. In this paper, we propose a supervised machine learning-based algorithm to improve threat information sharing by providing a sophisticated classification of cyber threats and data. Extensive simulations investigate the accuracy, precision, recall, f1-score, and support overall to validate the designed algorithm and to compare it with several supervised machine learning algorithms.

Keywords: threat information sharing, supervised learning, data classification, performance evaluation

Procedia PDF Downloads 149

Authors: Behzad Niknam, Kourosh Shahriar, Hassan Madani

Abstract:

This paper demonstrates the applicability of Hybrid Neural Networks that combine with back propagation networks (BPN) and Genetic Algorithms (GAs) for predicting the flame length of tunnel fire A hybrid neural network model has been developed to predict the flame length of tunnel fire based parameters such as Fire Heat Release rate, air velocity, tunnel width, height and cross section area. The network has been trained with experimental data obtained from experimental work. The hybrid neural network model learned the relationship for predicting the flame length in just 3000 training epochs. After successful learning, the model predicted the flame length.

Keywords: tunnel fire, flame length, ANN, genetic algorithm

Procedia PDF Downloads 643

Authors: Amirhossein Chambari

Abstract:

This article proposes a bi-objective model for the facility location problem subject to congestion (overcrowding). Motivated by implementations to locate servers in internet mirror sites, communication networks, one-server-systems, so on. This model consider for situations in which immobile (or fixed) service facilities are congested (or queued) by stochastic demand to behave as M/M/1/K queues. We consider for this problem two simultaneous perspectives; (1) Customers (desire to limit times of accessing and waiting for service) and (2) Service provider (desire to limit average facility idle-time). A bi-objective model is setup for facility location problem with two objective functions; (1) Minimizing sum of expected total traveling and waiting time (customers) and (2) Minimizing the average facility idle-time percentage (service provider). The proposed model belongs to the class of mixed-integer nonlinear programming models and the class of NP-hard problems. In addition, to solve the model, controlled elitist non-dominated sorting genetic algorithms (Controlled NSGA-II) and controlled elitist non-dominated ranking genetic algorithms (NRGA-I) are proposed. Furthermore, the two proposed metaheuristics algorithms are evaluated by establishing standard multiobjective metrics. Finally, the results are analyzed and some conclusions are given.

Keywords: bi-objective, facility location, queueing, controlled NSGA-II, NRGA-I

Procedia PDF Downloads 583

Authors: Fahimeh Palizban, Farshad Noravesh, Amir Hossein Saeidian, Mahya Mehrmohamadi

Abstract:

In the recent decade, the emergence of liquid biopsy has significantly improved cancer monitoring and detection. Dying cells, including those originating from tumors, shed their DNA into the blood and contribute to a pool of circulating fragments called cell-free DNA. Accordingly, identifying the tissue origin of these DNA fragments from the plasma can result in more accurate and fast disease diagnosis and precise treatment protocols. Open chromatin regions are important epigenetic features of DNA that reflect cell types of origin. Profiling these features by DNase-seq, ATAC-seq, and histone ChIP-seq provides insights into tissue-specific and disease-specific regulatory mechanisms. There have been several studies in the area of cancer liquid biopsy that integrate distinct genomic and epigenomic features for early cancer detection along with tissue of origin detection. However, multimodal analysis requires several types of experiments to cover the genomic and epigenomic aspects of a single sample, which will lead to a huge amount of cost and time. To overcome these limitations, the idea of predicting OCRs from WGS is of particular importance. In this regard, we proposed a computational approach to target the prediction of open chromatin regions as an important epigenetic feature from cell-free DNA whole genome sequence data. To fulfill this objective, local sequencing depth will be fed to our proposed algorithm and the prediction of the most probable open chromatin regions from whole genome sequencing data can be carried out. Our method integrates the signal processing method with sequencing depth data and includes count normalization, Discrete Fourie Transform conversion, graph construction, graph cut optimization by linear programming, and clustering. To validate the proposed method, we compared the output of the clustering (open chromatin region+, open chromatin region-) with previously validated open chromatin regions related to human blood samples of the ATAC-DB database. The percentage of overlap between predicted open chromatin regions and the experimentally validated regions obtained by ATAC-seq in ATAC-DB is greater than 67%, which indicates meaningful prediction. As it is evident, OCRs are mostly located in the transcription start sites (TSS) of the genes. In this regard, we compared the concordance between the predicted OCRs and the human genes TSS regions obtained from refTSS and it showed proper accordance around 52.04% and ~78% with all and the housekeeping genes, respectively. Accurately detecting open chromatin regions from plasma cell-free DNA-seq data is a very challenging computational problem due to the existence of several confounding factors, such as technical and biological variations. Although this approach is in its infancy, there has already been an attempt to apply it, which leads to a tool named OCRDetector with some restrictions like the need for highly depth cfDNA WGS data, prior information about OCRs distribution, and considering multiple features. However, we implemented a graph signal clustering based on a single depth feature in an unsupervised learning manner that resulted in faster performance and decent accuracy. Overall, we tried to investigate the epigenomic pattern of a cell-free DNA sample from a new computational perspective that can be used along with other tools to investigate genetic and epigenetic aspects of a single whole genome sequencing data for efficient liquid biopsy-related analysis.

Keywords: open chromatin regions, cancer, cell-free DNA, epigenomics, graph signal processing, correlation clustering

Procedia PDF Downloads 150

Authors: Ganesh Nandakumaran, Mehmet Karaata

Abstract:

A wave is a distributed execution, often made up of a broadcast phase followed by a feedback phase, requiring the participation of all the system processes before a particular event called decision is taken. Wave algorithms with one initiator such as the 1-wave algorithm have been shown to be very efficient for broadcasting messages in tree networks. Extensions of this algorithm broadcasting a sequence of waves using a single initiator have been implemented in algorithms such as the m-wave algorithm. However as the network size increases, having a single initiator adversely affects the message delivery times to nodes further away from the initiator. As a remedy, broadcast waves can be allowed to be initiated by multiple initiator nodes distributed across the network to reduce the completion time of broadcasts. These waves initiated by one or more initiator processes form a collection of waves covering the entire network. Solutions to global-snapshots, distributed broadcast and various synchronization problems can be solved efficiently using waves with multiple concurrent initiators. In this paper, we propose the first stabilizing multi-wave sequence algorithm implementing waves started by multiple initiator processes such that every process in the network receives at least one sequence of broadcasts. Due to being stabilizing, the proposed algorithm can withstand transient faults and do not require initialization. We view a fault as a transient fault if it perturbs the configuration of the system but not its program.

Keywords: distributed computing, multi-node broadcast, propagation of information with feedback and cleaning (PFC), stabilization, wave algorithms

Procedia PDF Downloads 504

Authors: Ahmed Negm, George Aggidis, Xiandong Ma

Abstract:

With the diverse nature of urban cities, customer demand patterns, landscape topologies or even seasonal weather trends; managing our water distribution networks (WDNs) has proved a complex task. These unpredictable circumstances manifest as pipe failures, intermittent supply and burst events thus adding to water loss, energy waste and increased carbon emissions. Whilst these events are unavoidable, advanced pressure management has proved an effective tool to control and mitigate them. Henceforth, water utilities have struggled with developing a real-time control method that is resilient when confronting the challenges of water distribution. In this paper we use deep reinforcement learning (DRL) algorithms as a novel pressure control strategy to minimise pressure violations and leakage under both burst and background leakage conditions. Agents based on asynchronous actor critic (A2C) and recurrent proximal policy optimisation (Recurrent PPO) were trained and compared to benchmarked optimisation algorithms (differential evolution, particle swarm optimisation. A2C manages to minimise leakage by 32.48% under burst conditions and 67.17% under background conditions which was the highest performance in the DRL algorithms. A2C and Recurrent PPO performed well in comparison to the benchmarks with higher processing speed and lower computational effort.

Keywords: deep reinforcement learning, pressure management, water distribution networks, leakage management

Procedia PDF Downloads 91

Authors: Haya Salah, Srinivas Sharan

Abstract:

Healthcare facilities use appointment systems to schedule their appointments and to manage access to their medical services. With the growing demand for outpatient care, it is now imperative to manage physician's time effectively. However, high variation in consultation duration affects the clinical scheduler's ability to estimate the appointment duration and allocate provider time appropriately. Underestimating consultation times can lead to physician's burnout, misdiagnosis, and patient dissatisfaction. On the other hand, appointment durations that are longer than required lead to doctor idle time and fewer patient visits. Therefore, a good estimation of consultation duration has the potential to improve timely access to care, resource utilization, quality of care, and patient satisfaction. Although the literature on factors influencing consultation length abound, little work has done to predict it using based data-driven approaches. Therefore, this study aims to predict consultation duration using supervised machine learning algorithms (ML), which predicts an outcome variable (e.g., consultation) based on potential features that influence the outcome. In particular, ML algorithms learn from a historical dataset without explicitly being programmed and uncover the relationship between the features and outcome variable. A subset of the data used in this study has been obtained from the electronic medical records (EMR) of four different outpatient clinics located in central Pennsylvania, USA. Also, publicly available information on doctor's characteristics such as gender and experience has been extracted from online sources. This research develops three popular ML algorithms (deep learning, random forest, gradient boosting machine) to predict the treatment time required for a patient and conducts a comparative analysis of these algorithms with respect to predictive performance. The findings of this study indicate that ML algorithms have the potential to predict the provider service time with superior accuracy. While the current approach of experience-based appointment duration estimation adopted by the clinic resulted in a mean absolute percentage error of 25.8%, the Deep learning algorithm developed in this study yielded the best performance with a MAPE of 12.24%, followed by gradient boosting machine (13.26%) and random forests (14.71%). Besides, this research also identified the critical variables affecting consultation duration to be patient type (new vs. established), doctor's experience, zip code, appointment day, and doctor's specialty. Moreover, several practical insights are obtained based on the comparative analysis of the ML algorithms. The machine learning approach presented in this study can serve as a decision support tool and could be integrated into the appointment system for effectively managing patient scheduling.

Keywords: clinical decision support system, machine learning algorithms, patient scheduling, prediction models, provider service time

Procedia PDF Downloads 121

Authors: Sara Hasani

Abstract:

This research focuses on natural sudden onset disasters characterised as ‘occurring with little or no warning and often cause excessive injuries far surpassing the national response capacities’. Based on the panel analysis of the historic record of 4,252 natural onset disasters between 1980 to 2015, a predictive method was developed to predict the human impact of the disaster (fatality, injured, homeless) with less than 3% of errors. The geographical dispersion of the disasters includes every country where the data were available and cross-examined from various humanitarian sources. The records were then filtered into 4252 records of the disasters where the five predictive variables (disaster type, HDI, DRI, population, and population density) were clearly stated. The procedure was designed based on a combination of pattern recognition techniques and rule-based clustering for prediction and discrimination analysis to validate the results further. The result indicates that there is a relationship between the disaster human impact and the five socio-economic characteristics of the affected country mentioned above. As a result, a framework was put forward, which could predict the disaster’s human impact based on their severity rank in the early hours of disaster strike. The predictions in this model were outlined in two worst and best-case scenarios, which respectively inform the lower range and higher range of the prediction. A necessity to develop the predictive framework can be highlighted by noticing that despite the existing research in literature, a framework for predicting the human impact and estimating the needs at the time of the disaster is yet to be developed. This can further be used to allocate the resources at the response phase of the disaster where the data is scarce.

Keywords: disaster management, natural disaster, pattern recognition, prediction

Procedia PDF Downloads 153

Authors: Seongho Han, Dongkwan Kim

Abstract:

The focus of regional innovation policies is shifting from a central government to local governments. The central government demands that regions enforce autonomous and responsible regional innovation policies and that regional governments seek for innovation policies fit for regional characteristics. However, the central government and local governments have not arrived yet at a conclusion on what innovation policies are appropriate for regional circumstances. In particular, even if each local government is trying to find regional innovation strategies that are based on the needs of a region, its innovation strategies turn out to be similar with those of other regions. This leads to a consequence that is inefficient not only at a national level, but also at a regional level. Existing researches on regional innovation types point out that there are remarkable differences in the types or characteristics of innovation among the regions of a nation. In addition they imply that there would be no expected innovation output in cases in which policies are enforced with ignoring such differences. This means that it is undesirable to enforce regional innovation policies under a single standard. This research, given this problem, aims to find out the characteristics and differences in innovation types among the regions in Korea and suggests appropriate policy implications by classifying such characteristics and differences. This research, given these objectives, classified regions in consideration of the various indicators that comprise the innovation suggested by existing related researches and illustrated policies based on such characteristics and differences. This research used recent data, mainly from 2012, and as a methodology, clustering analysis based on multiple factor analysis was applied. Supplementary researches on dynamically analyzing stability in regional innovation types, establishing systematic indicators based on the regional innovation theory, and developing additional indicators are necessary in the future.

Keywords: regional innovation policy, regional innovation type, region-based innovation, multiple factor analysis, clustering analysis

Procedia PDF Downloads 475

Authors: K. R. Roopesh Bharatwaj, Avinash Maharana, Favour Tobi Aborisade, Roger Young

Abstract:

Presently, there are few cases of complete automation of drones and its allied intelligence capabilities. In essence, the potential of the drone has not yet been fully utilized. This paper presents feasible methods to build an intelligent drone with smart capabilities such as self-driving, and obstacle avoidance. It does this through advanced Reinforcement Learning Techniques and performs object detection using latest advanced algorithms, which are capable of processing light weight models with fast training in real time instances. For the scope of this paper, after researching on the various algorithms and comparing them, we finally implemented the Deep-Q-Networks (DQN) algorithm in the AirSim Simulator. In future works, we plan to implement further advanced self-driving and object detection algorithms, we also plan to implement voice-based speech recognition for the entire drone operation which would provide an option of speech communication between users (People) and the drone in the time of unavoidable circumstances. Thus, making drones an interactive intelligent Robotic Voice Enabled Service Assistant. This proposed drone has a wide scope of usability and is applicable in scenarios such as Disaster management, Air Transport of essentials, Agriculture, Manufacturing, Monitoring people movements in public area, and Defense. Also discussed, is the entire drone communication based on the satellite broadband Internet technology for faster computation and seamless communication service for uninterrupted network during disasters and remote location operations. This paper will explain the feasible algorithms required to go about achieving this goal and is more of a reference paper for future researchers going down this path.

Keywords: convolution neural network, natural language processing, obstacle avoidance, satellite broadband technology, self-driving

Procedia PDF Downloads 251

Authors: Mohammad Ghavami, Reza S. Dilmaghani

Abstract:

This paper presents an adaptive framework for modelling financial markets using equity risk premiums, risk free rates and volatilities. The recorded economic factors are initially used to train four adaptive filters for a certain limited period of time in the past. Once the systems are trained, the adjusted coefficients are used for modelling and prediction of an important financial market index. Two different approaches based on least mean squares (LMS) and recursive least squares (RLS) algorithms are investigated. Performance analysis of each method in terms of the mean squared error (MSE) is presented and the results are discussed. Computer simulations carried out using recorded data show MSEs of 4% and 3.4% for the next month prediction using LMS and RLS adaptive algorithms, respectively. In terms of twelve months prediction, RLS method shows a better tendency estimation compared to the LMS algorithm.

Keywords: adaptive methods, LSE, MSE, prediction of financial Markets

Procedia PDF Downloads 336

Authors: Rohit Dey, Sailendra Karra

Abstract:

This paper aims at solving the problem of multi-target searching in a Global Positioning System (GPS) denied environment using swarm robots with limited sensing and communication abilities. Typically, existing swarm-based search algorithms rely on the presence of a global coordinate system (vis-à-vis, GPS) that is shared by the entire swarm which, in turn, limits its application in a real-world scenario. This can be attributed to the fact that robots in a swarm need to share information among themselves regarding their location and signal from targets to decide their future course of action but this information is only meaningful when they all share the same coordinate frame. The paper addresses this very issue by eliminating any dependency of a search algorithm on the need of a predetermined global coordinate frame by the unification of the relative coordinate of individual robots when within the communication range, therefore, making the system more robust in real scenarios. Our algorithm assumes that all the robots in the swarm are equipped with range and bearing sensors and have limited sensing range and communication abilities. Initially, every robot maintains their relative coordinate frame and follow Levy walk random exploration until they come in range with other robots. When two or more robots are within communication range, they share sensor information and their location w.r.t. their coordinate frames based on which we unify their coordinate frames. Now they can share information about the areas that were already explored, information about the surroundings, and target signal from their location to make decisions about their future movement based on the search algorithm. During the process of exploration, there can be several small groups of robots having their own coordinate systems but eventually, it is expected for all the robots to be under one global coordinate frame where they can communicate information on the exploration area following swarm search techniques. Using the proposed method, swarm-based search algorithms can work in a real-world scenario without GPS and any initial information about the size and shape of the environment. Initial simulation results show that running our modified-Particle Swarm Optimization (PSO) without global information we can still achieve the desired results that are comparable to basic PSO working with GPS. In the full paper, we plan on doing the comparison study between different strategies to unify the coordinate system and to implement them on other bio-inspired algorithms, to work in GPS denied environment.

Keywords: bio-inspired search algorithms, decentralized control, GPS denied environment, swarm robotics, target searching, unifying coordinate systems

Procedia PDF Downloads 137

Authors: Bliss Singhal

Abstract:

Epilepsy is a prevalent neurological disorder affecting approximately 50 million individuals worldwide and 1.2 million Americans. There exist millions of pediatric patients with intractable epilepsy, a condition in which seizures fail to come under control. The occurrence of seizures can result in physical injury, disorientation, unconsciousness, and additional symptoms that could impede children's ability to participate in everyday tasks. Predicting seizures can help parents and healthcare providers take precautions, prevent risky situations, and mentally prepare children to minimize anxiety and nervousness associated with the uncertainty of a seizure. This research proposes a comprehensive framework to predict seizures in pediatric patients by evaluating machine learning algorithms on unimodal neuroimaging data consisting of electroencephalogram signals. The bandpass filtering and independent component analysis proved to be effective in reducing the noise and artifacts from the dataset. Various machine learning algorithms’ performance is evaluated on important metrics such as accuracy, precision, specificity, sensitivity, F1 score and MCC. The results show that the deep learning algorithms are more successful in predicting seizures than logistic Regression, and k nearest neighbors. The recurrent neural network (RNN) gave the highest precision and F1 Score, long short-term memory (LSTM) outperformed RNN in accuracy and convolutional neural network (CNN) resulted in the highest Specificity. This research has significant implications for healthcare providers in proactively managing seizure occurrence in pediatric patients, potentially transforming clinical practices, and improving pediatric care.

Keywords: intractable epilepsy, seizure, deep learning, prediction, electroencephalogram channels

Procedia PDF Downloads 84

Authors: Ali Almadan, Anoop Krishnan, Ajita Rattani

Abstract:

Facial recognition is a widely adopted technology in surveillance, border control, healthcare, banking services, and lately, in mobile user authentication with Apple introducing “Face ID” moniker with iPhone X. A lot of research has been conducted in the area of face recognition on datasets captured by surveillance cameras, DSLR, and mobile devices. Recently, face recognition technology has also been deployed on body-worn cameras to keep officers safe, enabling situational awareness and providing evidence for trial. However, limited academic research has been conducted on this topic so far, without the availability of any publicly available datasets with a sufficient sample size. This paper aims to advance research in the area of face recognition using body-worn cameras. To this aim, the contribution of this work is two-fold: (1) collection of a dataset consisting of a total of 136,939 facial images of 102 subjects captured using body-worn cameras in in-door and daylight conditions and (2) evaluation of various deep-learning architectures for face identification on the collected dataset. Experimental results suggest a maximum True Positive Rate(TPR) of 99.86% at False Positive Rate(FPR) of 0.000 obtained by SphereFace based deep learning architecture in daylight condition. The collected dataset and the baseline algorithms will promote further research and development. A downloadable link of the dataset and the algorithms is available by contacting the authors.

Keywords: face recognition, body-worn cameras, deep learning, person identification

Procedia PDF Downloads 163

Authors: Victor Bodell, Lukas Ekstrom, Somayeh Aghanavesi

Abstract:

Fuel consumption (FC) is one of the key factors in determining expenses of operating a heavy-duty vehicle. A customer may therefore request an estimate of the FC of a desired vehicle. The modular design of heavy-duty vehicles allows their construction by specifying the building blocks, such as gear box, engine and chassis type. If the combination of building blocks is unprecedented, it is unfeasible to measure the FC, since this would first r equire the construction of the vehicle. This paper proposes a machine learning approach to predict FC. This study uses around 40,000 vehicles specific and o perational e nvironmental c onditions i nformation, such as road slopes and driver profiles. A ll v ehicles h ave d iesel engines and a mileage of more than 20,000 km. The data is used to investigate the accuracy of machine learning algorithms Linear regression (LR), K-nearest neighbor (KNN) and Artificial n eural n etworks (ANN) in predicting fuel consumption for heavy-duty vehicles. Performance of the algorithms is evaluated by reporting the prediction error on both simulated data and operational measurements. The performance of the algorithms is compared using nested cross-validation and statistical hypothesis testing. The statistical evaluation procedure finds that ANNs have the lowest prediction error compared to LR and KNN in estimating fuel consumption on both simulated and operational data. The models have a mean relative prediction error of 0.3% on simulated data, and 4.2% on operational data.

Keywords: artificial neural networks, fuel consumption, friedman test, machine learning, statistical hypothesis testing

Procedia PDF Downloads 178

Authors: James Jr. Mashiyane, Risuna Nkolele, Stephanie J. Müller, Gciniwe S. Dlamini, Rebone L. Meraba, Darlington S. Mapiye

Abstract:

When performing language tasks in natural language processing (NLP), the dimensionality of word embeddings is chosen either ad-hoc or is calculated by optimizing the Pairwise Inner Product (PIP) loss. The PIP loss is a metric that measures the dissimilarity between word embeddings, and it is obtained through matrix perturbation theory by utilizing the unitary invariance of word embeddings. Unlike in natural language, in genomics, especially in genome sequence processing, unlike in natural language processing, there is no notion of a “word,” but rather, there are sequence substrings of length k called k-mers. K-mers sizes matter, and they vary depending on the goal of the task at hand. The dimensionality of word embeddings in NLP has been studied using the matrix perturbation theory and the PIP loss. In this paper, the sufficiency and reliability of applying word-embedding algorithms to various genomic sequence datasets are investigated to understand the relationship between the k-mer size and their embedding dimension. This is completed by studying the scaling capability of three embedding algorithms, namely Latent Semantic analysis (LSA), Word2Vec, and Global Vectors (GloVe), with respect to the k-mer size. Utilising the PIP loss as a metric to train embeddings on different datasets, we also show that Word2Vec outperforms LSA and GloVe in accurate computing embeddings as both the k-mer size and vocabulary increase. Finally, the shortcomings of natural language processing embedding algorithms in performing genomic tasks are discussed.

Keywords: word embeddings, k-mer embedding, dimensionality reduction

Procedia PDF Downloads 138