Search results for: optimization for learning and data analysis

Authors: Xiaocong Liu, Huazhen Wang, Ting He, Xiaozheng Li, Weihan Zhang, Jian Chen

Abstract:

The utilization of electronic medical record (EMR) data to establish the disease diagnosis model has become an important research content of biomedical informatics. Deep learning can automatically extract features from the massive data, which brings about breakthroughs in the study of EMR data. The challenge is that deep learning lacks semantic knowledge, which leads to impracticability in medical science. This research proposes a method of incorporating lexical-semantic knowledge from abundant entities into a convolutional neural network (CNN) framework for pediatric disease diagnosis. Firstly, medical terms are vectorized into Lexical Semantic Vectors (LSV), which are concatenated with the embedded word vectors of word2vec to enrich the feature representation. Secondly, the semantic distribution of medical terms serves as Semantic Decision Guide (SDG) for the optimization of deep learning models. The study evaluates the performance of LSV-SDG-CNN model on four kinds of Chinese EMR datasets. Additionally, CNN, LSV-CNN, and SDG-CNN are designed as baseline models for comparison. The experimental results show that LSV-SDG-CNN model outperforms baseline models on four kinds of Chinese EMR datasets. The best configuration of the model yielded an F1 score of 86.20%. The results clearly demonstrate that CNN has been effectively guided and optimized by lexical-semantic knowledge, and LSV-SDG-CNN model improves the disease classification accuracy with a clear margin.

Keywords: lexical semantics, feature representation, semantic decision, convolutional neural network, electronic medical record

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 539

Authors: Zeshan Ahmad, Matteo Zoppi, Rezia Molfino

Abstract:

The geometric errors in the manufacturing process can be reduced by optimal positioning of the fixture elements in the fixture to make the workpiece stiff. We propose a new fixture layout optimization method N-3-2-1 for large metal sheets in this paper that combines the genetic algorithm and finite element analysis. The objective function in this method is to minimize the sum of the nodal deflection normal to the surface of the workpiece. Two different kinds of case studies are presented, and optimal position of the fixturing element is obtained for different cases.

Keywords: Fixture layout, optimization, fixturing element, genetic algorithm.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2245

Authors: Konstantinos Sofianos, Michail Stefanidakis

Abstract:

Nowadays, despite the great advances in technology, most educational frameworks lack a strong educational design basis. E-learning has become prevalent, but it faces various challenges such as student isolation and lack of quality in the learning process. An intelligent learning system provides a student with educational material according to their learning background and learning preferences. It records full information about the student, such as demographic information, learning styles, and academic performance. This information allows the system to be fully adapted to the student’s needs. In this paper, we propose a framework and an ontology for music education, consisting of the learner model and all elements of the learning process (learning objects, teaching methods, learning activities, assessment). This framework can be integrated into an intelligent learning system and used for music education in schools for the development of professional skills and beyond.

Keywords: Intelligent learning systems, e-learning, music education, ontology, semantic web.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 394

Authors: M.Victoria Garcia-Camba, M.Isabel Garcia-Planas

Abstract:

The use of brain stem auditory evoked potential (BAEP) is a common way to study the hearing function of people, a way to learn the functionality of a part of the brain neuronal groups that intervene in the learning process by studying the behaviour of wave VI. The latest advances in neuroscience have revealed the existence of different brain activity in the learning process that can be highlighted through the use of innocuous, low-cost and easy-access techniques such as, among others, the BAEP that can help us to detect early possible neurodevelopmental difficulties for their subsequent assessment and cure. To date and the authors best knowledge, only the latency data obtained, observing the first to V waves and mainly in the left ear, were taken into account. This work shows that it is essential to consider both ears; with these latest data, it has been possible to diagnose more precisely some cases than with the previous data had been diagnosed as “normal”despite showing signs of some alteration that motivated the new consultation to the specialist.

Keywords: Ear, neurodevelopment, auditory evoked potentials, intervals of normality, learning disabilities.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 453

Authors: Min Kyeong Cha, Ju Yeon Mun, Seong Baeg Kim

Abstract:

Recently, ‘play of learning’ becomes important and is  emphasized as a useful learning tool. Therefore, interest in  edutainment contents is growing. Storytelling is considered first as a  method that improves the transmission of information and learner's  interest when planning edutainment contents. In this study, we  designed edutainment contents in the form of an adventure game that  applies the storytelling method. This content provides questions and  items constituted dynamically and reorganized learning contents  through analysis of test results. It allows learners to solve various  questions through effective iterative learning. As a result, the learners  can reach mastery learning.

 

Keywords: Storytelling, edutainment, mastery learning, computer operating principle.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1814

Authors: Hesheng Wang, Haoyu Wang, Chungang Zhuang

Abstract:

Estimating the 6D pose of objects is a core step for robot bin-picking tasks. The problem is that various objects are usually randomly stacked with heavy occlusion in real applications. In this work, we propose a method to regress 6D poses by predicting three points for each object in the 3D point cloud through deep learning. To solve the ambiguity of symmetric pose, we propose a labeling method to help the network converge better. Based on the predicted pose, an iterative method is employed for pose optimization. In real-world experiments, our method outperforms the classical approach in both precision and recall.

Keywords: Pose estimation, deep learning, point cloud, bin-picking, 3D computer vision.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1753

Authors: M. Hosseinzadeh, S. A. Zamani, A. Taheri

Abstract:

During the last few years, several sheet hydroforming processes have been introduced. Despite the advantages of these methods, they have some limitations. Of the processes, the two main ones are the standard hydroforming and hydromechanical deep drawing. A new sheet hydroforming die set was proposed that has the advantages of both processes and eliminates their limitations. In this method, a polyurethane plate was used as a part of the die-set to control the blank holder force. This paper outlines the Taguchi optimization methodology, which is applied to optimize the effective parameters in forming cylindrical cups by the new die set of sheet hydroforming process. The process parameters evaluated in this research are polyurethane hardness, polyurethane thickness, forming pressure path and polyurethane hole diameter. The design of experiments based upon L9 orthogonal arrays by Taguchi was used and analysis of variance (ANOVA) was employed to analyze the effect of these parameters on the forming pressure. The analysis of the results showed that the optimal combination for low forming pressure is harder polyurethane, bigger diameter of polyurethane hole and thinner polyurethane. Finally, the confirmation test was derived based on the optimal combination of parameters and it was shown that the Taguchi method is suitable to examine the optimization process.

Keywords: Sheet Hydroforming, Optimization, Taguchi Method

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2563

Authors: O. Hasançebi, S. Kazemzadeh Azad

Abstract:

In the present study the efficiency of Big Bang-Big Crunch (BB-BC) algorithm is investigated in discrete structural design optimization. It is shown that a standard version of the BB-BC algorithm is sometimes unable to produce reasonable solutions to problems from discrete structural design optimization. Two reformulations of the algorithm, which are referred to as modified BB-BC (MBB-BC) and exponential BB-BC (EBB-BC), are introduced to enhance the capability of the standard algorithm in locating good solutions for steel truss and frame type structures, respectively. The performances of the proposed algorithms are experimented and compared to its standard version as well as some other algorithms over several practical design examples. In these examples, steel structures are sized for minimum weight subject to stress, stability and displacement limitations according to the provisions of AISC-ASD.

Keywords: Structural optimization, discrete optimization, metaheuristics, big bang-big crunch (BB-BC) algorithm, design optimization of steel trusses and frames.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2354

Authors: D. K. Lee, S. M. Shin, J. H. Lee

Abstract:

Topology optimization technique utilizes constant element densities as design parameters. Finally, optimal distribution contours of the material densities between voids (0) and solids (1) in design domain represent the determination of topology. It means that regions with element density values become occupied by solids in design domain, while there are only void phases in regions where no density values exist. Therefore the void regions of topology optimization results provide design information to decide appropriate depositions of web-opening in structure. Contrary to the basic objective of the topology optimization technique which is to obtain optimal topology of structures, this present study proposes a new idea that topology optimization results can be also utilized for decision of proper web-opening’s position. Numerical examples of linear elastostatic structures demonstrate efficiency of methodological design processes using topology optimization in order to determinate the proper deposition of web-openings.

Keywords: Topology optimization, web-opening, structure, element density, material.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1395

Authors: Mária Stachová, Lukáš Sobíšek

Abstract:

Data mining incorporates a group of statistical methods used to analyze a set of information, or a data set. It operates with models and algorithms, which are powerful tools with the great potential. They can help people to understand the patterns in certain chunk of information so it is obvious that the data mining tools have a wide area of applications. For example in the theoretical chemistry data mining tools can be used to predict moleculeproperties or improve computer-assisted drug design. Classification analysis is one of the major data mining methodologies. The aim of thecontribution is to create a classification model, which would be able to deal with a huge data set with high accuracy. For this purpose logistic regression, Bayesian logistic regression and random forest models were built using R software. TheBayesian logistic regression in Latent GOLD software was created as well. These classification methods belong to supervised learning methods. It was necessary to reduce data matrix dimension before construct models and thus the factor analysis (FA) was used. Those models were applied to predict the biological activity of molecules, potential new drug candidates.

Keywords: data mining, classification, drug design, QSAR

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2813

Authors: H. Shayeghi, A. Safari, H. A. Shayanfar

Abstract:

In this paper, multiobjective design of multi-machine Power System Stabilizers (PSSs) using Particle Swarm Optimization (PSO) is presented. The stabilizers are tuned to simultaneously shift the lightly damped and undamped electro-mechanical modes of all machines to a prescribed zone in the s-plane. A multiobjective problem is formulated to optimize a composite set of objective functions comprising the damping factor, and the damping ratio of the lightly damped electromechanical modes. The PSSs parameters tuning problem is converted to an optimization problem which is solved by PSO with the eigenvalue-based multiobjective function. The proposed PSO based PSSs is tested on a multimachine power system under different operating conditions and disturbances through eigenvalue analysis and some performance indices to illustrate its robust performance.

Keywords: PSS Design, Particle Swarm Optimization, Dynamic Stability, Multiobjective Optimization.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2613

Authors: Phakhapong Thanitakul, Worawat Sa-ngiamvibool, Apinan Aurasopon, Saravuth Pothiya

Abstract:

This paper presents an improved ant colony optimization (IACO) for solving the reliability redundancy allocation problem (RAP) in order to maximize system reliability. To improve the performance of ACO algorithm, two additional techniques, i.e. neighborhood search, and re-initialization process are presented. To show its efficiency and effectiveness, the proposed IACO is applied to solve three RAPs. Additionally, the results of the proposed IACO are compared with those of the conventional heuristic approaches i.e. genetic algorithm (GA), particle swarm optimization (PSO) and ant colony optimization (ACO). The experimental results show that the proposed IACO approach is comparatively capable of obtaining higher quality solution and faster computational time.

Keywords: Ant colony optimization, Heuristic algorithm, Mixed-integer nonlinear programming, Redundancy allocation problem, Reliability optimization.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2068

Authors: Joko Pitono, Adi Soeprijanto, Takashi Hiyama

Abstract:

This paper present an efficient and reliable technique of optimization which combined fuel cost economic optimization and emission dispatch using the Sigmoid Decreasing Inertia Weight Particle Swarm Optimization algorithm (PSO) to reduce the cost of fuel and pollutants resulting from fuel combustion by keeping the output of generators, bus voltages, shunt capacitors and transformer tap settings within the security boundary. The performance of the proposed algorithm has been demonstrated on IEEE 30-bus system with six generating units. The results clearly show that the proposed algorithm gives better and faster speed convergence then linearly decreasing inertia weight.

Keywords: Optimal Power Flow, Combined Economic Emission Dispatch, Sigmoid decreasing Inertia Weight, Particle Swarm Optimization.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1653

Authors: Burcu Kaya, Jan-Martin Kaiser, Karl-Friedrich Becker, Tanja Braun, Klaus-Dieter Lang

Abstract:

Quality of the molded packages is strongly influenced by the process parameters of the transfer molding. To achieve a better package quality and a stable transfer molding process, it is necessary to understand the influence of the process parameters on the package quality. This work aims to comprehend the relationship between the process parameters, and to identify the optimum process parameters for the transfer molding process in order to achieve less voids and wire sweep. To achieve this, a DoE is executed for process optimization and a regression analysis is carried out. A systematic approach is represented to generate models which enable an estimation of the number of voids and wire sweep. Validation experiments are conducted to verify the model and the results are presented.

Keywords: Epoxy molding compounds, optimization, regression analysis, transfer molding process, voids, wire sweep.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1484

Authors: Suraiya Jabin, Kamal K. Bharadwaj

Abstract:

This research presents a system for post processing of data that takes mined flat rules as input and discovers crisp as well as fuzzy hierarchical structures using Learning Classifier System approach. Learning Classifier System (LCS) is basically a machine learning technique that combines evolutionary computing, reinforcement learning, supervised or unsupervised learning and heuristics to produce adaptive systems. A LCS learns by interacting with an environment from which it receives feedback in the form of numerical reward. Learning is achieved by trying to maximize the amount of reward received. Crisp description for a concept usually cannot represent human knowledge completely and practically. In the proposed Learning Classifier System initial population is constructed as a random collection of HPR–trees (related production rules) and crisp / fuzzy hierarchies are evolved. A fuzzy subsumption relation is suggested for the proposed system and based on Subsumption Matrix (SM), a suitable fitness function is proposed. Suitable genetic operators are proposed for the chosen chromosome representation method. For implementing reinforcement a suitable reward and punishment scheme is also proposed. Experimental results are presented to demonstrate the performance of the proposed system.

Keywords: Hierarchical Production Rule, Data Mining, Learning Classifier System, Fuzzy Subsumption Relation, Subsumption matrix, Reinforcement Learning.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1428

Authors: R. Akbari, M. A. Ehyaei, R. Shahi Shavvon

Abstract:

Nowadays, solar energy is used for energy purposes such as the use of thermal energy for domestic, industrial and power applications, as well as the conversion of the sunlight into electricity by photovoltaic cells. In this study, the thermodynamic simulation of the solar Rankin cycle with phase change material (paraffin) was first studied. Then energy and exergy analyses were performed. For optimization, a single and multi-objective genetic optimization algorithm to maximize thermal and exergy efficiency was used. The parameters discussed in this paper included the effects of input pressure on turbines, input mass flow to turbines, the surface of converters and collector angles on thermal and exergy efficiency. In the organic Rankin cycle, where solar energy is used as input energy, the fluid selection is considered as a necessary factor to achieve reliable and efficient operation. Therefore, silicon oil is selected for a high-temperature cycle and water for a low-temperature cycle as an operating fluid. The results showed that increasing the mass flow to turbines 1 and 2 would increase thermal efficiency, while it reduces and increases the exergy efficiency in turbines 1 and 2, respectively. Increasing the inlet pressure to the turbine 1 decreases the thermal and exergy efficiency, and increasing the inlet pressure to the turbine 2 increases the thermal efficiency and exergy efficiency. Also, increasing the angle of the collector increased thermal efficiency and exergy. The thermal efficiency of the system was 22.3% which improves to 33.2 and 27.2% in single-objective and multi-objective optimization, respectively. Also, the exergy efficiency of the system was 1.33% which has been improved to 1.719 and 1.529% in single-objective and multi-objective optimization, respectively. These results showed that the thermal and exergy efficiency in a single-objective optimization is greater than the multi-objective optimization.

Keywords: Exergy analysis, Genetic algorithm, Rankine cycle, Single and Multi-objective function.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 597

Authors: Maumita Bhattacharya

Abstract:

Optimization is often a critical issue for most system design problems. Evolutionary Algorithms are population-based, stochastic search techniques, widely used as efficient global optimizers. However, finding optimal solution to complex high dimensional, multimodal problems often require highly computationally expensive function evaluations and hence are practically prohibitive. The Dynamic Approximate Fitness based Hybrid EA (DAFHEA) model presented in our earlier work [14] reduced computation time by controlled use of meta-models to partially replace the actual function evaluation by approximate function evaluation. However, the underlying assumption in DAFHEA is that the training samples for the meta-model are generated from a single uniform model. Situations like model formation involving variable input dimensions and noisy data certainly can not be covered by this assumption. In this paper we present an enhanced version of DAFHEA that incorporates a multiple-model based learning approach for the SVM approximator. DAFHEA-II (the enhanced version of the DAFHEA framework) also overcomes the high computational expense involved with additional clustering requirements of the original DAFHEA framework. The proposed framework has been tested on several benchmark functions and the empirical results illustrate the advantages of the proposed technique.

Keywords: Evolutionary algorithm, Fitness function, Optimization, Meta-model, Stochastic method.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1543

Authors: Ehi E. Aimiuwu

Abstract:

As of 2017, many researchers in educational journals are still wondering if students are effectively and efficiently engaged in active learning in the online learning environment. The goal of this qualitative single case study and narrative research is to explore if students are actively engaged in their online learning. Seven online students in the United States from LinkedIn and residencies were interviewed for this study. Eleven online learning techniques from research were used as a framework.  Data collection tools were used for the study that included a digital audiotape, observation sheet, interview protocol, transcription, and NVivo 12 Plus qualitative software.  Data analysis process, member checking, and key themes were used to reach saturation. About 85.7% of students preferred individual grading. About 71.4% of students valued professor’s interacting 2-3 times weekly, participating through posts and responses, having good internet access, and using email.  Also, about 57.1% said students log in 2-3 times weekly to daily, professor’s social presence helps, regular punctuality in work submission, and prefer assessments style of research, essay, and case study.  About 42.9% appreciated syllabus usefulness and professor’s expertise.

Keywords: Class facilitation, course management, online teaching, online education, student engagement.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 654

Authors: Phanu Waraporn

Abstract:

This paper discusses the designing of knowledge integration of clinical information extracted from distributed medical ontologies in order to ameliorate a machine learning-based multilabel coding assignment system. The proposed approach is implemented using a decision tree technique of the machine learning on the university hospital data for patients with Coronary Heart Disease (CHD). The preliminary results obtained show a satisfactory finding that the use of medical ontologies improves the overall system performance.

Keywords: Medical Ontology, Knowledge Integration, Machine Learning, Medical Coding, Text Assignment.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1814

Authors: A. Rajagopal, S. Somasundaram, B. Sowmya, T. Suguna

Abstract:

Wireless Sensor Networks (WSNs) enable new applications and need non-conventional paradigms for the protocol because of energy and bandwidth constraints, In WSN, sensor node’s life is a critical parameter. Research on life extension is based on Low-Energy Adaptive Clustering Hierarchy (LEACH) scheme, which rotates Cluster Head (CH) among sensor nodes to distribute energy consumption over all network nodes. CH selection in WSN affects network energy efficiency greatly. This study proposes an improved CH selection for efficient data aggregation in sensor networks. This new algorithm is based on Bacterial Foraging Optimization (BFO) incorporated in LEACH.

Keywords: Bacterial Foraging Optimization (BFO), Cluster Head (CH), Data-aggregation protocols, Low-Energy Adaptive Clustering Hierarchy (LEACH).

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 3445

Authors: Anatoli Torokhti, Stan Miklavcic

Abstract:

We study a new technique for optimal data compression subject to conditions of causality and different types of memory. The technique is based on the assumption that some information about compressed data can be obtained from a solution of the associated problem without constraints of causality and memory. This allows us to consider two separate problem related to compression and decompression subject to those constraints. Their solutions are given and the analysis of the associated errors is provided.

Keywords: stochastic signals, optimization problems in signal processing.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1305

Authors: Vilas P. Mahatme, K. K. Bhoyar

Abstract:

Nowadays, education cannot be imagined without digital technologies. It broadens the horizons of teaching learning processes. Several universities are offering online courses. For evaluation purpose, e-examination systems are being widely adopted in academic environments. Multiple-choice tests are extremely popular. Moving away from traditional examinations to e-examination, Moodle as Learning Management Systems (LMS) is being used. Moodle logs every click that students make for attempting and navigational purposes in e-examination. Data mining has been applied in various domains including retail sales, bioinformatics. In recent years, there has been increasing interest in the use of data mining in e-learning environment. It has been applied to discover, extract, and evaluate parameters related to student’s learning performance. The combination of data mining and e-learning is still in its babyhood. Log data generated by the students during online examination can be used to discover knowledge with the help of data mining techniques. In web based applications, number of right and wrong answers of the test result is not sufficient to assess and evaluate the student’s performance. So, assessment techniques must be intelligent enough. If student cannot answer the question asked by the instructor then some easier question can be asked. Otherwise, more difficult question can be post on similar topic. To do so, it is necessary to identify difficulty level of the questions. Proposed work concentrate on the same issue. Data mining techniques in specific clustering is used in this work. This method decide difficulty levels of the question and categories them as tough, easy or moderate and later this will be served to the desire students based on their performance. Proposed experiment categories the question set and also group the students based on their performance in examination. This will help the instructor to guide the students more specifically. In short mined knowledge helps to support, guide, facilitate and enhance learning as a whole.

Keywords: Data mining, e-examination, e-learning, moodle.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2037

Authors: Rishi Ruttun

Abstract:

One of the major features of hypermedia learning is its non-linear structure, allowing learners, the opportunity of flexible navigation to accommodate their own needs. Nevertheless, such flexibility can also cause problems such as insufficient navigation and disorientation for some learners, especially those with Field Dependent cognitive styles. As a result students learning performance can be deteriorated and in turn, they can have negative attitudes with hypermedia learning systems. It was suggested that visual elements can be used to compensate dilemmas. However, it is unclear whether these visual elements improve their learning or whether problems still exist. The aim of this study is to investigate the effect of students cognitive styles and visual elements on students learning performance and attitudes in hypermedia learning environment. Cognitive Style Analysis (CSA), Learning outcome in terms of pre and post-test, practical task, and Attitude Questionnaire (AQ) were administered to a sample of 60 university students. The findings revealed that FD students preformed equally to those of FI. Also, FD students experienced more disorientation in the hypermedia learning system where they depend a lot on the visual elements for navigation and orientation purposes. Furthermore, they had more positive attitudes towards the visual elements which escape them from experiencing navigation and disorientation dilemmas. In contrast, FI students were more comfortable, did not get disturbed or did not need some of the visual elements in the hypermedia learning system.

Keywords: Hypermedia learning, cognitive styles, visual elements, support, learning performance, attitudes and perceptions

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1648

Authors: Victor Breux, Jérôme Boutet, Alain Goret, Viviane Cattin

Abstract:

Early detection of anomalies in data centers is important to reduce downtimes and the costs of periodic maintenance. However, there is little research on this topic and even fewer on the fusion of sensor data for the detection of abnormal events. The goal of this paper is to propose a method for anomaly detection in data centers by combining sensor data (temperature, humidity, power) and deep learning models. The model described in the paper uses one autoencoder per sensor to reconstruct the inputs. The auto-encoders contain Long-Short Term Memory (LSTM) layers and are trained using the normal samples of the relevant sensors selected by correlation analysis. The difference signal between the input and its reconstruction is then used to classify the samples using feature extraction and a random forest classifier. The data measured by the sensors of a data center between January 2019 and May 2020 are used to train the model, while the data between June 2020 and May 2021 are used to assess it. Performances of the model are assessed a posteriori through F1-score by comparing detected anomalies with the data center’s history. The proposed model outperforms the state-of-the-art reconstruction method, which uses only one autoencoder taking multivariate sequences and detects an anomaly with a threshold on the reconstruction error, with an F1-score of 83.60% compared to 24.16%.

Keywords: Anomaly detection, autoencoder, data centers, deep learning.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 683

Authors: Panpan Xu, Shulin Sui, Zongjie Du

Abstract:

Genetic algorithm is widely used in optimization problems for its excellent global search capabilities and highly parallel processing capabilities; but, it converges prematurely and has a poor local optimization capability in actual operation. Simulated annealing algorithm can avoid the search process falling into local optimum. A hybrid genetic algorithm based on simulated annealing is designed by combining the advantages of genetic algorithm and simulated annealing algorithm. The numerical experiment represents the hybrid genetic algorithm can be applied to solve the function optimization problems efficiently.

Keywords: Genetic algorithm, Simulated annealing, Hybrid genetic algorithm, Function optimization.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2512

Authors: Sajjad Ahmed, Knut Hinkelmann, Flavio Corradini

Abstract:

Fake news detection research is still in the early stage as this is a relatively new phenomenon in the interest raised by society. Machine learning helps to solve complex problems and to build AI systems nowadays and especially in those cases where we have tacit knowledge or the knowledge that is not known. We used machine learning algorithms and for identification of fake news; we applied three classifiers; Passive Aggressive, Naïve Bayes, and Support Vector Machine. Simple classification is not completely correct in fake news detection because classification methods are not specialized for fake news. With the integration of machine learning and text-based processing, we can detect fake news and build classifiers that can classify the news data. Text classification mainly focuses on extracting various features of text and after that incorporating those features into classification. The big challenge in this area is the lack of an efficient way to differentiate between fake and non-fake due to the unavailability of corpora. We applied three different machine learning classifiers on two publicly available datasets. Experimental analysis based on the existing dataset indicates a very encouraging and improved performance.

Keywords: Fake news detection, types of fake news, machine learning, natural language processing, classification techniques.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1461

Authors: Jeff Clarine, Chang-Shyh Peng, Daisy Sang

Abstract:

Bioassay is the measurement of the potency of a chemical substance by its effect on a living animal or plant tissue. Bioassay data and chemical structures from pharmacokinetic and drug metabolism screening are mined from and housed in multiple databases. Bioassay prediction is calculated accordingly to determine further advancement. This paper proposes a four-step preprocessing of datasets for improving the bioassay predictions. The first step is instance selection in which dataset is categorized into training, testing, and validation sets. The second step is discretization that partitions the data in consideration of accuracy vs. precision. The third step is normalization where data are normalized between 0 and 1 for subsequent machine learning processing. The fourth step is feature selection where key chemical properties and attributes are generated. The streamlined results are then analyzed for the prediction of effectiveness by various machine learning algorithms including Pipeline Pilot, R, Weka, and Excel. Experiments and evaluations reveal the effectiveness of various combination of preprocessing steps and machine learning algorithms in more consistent and accurate prediction.

Keywords: Bioassay, machine learning, preprocessing, virtual screen.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 952

Authors: Javad Marzbanrad, Mostafa Pahlavani

Abstract:

This study has investigated a vehicle Lumped Parameter Model (LPM) in frontal crash. There are several ways for determining spring and damper characteristics and type of problem shall be considered as system identification. This study use Genetic Algorithm (GA) procedure, being an effective procedure in case of optimization issues, for optimizing errors, between target data (experimental data) and calculated results (being obtained by analytical solving). In this study analyzed model in 5-DOF then compared our results with 5-DOF serial model. Finally, the response of model due to external excitement is investigated.

Keywords: Vehicle, Lumped-Parameter Model, GeneticAlgorithm, Optimization

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1644

Authors: Thi Nguyen, Lee Gordon-Brown, Jim Peterson, Peter Wheeler

Abstract:

An additive fuzzy system comprising m rules with n inputs and p outputs in each rule has at least t m(2n + 2 p + 1) parameters needing to be tuned. The system consists of a large number of if-then fuzzy rules and takes a long time to tune its parameters especially in the case of a large amount of training data samples. In this paper, a new learning strategy is investigated to cope with this obstacle. Parameters that tend toward constant values at the learning process are initially fixed and they are not tuned till the end of the learning time. Experiments based on applications of the additive fuzzy system in function approximation demonstrate that the proposed approach reduces the learning time and hence improves convergence speed considerably.

Keywords: Additive fuzzy system, improving convergence, parameter learning process, unsupervised learning.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1486

Authors: Jianwei Wu

Abstract:

Based on the one-bit-matching principle and by turning the de-mixing matrix into an orthogonal matrix via certain normalization, Ma et al proposed a one-bit-matching learning algorithm on the Stiefel manifold for independent component analysis [8]. But this algorithm is not adaptive. In this paper, an algorithm which can extract kurtosis and its sign of each independent source component directly from observation data is firstly introduced.With the algorithm , the one-bit-matching learning algorithm is revised, so that it can make the blind separation on the Stiefel manifold implemented completely in the adaptive mode in the framework of natural gradient.

Keywords: Independent component analysis, kurtosis, Stiefel manifold, super-gaussians or sub-gaussians.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1468