Amir Farzad and T. Aaron Gulliver Department of Electrical and Computer Engineering, University of Victoria, PO Box 1700, STN CSC, Victoria, BC Canada V8W 2Y2

ABSTRACT

Dealing with imbalanced data is one of the main challenges in machine/deep learning algorithms for classification. This issue is more important with log message data as it is typically very imbalanced and negative logs are rare. In this paper, a model is proposed to generate text log messages using a SeqGAN network. Then features are extracted using an Autoencoder and anomaly detection is done using a GRU network. The proposed model is evaluated with two imbalanced log data sets, namely BGL and Openstack. Results are presented which show that oversampling and balancing data increases the accuracy of anomaly detection and classification.

KEYWORDS

Deep Learning, Oversampling, Log messages, Anomaly detection, Classification

For More Details :
https://aircconline.com/csit/papers/vol10/csit100515.pdf

Volume Link :
http://airccse.org/csit/V10N05.html

Mashael Maashi, Nujood Alwhibi, Fatima Alamr, Rehab Alzahrani, Alanoud Alhamid and Nourah Altawallah Department of Software Engineering, College of Computer and Information Sciences, King Saud University, Riyadh, Kingdom of Saudi Arabia

ABSTRACT

When manufacturers equipment encounters an unexpected failure, or undergo unnecessary maintenance pre-scheduled plan, which happens for a total of millions of hours worldwide annually, this is time-consuming and costly. Predictive maintenance can help with the use of modern sensing technology and sophisticated data analytics to predict the maintenance required for machinery and devices. The demands of modern maintenance solutions have never been greater. The constant pressure to demonstrate enhanced cost-effectiveness return on investment and improve the competitiveness of the organization is always combined with the pressure of improving equipment productivity and keep machines running at the maximum output. In this paper, we propose maintenance prediction approach based on a machine learning technique namely random forest algorithm. The main focus is on the industrial duct fans as it is one of the most common equipment in most manufacturing industries. The experimental results show the accuracy, reliability of proposed Predictive Maintenance approach.

KEYWORDS

Predictive Maintenance, Maintenance, Random Forest, Duct Fan, Machine Learning & Artificial Intelligence

For More Details :
https://aircconline.com/csit/papers/vol10/csit100516.pdf

Volume Link :
http://airccse.org/csit/V10N05.html

Courtney Foots¹ , Palash Pal² , Rituparna Datta¹ and Aviv Segev^{1 1}Department of Computer Science, University of South Alabama, Mobile, USA ²University Institute of Technology, Burdwan University, India

ABSTRACT

We propose a set of methods to classify vendors based on estimated central processing unit (CPU) performance and predict CPU performance based on hardware components. For vendor classification, we use the highest and lowest estimated performance and frequency of occurrences of each vendor in the dataset to create classification zones. These zones can be used to list vendors who manufacture hardware that satisfy given performance requirements. We use multi-layered neural networks for performance prediction, which accounts for nonlinearity in performance data. Several neural network architectures are analysed in comparison to linear, quadratic, and cubic regression. Experiments show that neural networks can be used to obtain low prediction error and high correlation between predicted and published performance values, while cubic regression can produce higher correlation than neural networks when more data is used for training than testing. The proposed methods can be used to identify suitable hardware replacements.

KEYWORDS

Computer Hardware, Performance Prediction and Classification, Neural Networks, Statistical Learning, Regression

For More Details :
https://aircconline.com/csit/papers/vol10/csit100517.pdf

Volume Link :
http://airccse.org/csit/V10N05.html

Vijay Biju and Shahid Ali Department of Information Technology, AGI Institute, Auckland, New Zealand

ABSTRACT

Regression testing is very important for dynamic verification. It helps to simulate a suite of test cases periodically and after major changes in the design or its environment, to check that no new bugs were introduced. Evidences regarding benefit of implementing automation testing which includes saves of time and cost as it can re-run test scripts again and again and hence is much quicker than manual testing, providing more confidence in the quality of the product and increasing the ability to meet schedules and significantly reducing the effort that automation requires from testers are provided on the basis of survey of 115 software professionals. In addition to this, automated regression suite has an ability to explore the whole software every day without requiring much of manual effort. Also, bug identification is easier after the incorrect changes have been made. Genius is going through continuous development and requires testing again and again to check if new feature implementation have affected the existing functionality. In addition to this, Erudite is facing issue in validation of the Genius installation at client site since it requires availability of testers to check the critical functionality of the software manually. Erudite wants to create an automated regression suite for Genius which can be executed at client site for checking the functionality of the software. In addition to this, this suite will also help the testing team to validate if the new features which have been added to the existing software are affecting the existing system or not. Visual studio, Selenium Webdriver, Visual SVN and Trello are the tools which have been used to achieve the creation of automation regression suite. The current research will provide guidelines to the future researchers on how to create an automated regression suite for any web application using open source tools.

KEYWORDS

Automation testing, Regression testing, Visual Studio, C#, Selenium Webdriver, Agile- Scrum

For More Details :
https://aircconline.com/csit/papers/vol10/csit100610.pdf

Volume Link :
http://airccse.org/csit/V10N06.html

Sushma Suryadevara and Shahid Ali Department of Information Technology, AGI Institute, Auckland, New Zealand

ABSTRACT

This study was conducted on the importance of performance testing of web applications and analyzing the bottleneck applications. This paper highlights performance testing based on load tests. Everyone wants the application to be very fast, at the same time, reliability of the application also plays an important role, such that user’s satisfaction is the push for performance testing of a given application. Performance testing determines a few aspects of system performance under the pre-defined workload. In this study JMeter performance testing tool was used to implement and execute the test cases. The first load test was calculated with 200 users which was increased to 500 users and their throughput, median, average response time and deviation were calculated.

KEYWORDS

Performance testing, load balancing, threads, throughput, JMeter, load test

For More Details :
https://aircconline.com/csit/papers/vol10/csit100703.pdf

Volume Link :
http://airccse.org/csit/V10N07.html

Damir Krstinić, Maja Braović, Ljiljana Šerić and Dunja Božić-Štulić Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of Split, R. Boškovića 32, Split 21000, Croatia

ABSTRACT

Confusion matrix is a useful and comprehensive presentation of the classifier performance. It is commonly used in the evaluation of multi-class, single-label classification models, where each data instance can belong to just one class at any given point in time. However, the real world is rarely unambiguous and hard classification of data instance to a single class, i.e. defining its properties with single distinctive feature, is not always possible. For example, an image can contain multiple objects and regions which makes multi-class classification inappropriate to describe its content. Proposed solutions to this set of problems are based on multi-label classification model where each data instance is assigned one or more labels describing its features. While most of the evaluation measures used to evaluate single-label classifier can be adapted to a multi-label classification model, presentation and evaluation of the obtained results using standard confusion matrices cannot be expanded to this case.
In this paper we propose a novel method for the computation of a confusion matrix for multi-label classification. The proposed algorithm overcomes the limitations of the existing approaches in modeling relations between the classifier output and the Ground Truth (i.e. hand-labeled) classification, and due to its versatility can be used in many different research fields.

KEYWORDS

Classification, multi label classifier, performance evaluation, confusion matrix

For More Details :
https://aircconline.com/csit/papers/vol10/csit100801.pdf

Volume Link :
http://airccse.org/csit/V10N08.html

Piya Techateerawat Engineering Faculty, Thammasat University, Thailand

ABSTRACT

E-learning is a common tool to support the education in variety of scenarios. As the education content can be prepared by the group of specialists, but the skilled teachers are limited in remote area. Also, the contents in most curriculum are planned to distribute to limited skilled people. The gap of education can be full-filled with E- learning system. However, the conventional E-learning is high cost system and not appropriated for rural area. Also, opensource system is complicated to implement and configure in dedicated curriculum. This research is proposed the customized design of E-learning system for primary Thai education curriculum. As Thai education curriculum was updated in 2017, school and teacher need to update the plan and methodology accordingly. Our research is based on the actual Thai school in Ratchaburi province by using the Android framework. The system is designed and implemented from actual requirements from teachers and students in grade 3. As a result, the result show student involvement and continue using of system in both school and extra hours. The feedback from actual usage also is evaluated.

KEYWORDS

E-learning, Android, Thai Curriculum.

For More Details :
https://aircconline.com/csit/papers/vol10/csit100802.pdf

Volume Link :
http://airccse.org/csit/V10N08.html

Gopi Krishna Erabati, Nuno Gonçalves and Hélder Araújo Institute of Systems and Robotics, University of Coimbra, Portugal

ABSTRACT

In the area of computer vision, research on object detection algorithms has grown rapidly as it is the fundamental step for automation, specifically for self-driving vehicles. This work presents a comparison of traditional and deep learning approaches for the task of object detection in traffic scenarios. The handcrafted feature descriptor like Histogram of oriented Gradients (HOG) with a linear Support Vector Machine (SVM) classifier is compared with deep learning approaches like Single Shot Detector (SSD) and You Only Look Once (YOLO), in terms of mean Average Precision (mAP) and processing speed. SSD algorithm is implemented with different backbone architectures like VGG16, MobileNetV2 and ResNeXt50, similarly YOLO algorithm with MobileNetV1 and ResNet50, to compare the performance of the approaches. The training and inference is performed on PASCAL VOC 2007 and 2012 training, and PASCAL VOC 2007 test data respectively. We consider five classes relevant for traffic scenarios, namely, bicycle, bus, car, motorbike and person for the calculation of mAP. Both qualitative and quantitative results are presented for comparison. For the task of object detection, the deep learning approaches outperform the traditional approach both in accuracy and speed. This is achieved at the cost of requiring large amount of data, high computation power and time to train a deep learning approach.

KEYWORDS

Object Detection, Deep Learning, SVM, SSD & YOLO.

For More Details :
https://aircconline.com/csit/papers/vol10/csit100918.pdf

Volume Link :
http://airccse.org/csit/V10N09.html

Yusuf Yazici Department of Computer Engineering, Istanbul Technical University, Istanbul, Turkey

ABSTRACT

Credit card fraud is an ongoing problem for almost all industries in the world, and it raises millions of dollars to the global economy each year. Therefore, there is a number of research either completed or proceeding in order to detect these kinds of frauds in the industry. These researches generally use rule-based or novel artificial intelligence approaches to find eligible solutions. The ultimate goal of this paper is to summarize state-of-the-art approaches to fraud detection using artificial intelligence and machine learning techniques. While summarizing, we will categorize the common problems such as imbalanced dataset, real time working scenarios, and feature engineering challenges that almost all research works encounter, and identify general approaches to solve them. The imbalanced dataset problem occurs because the number of legitimate transactions is much higher than the fraudulent ones whereas applying the right feature engineering is substantial as the features obtained from the industries are limited, and applying feature engineering methods and reforming the dataset is crucial. Also, adapting the detection system to real time scenarios is a challenge since the number of credit card transactions in a limited time period is very high. In addition, we will discuss how evaluation metrics and machine learning methods differentiate among each research.

KEYWORDS

Credit Card, Fraud Detection, Machine Learning, Survey, Artificial Intelligence.

For More Details :
https://aircconline.com/csit/papers/vol10/csit101018.pdf

Volume Link :
http://airccse.org/csit/V10N10.html

Fabio Fassetti¹ and Ilaria Fassetti^{2 1}DIMES Dept., University of Calabria, Italy ²Therapeia, Rehabilitation Center, Italy

ABSTRACT

The paper presents a framework to evaluate the adequateness of a written text with respect to age or in presence of pathologies like deafness. This work aims at providing insights about verbal production level of an individual in order for a therapist to evaluate the adequateness of such level. The verbal production is analyzed by several points of view, categorized in six families: orthography, syntax, lexicon, lemmata, morphology, discourse. The proposed approach extract several features belonging to these categories through ad-hoc algorithms and exploits such features to train a learner able to classify verbal production in levels. This study is conducted in conjunction with a speech rehabilitation center. The technique is precisely designed for Italian language, however the methodology is more widely applicable. The proposed technique has a twofold aim. Other than the main goal of providing the therapist with an evaluation of the provided essay, the framework could spread lights on relationship between capabilities and ages.To the best of our knowledge, this is the first attempt to perform these evaluations through an automatic system.

KEYWORDS

Verbal production, Feature Extraction, Deep Learning.

For More Details :
https://aircconline.com/csit/papers/vol10/csit100109.pdf

Volume Link :
http://airccse.org/csit/V10N01.html

Malobika Roy Choudhury
Innovation and Technology, SAP Labs India Pvt Lmt., Bengaluru, Karnataka, India

ABSTRACT

Every product industry goes through the process of product validation before its release.Validation could be effortless or laborious depending upon the process. Here in this paper, aprocess is defined that can make the task-independent of constant monitoring. This method will not only make the work of test engineers easier it will also help the company meet stringent release deadlines with ease. The method explores how to complete visual validation of the display screen using deep learning and image processing. In the example, a method is discussed wrt a car-cluster display screen. The method breaks down the components of the screen then validates each component against its design and outputs a result predicting whether the displayed content is correct or incorrect. The models like You-Only-Live-Once, Machine Learning, Convolution Neural Networks-Conv2D, and image processing techniques like Hough circle/Hough lines are used to predict the accuracy of each display component. These sets of algorithms compile to provide consistent results throughout and are being currently used to generate results for the validation process.

KEYWORDS

Convolution Neural Networks, You-Only-Live-Once, display-validation

For More Details :
https://aircconline.com/csit/papers/vol10/csit101112.pdf

Volume Link :
http://airccse.org/csit/V10N11.html

Tatiana Tambouratzis^a, Laurent Pantera^b and Petr Stulik^c
aDepartment of Industrial Management & Technology, University of Piraeus, Piraeus 185 34, Greece ^bLaboratoire des Programmes Expérimentaux et des Essais en Sûreté, CEA/DES/IRESNE/DER/SPESI/LP2E/, Cadarache, F-13108 Saint-Paul-Lez-Durance, France ^cDiagnostics and Radiation Safety Department, ÚJV Řeža.s., Hlavní 130, Řež, 250 68 Husinec,Czech Republic

ABSTRACT

On-line monitoring (OLM) of nuclear reactors (NRs) incorporates – among other priorities – the concurrent verification of (i) valid operation of the NR neutron detectors (NDs) and (ii) soundness of the captured neutron noise (NN) signals (NSs) per se. In this piece of research, efficient, timely, directly reconfigurable and non-invasive OLM is implemented for providing swift – yet precise – decisions upon the (i) identities of malfunctioning NDs and(ii) locations of NR instability/unexpected operation. The use of Harmony Theory Networks (HTNs)is put forward to this end, with the results demonstrating the ability of these constraint-satisfaction artificial neural networks (ANNs) to identify(a) the smallest possible set of NDs which, configured into (b) the minimum number of 3-tuples of NDs operating on(c) the shortest NS time-window possible,instigate maximally efficient and accurate OLM. A proof-of-concept demonstration on the set of eight ex-core NDs and corresponding NSs of a simulated Pressurized Water nuclear Reactor (PWR) exhibits(i) significantly higher efficiency, at(ii) no detriment to localization accuracy, when employing only (iii) half of the original NDs and corresponding NSs, which are configured in (iv)a total of only two (out of the 56 combinatorially possible)3-tuples of NDs. Follow-up research shall investigate the scalability of the proposed methodology on the more extensive and homogeneous (i.e. “harder” in terms of ND/NS cardinality as well as of ranking/selection) dataset of the 36 in-core NSs of the same simulated NR.

KEYWORDS

Nuclear Reactor (NR),On-Line Monitoring (OLM), Neutron Noise (NN), Neutron Noise Signal (NS), Neutron Detector (ND), Computational Intelligence (CI),Artificial Neural Network (ANN),Harmony Theory Network (HTN), 3-tuple of NDs/NSs

For More Details :
https://aircconline.com/csit/papers/vol10/csit101113.pdf

Volume Link :
http://airccse.org/csit/V10N11.html

Christopher Gebauer and Maren Bennewitz
Humanoid Robots Lab, University of Bonn, Bonn, Germany

ABSTRACT

The recent progress in reinforcement learning algorithms enabled more complex tasks and, at the same time, enforced the need for a careful balance between exploration and exploitation. Enhanced exploration supersedes the requirement to hardly constrain the agent, e.g., with complex reward functions. This seems highly promising as it reduces the work for learning new tasks, while improving the agents performance. In this paper, we address deep exploration in reinforcement learning. Our approach is based on Thompson sampling and keeps multiple hypotheses of the posterior knowledge. We maintain the distribution over the hypotheses by a potential field based penalty function. The resulting policy is more performant in terms of collected reward. Furthermore, is our method faster in application and training than the current state of the art. We evaluate our approach in low-level robot control tasks to back up our claims of a more performant policy and faster training procedure

KEYWORDS

Deep Reinforcement Learning, Deep Exploration, Thompson Sampling, Bootstrapping

For More Details :
https://aircconline.com/csit/papers/vol10/csit101114.pdf

Volume Link :
http://airccse.org/csit/V10N11.html

Flavio Amadeu Bernardini¹ , Marcia Terra da Silva¹ , Jair Minoro Abe¹ , Luiz Antonio de Lima¹ and Kanstantsin Miatluk^2
1Graduate Program in Production Engineering Paulista University, Sao Paulo, Brazil ²Bialystok University of Technology, Bialystok, Poland

ABSTRACT

The recent progress in reinforcement learning algorithms enabled more complex tasks and, at the same time, enforced the need for a careful balance between exploration and exploitation. Enhanced exploration supersedes the requirement to hardly constrain the agent, e.g., with complex reward functions. This seems highly promising as it reduces the work for learning new tasks, while improving the agents performance. In this paper, we address deep exploration in reinforcement learning. Our approach is based on Thompson sampling and keeps multiple hypotheses of the posterior knowledge. We maintain the distribution over the hypotheses by a potential field based penalty function. The resulting policy is more performant in terms of collected reward. Furthermore, is our method faster in application and training than the current state of the art. We evaluate our approach in low-level robot control tasks to back up our claims of a more performant policy and faster training procedure

KEYWORDS

Deep Reinforcement Learning, Deep Exploration, Thompson Sampling, Bootstrapping

For More Details :
https://aircconline.com/csit/papers/vol10/csit101115.pdf

Volume Link :
http://airccse.org/csit/V10N11.html

Motoki Seguchi¹ , Fumiko Harada² and Hiromitsu Shimakawa^1
1College of Information Science and Engineering, Ritsumeikan University, Kusatsu, Shiga, Japan ²Connect Dot Ltd., Kyoto, Japan

ABSTRACT

In this research, we propose a method of recommending the best sightseeing spot through watching stories of sightseeing spots. It predicts the rating for each sightseeing spot of a target tourist based on Non-negative Matrix Factorization on the story watching times and ratings of tourists. We also propose to estimate the degree of the target tourist’s preference for a sightseeing spot. Tourists visit a sightseeing spot for a certain purpose of tourism. The preferences of tourists appear prominently in their purposes of tourism. In addition, the degree of the tourists’ preferences for sightseeing spots differs depending on the sightseeing spot. If we can estimate the degree of preference of a tourist, it will be possible to recommend a sightseeing spot that can achieve his purpose of tourism

KEYWORDS

Sightseeing, Recommendation, Interest Estimation, Story Watching, Preference

For More Details :
https://aircconline.com/csit/papers/vol10/csit101116.pdf

Volume Link :
http://airccse.org/csit/V10N11.html

Peter Regier Lukas Gesing Maren Bennewitz
Humanoid Robots Lab, University of Bonn, Bonn, Germany

ABSTRACT

Collision-free motion is essential for mobile robots. Most approaches to collision-free and efficient navigation with wheeled robots require parameter tuning by experts to obtain good navigation behavior. In this paper, we aim at learning an optimal navigation policy by deep reinforcement learning to overcome this manual parameter tuning. Our approach uses proximal policy optimization to train the policy and achieve collision-free and goal-directed behavior. The output of the learned network are the robot’s translational and angular velocities for the next time step. Our method combines path planning on a 2D grid with reinforcement learning and does not need any supervision. Our network is first trained in a simple environment and then transferred to scenarios of increasing complexity. We implemented our approach in C++ and Python for the Robot Operating System (ROS) and thoroughly tested it in several simulated as well as real-world experiments. The experiments illustrate that our trained policy can be applied to solve complex navigation tasks. Furthermore, we compare the performance of our learned controller to the popular dynamic window approach (DWA) of ROS. As the xperimental results show, a robot controlled by our learned policy reaches the goal significantly faster compared to using the DWA by closely bypassing obstacles and thus saving time.

For More Details :
https://aircconline.com/csit/papers/vol10/csit101117.pdf

Volume Link :
http://airccse.org/csit/V10N11.html

Khalid Amen¹ , Mohamed Zohdy¹ and Mohammed Mahmoud^2
1Department of Electrical and Computer Engineering, Oakland University, Rochester, MI, USA ²Department of Computer Science and Engineering, Oakland University, Rochester, MI, USA

ABSTRACT

According to the Centers for Disease Control and Prevention (CDC), heart disease is the number one cause of death for men, women, and people of most racial and ethnic groups in the United States. More than one person dies every minute and nearly half a million die each year from it, costing billions of dollars annually. Previous machine learning approaches have been used to predict whether patients have heart disease. The purpose of this work is to predict the five stages of heart disease starting from no disease, stage 1, stage 2, stage 3, and advance condition or severe heart disease. We investigate different potential supervised models that are trained by machine learning algorithms and find out which of these models has better accuracy. In this paper, we describe and investigate five machine learning algorithms (SVM, LR, RF, GTB, ERF) with hyper parameters that maximize classifier performance to show which one is the best to predict the stage at which a person is determined to have heart disease. We found that the LR algorithm performs better compared to the other four algorithms. The experiment results show that LR performs the best with an accuracy of 82%, followed by SVM with an accuracy of 80% when all five classifiers are compared and evaluated for performance based on accuracy, precision, recall, and F measure. This predication can facilitate every step of patient care, reducing the margin of error and contributing to precision medicine. Lastly, this paper aims to improve heart disease prediction accuracy, precision, recall and F measure using UCI heart disease dataset. For this, multiple machine learning approaches were used to understand the data and predict the chances of heart disease in a medical database

KEYWORDS

machine learning, ml, cnn, dnn, rnn, jupyter, python, cleveland dataset, gradient tree boosting, gtb, random forest, rf, support vector machine, svm, extra random forest, erf, logistic regression, lr.

For More Details :
https://aircconline.com/csit/papers/vol10/csit101118.pdf

Volume Link :
http://airccse.org/csit/V10N11.html

Reda Khalaf and Mireille Makary
Department of Computer Science and Information Technology, Lebanese International University, Beirut, Lebanon

ABSTRACT

Recent studies showed a huge interest in social networks sentiment analysis such as Twitter, to study how the users feel about a certain topic. In this paper, we conducted a sentiment analysis study for the tweets in spoken Lebanese Arabic related to the Lebanon Uprising hash tag (ينتفض_لبنان ,(#which was trending upon a socio-economic revolution that started in October, using different machine learning algorithms. The dataset was manually labelled to measure the precision and recall metrics and to compare between the different algorithms. Furthermore, the work completed in this paper provides two more contributions. The first is related to building a Lebanese – Modern Standard Arabic (فصحة (mapping dictionary and the second is an attempt to detect sarcastic and funny emotions in the tweets using emojis. The results we obtained seem satisfactory especially considering that there was no previous or similar work done involving Lebanese Arabic tweets, to our knowledge.

KEYWORDS

Lebanese Arabic tweets, sentiment analysis, machine learning, emotions, emojis

For More Details :
https://aircconline.com/csit/papers/vol10/csit101119.pdf

Volume Link :
http://airccse.org/csit/V10N11.html

Sheela Raju Kurupathi¹ and Wolfgang Maass¹, ^2
1German Research Center for Artificial Intelligence, Saarbrücken, Germany ²Saarland University, Saarbrücken, Germany

ABSTRACT

One of the significant challenges of Artificial Intelligence (AI) and Machine learning models is to preserve data privacy and to ensure data security. Addressing this problem lead to the application of Federated Learning (FL) mechanism towards preserving data privacy. Preserving user privacy in the European Union (EU) has to abide by the General Data Protection Regulation (GDPR). Therefore, exploring the machine learning models for preserving data privacy has to take into consideration of GDPR. In this paper, we present in detail understanding of Federated Machine Learning, various federated architectures along with different privacy-preserving mechanisms. The main goal of this survey work is to highlight the existing privacy techniques and also propose applications of Federated Learning in Industries. Finally, we also depict how Federated Learning is an emerging area of future research that would bring a new era in AI and Machine learning

KEYWORDS

Federated Learning, Artificial Intelligence, Machine Learning, Privacy, Security, Distributed Learning.

For More Details :
https://aircconline.com/csit/papers/vol10/csit101120.pdf

Volume Link :
http://airccse.org/csit/V10N11.html

Breno W. S. R. Carvalho¹ , Aline Paes² and Bernardo Gonçalves^3
1 IBM Research, Brazil. Institute of Computing, Universidade Federal Fluminense (UFF), Niterói, RJ, Brazil. ² Institute of Computing, Universidade Federal Fluminense (UFF), Niterói, RJ, Brazil. ³ IBM Research, Brazil

ABSTRACT

Semantic Role Labelling (SRL) is the process of automatically finding the semantic roles of terms in a sentence. It is an essential task towards creating a machine-meaningful representation of textual information. One public linguistic resource commonly used for this task is the FrameNet Project. FrameNet is a human and machine-readable lexical database containing a considerable number of annotated sentences, those annotations link sentence fragments to semantic frames. However, while the annotations across all the documents covered in the dataset link to most of the frames, a large group of frames lack annotations in the documents pointing to them. In this paper, we present a data augmentation method for FrameNet documents that increases by over 13% the total number of annotations. Our approach relies on lexical, syntactic, and semantic aspects of the sentences to provide additional annotations. We evaluate the proposed augmentation method by comparing the performance of a state-of-the-art semantic-role-labelling system, trained using a dataset with and without augmentation

KEYWORDS

FrameNet, Frame Semantic Parsing, Semantic Role Labelling, Data Augmentation.

For More Details :
https://aircconline.com/csit/papers/vol10/csit101201.pdf

Volume Link :
http://airccse.org/csit/V10N12.html

Chun-peng Chang, Wen-Jen Ho, Yung-chieh Hung, Kuei-Chun Chiang and Bill Zhao
Institute for Information Industry, Taipei, Taiwan

ABSTRACT

Based on neural network and machine learning, we apply the energy disaggregation for both classification (prediction on usage time) and estimation (prediction on usage amount) on 150 AMI (Advanced Metering Infrastructure) smart meters and a small amount of HEMS (Home Energy Management System) smart plugs in a community in New Taipei City, Taiwan. The aim of this paper is to clarify how we lower the cost, obtain the model of appliance usage from only a small portion of households, improve it with simple questionnaire, and generalize it for prediction on collective households. Our investigation demonstrates the benefits and various possibilities for power suppliers and the government, and won the Elite Award in the Presidential Hackathon 2020, Taiwan

KEYWORDS

Energy Disaggregation, Non-intrusive Load Monitoring, Deep Learning, Autoencoder

For More Details :
https://aircconline.com/csit/papers/vol10/csit101204.pdf

Volume Link :
http://airccse.org/csit/V10N12.html

Lirui (Harrison) Huang¹ , Yu Sun² , Fangyan Zhang^3
1University High School, Irvine, CA 92612 ²California State Polytechnic University, Pomona, CA, 91768 ³ASML, San Jose, CA, 95131

ABSTRACT

Since the moment human step into the high technology era, the world that people live in has changed subversively. As more and more unprecedented advancement being discovered, modern life of today’s people is now incredibly convenience. However, when the internet has played a pivotal role in everyday life, among the information that we are getting, countless of them are fraudulent. This paper designs a website to filter the influence of fake comments made to the product. We applied our application to the most commonly used shopping platform, Amazon, and conducted a qualitative evaluation of the approach. With a large amount of trained data, the sentiment analysis program can filter the fraudulent and odd comments and to give a more accurate score regarding a product by reading through each comment. This will definitely help consumers determine whether a product is fine or not.

KEYWORDS

Product review, website, machine learning, JavaScript, HTML

For More Details :
https://aircconline.com/csit/papers/vol10/csit101209.pdf

Volume Link :
http://airccse.org/csit/V10N12.html

Mahbubur Rahman
Department of Computer Science, North American University, Stafford, Texas, USA

ABSTRACT

Learning from the multidimensional data has been an interesting concept in the field of machine learning. However, such learning can be difficult, complex, expensive because of expensive data processing, manipulations as the number of dimension increases. As a result, we have introduced an ordered index-based data organization model as the ordered data set provides easy and efficient access than the unordered one and finally, such organization can improve the learning. The ordering maps the multidimensional dataset in the reduced space and ensures that the information associated with the learning can be retrieved back and forth efficiently. We have found that such multidimensional data storage can enhance the predictability for both the unsupervised and supervised machine learning algorithms.

KEYWORDS

Multidimensional, Euclidean norm, cosine similarity, database, model, hash table, index, Knearest neighbour, K-means clustering.

For More Details :
https://aircconline.com/csit/papers/vol10/csit101210.pdf

Volume Link :
http://airccse.org/csit/V10N12.html

Wilson Zhu¹ and Yu Sun^2
1Diamond Bar High School, Diamond Bar, California, USA ²California State Polytechnic University, Pomona, California, USA

ABSTRACT

Standardized testing such as the SAT often requires students to write essays and hires a large number of graders to evaluate these essays which can be time and cost consuming. Using natural language processing tools such as Global Vectors for word representation (GloVe), and various types of neural networks designed for picture classification, we developed an automatic grading system that is more time- and cost-efficient compared to human graders. We applied our application to a set of manually graded essays provided by a previous competition on Kaggle in 2012 on automated essay grading and conducted a qualitative evaluation of the approach. The result shows that the program is able to correctly score most of the essay and give an evaluation close to that of a human grader on the rest. The system proves itself to be effective in evaluating various essay prompts and capable of real-life application such as assisting another grader or even used as a standalone grader

KEYWORDS

Automated Essay Scoring System, Natural Language Processing, Multi-Model Machine Learning.

For More Details :
https://aircconline.com/csit/papers/vol10/csit101211.pdf

Volume Link :
http://airccse.org/csit/V10N12.html

Justin Kim¹ , Yu Sun² and Fangyan Zhang^3
1Los Osos High School, Rancho Cucamonga, CA 91739, USA ²California State Polytechnic University, Pomona, CA 91768, USA ³ASML, San Jose, CA 95131, USA

ABSTRACT

Recent years have seen a large increase in the number of programmers, especially as more online resources to learn became available. Many beginner coders struggle with bugs in their code, mostly as a result of a lack of knowledge and experience. The common approach is to have plenty of online resources that can address these issues. However, this is inconvenient to the coder, who may not have the time or patience to look for a solution. In this project, we address this problem by integrating the coding and error resolving environment. A website has been developed that examines code and provides simpler error messages that give a more comprehensive understanding of the bug. Once an error has been added to the database, the program can display the error more understandably. Experiments show that given several sample programs, our tool can extract the errors and report a more easily understandable solution.

KEYWORDS

Programming Environment, Python, Server, Database

For More Details :
https://aircconline.com/csit/papers/vol10/csit101216.pdf

Volume Link :
http://airccse.org/csit/V10N12.html

Ruichu (Eric) Xia¹ , Yu Sun² , Fangyan Zhang^3
1Santa Margarita Catholic High School, Rancho Santa Margarita, CA 92688, USA ²California State Polytechnic University, Pomona, CA, 91768, USA ³ASML, San Jose, CA, 95131, USA

ABSTRACT

Many people today suffer from the negative effects of procrastination and poor time management which includes lower productivity, missing opportunities, lower self-esteem and increased levels of guilt, stress, frustration, and anxiety. Although people can often recognize their tendency to procrastinate and the need to change this bad habit, the majority of them still do not take meaningful actions to prevent themselves from procrastinating. To help people fix this problem, we created a goal tracking mobile application called iProgress that aims to assist and motivate people to better manage their time by allowing them to create short-term and long-term goals that they want to achieve, and encouraging them to complete those goals through a rank/reward system that provides them with the opportunity to compete with other users by completing more goals.

KEYWORDS

Procrastination, iProgreass, flutter, iOS, Android

For More Details :
https://aircconline.com/csit/papers/vol10/csit101217.pdf

Volume Link :
http://airccse.org/csit/V10N12.html

Bill Zheng
Webb School of California, USA

ABSTRACT

In the current political climate, mass media was depicted as highly divisive and inaccurate while many cannot efficiently identify its bias presented in the news. Using research regarding keywords in the current political environment, we have designed an algorithm that detects and quantifies political, opinion, and satirical biases present in current day articles. Our algorithm makes use of scipy’ssklearn linear regression model and multiple regression model to automatically identify the bias of a news article based on a scale of 0 to 3 (-3 to 3 in political bias detection) to automatically detect the bias presented in a news source. The usage of this algorithm on all three segments, politics, opinion, and satire has been proven effective, and it enables an average reader to accurately evaluate the bias in a news source.

For More Details :
https://aircconline.com/csit/papers/vol10/csit101218.pdf

Volume Link :
http://airccse.org/csit/V10N12.html

Wesley Fan¹ , Eric Wasserman¹ , Eiffel Vuong¹ , Dylan Lazar¹ , Matthew Haase¹ , and Yu Sun^2
1Portola High School, 1001 Cadence, Irvine, CA 92618 ²California State Polytechnic University, Pomona, CA, 91768

ABSTRACT

In the recent decades, an increasing number of people become overweight, ranging from children to elders. Consequently, a series of diseases come along with obesity. How to control weight effectively is a big concern for most people. In order to improve the awareness of people’s diets and calorie intake, this paper develops an application–Fitable, which can help users by calculating calories burned in a particular workout. The foods that Fitable recommends are all based on the lifestyle the user is aiming to achieve. Until now, the app is accessible to Android users.

KEYWORDS

Android, flutter, firebase, machine learning

For More Details :
https://aircconline.com/csit/papers/vol10/csit101219.pdf

Volume Link :
http://airccse.org/csit/V10N12.html

Aman Gupta and Yadul Raghav
Department of Computer Science and Engineering Indian Institute of Technology (BHU) Varanasi, India 221–005

ABSTRACT

Meta-Learning, the ability of learning to learn, helps to train a model to learn very quickly on a variety of learning tasks; adapting to any new environment with a minimal number of examples allows us to speed up the performance and training of the model. It solves the traditional machine learning paradigm problem, where it needed a vast dataset to learn any task to train the model from scratch. Much work has already been done on meta-learning in various learning environments, including reinforcement learning, regression task, classification task with image, and other datasets, but it is yet to be explored with the time-series domain. In this work, we aimed to understand the effectiveness of meta-learning algorithms in time series classification task with multivariate time-series datasets. We present the algorithm’s performance on the time series archive, where the result shows that using meta-learning algorithms leads to faster convergence with fewer iteration over the non-meta-learning equivalent.

KEYWORDS

Time Series, Classification, Meta Learning, Few Shot Learning, Convolutional Neural Network

For More Details :
https://aircconline.com/csit/papers/vol10/csit101415.pdf

Volume Link :
http://airccse.org/csit/V10N14.html

Aman Gupta and Yadul Raghav
Department of Computer Science and Engineering Indian Institute of Technology (BHU) Varanasi, India 221–005

ABSTRACT

The problem of predicting links has gained much attention in recent years due to its vast application in various domains such as sociology, network analysis, information science, etc. Many methods have been proposed for link prediction such as RA, AA, CCLP, etc. These methods required hand-crafted structural features to calculate the similarity scores between a pair of nodes in a network. Some methods use local structural information while others use global information of a graph. These methods do not tell which properties are better than others. With an in-depth analysis of these methods, we understand that one way to overcome this problem is to consider network structure and node attribute information to capture the discriminative features for link prediction tasks. We proposed a deep learning Autoencoder based Link Prediction (ALP) architecture for the latent representation of a graph, unified with non-negative matrix factorization to automatically determine the underlying roles in a network, after that assigning a mixed-membership of these roles to each node in the network. The idea is to transfer these roles as a feature vector for the link prediction task in the network. Further, cosine similarity is applied after getting the required features to compute the pairwise similarity score between the nodes. We present the performance of the algorithm on the real-world datasets, where it gives the competitive result compared to other algorithms.

KEYWORDS

Keywords: Link Prediction, Deep Learning, Autoencoder, Latent Representation, Non-Negative Matrix Factorization

For More Details :
https://aircconline.com/csit/papers/vol10/csit101416.pdf

Volume Link :
http://airccse.org/csit/V10N14.html

Sergio Hidalgo-Espinoza, Kevin Chamorro-Cupuerán and Oscar Chang-Tortolero
School of Mathematics and Computer Science, University of Yachay Tech, Ecuador

ABSTRACT

Intrusion detection into computer networks has become one of the most important issues in cybersecurity. Attackers keep on researching and coding to discover new vulnerabilities to penetrate information security system. In consequence computer systems must be daily upgraded using up-to-date techniques to keep hackers at bay. This paper focuses on the design and implementation of an intrusion detection system based on Deep Learning architectures. As a first step, a shallow network is trained with labelled log-in [into a computer network] data taken from the Dataset CICIDS2017. The internal behaviour of this network is carefully tracked and tuned by using plotting and exploring codes until it reaches a functional peak in intrusion prediction accuracy. As a second step, an autoencoder, trained with big unlabelled data, is used as a middle processor which feeds compressed information and abstract representation to the original shallow network. It is proven that the resultant deep architecture has a better performance than any version of the shallow network alone. The resultant functional code scripts, written in MATLAB, represent a re-trainable system which has been proved using real data, producing good precision and fast response.

KEYWORDS

Keywords: Artificial Neural Networks, Information Security, Deep Learning, intrusion detection & hacking attacks

For More Details :
https://aircconline.com/csit/papers/vol10/csit101501.pdf

Volume Link :
http://airccse.org/csit/V10N15.html

Chie Ikeda, Karim Ouazzane and Qicheng Yu
School of Computing and Digital Media, London Metropolitan University, London, UK

ABSTRACT

Financial fraud activities have soared despite the advancement of fraud detection models empowered by machine learning (ML). To address this issue, we propose a new framework of feature engineering for ML models. The framework consists of feature creation that combines feature aggregation and feature transformation, and feature selection that accommodates a variety of ML algorithms. To illustrate the effectiveness of the framework, we conduct an experiment using an actual financial transaction dataset and show that the framework significantly improves the performance of ML fraud detection models. Specifically, all the ML models complemented by a feature set generated from our framework surpass the same models without such a feature set by nearly 40% on the F1-measure and 20% on the Area Under the Curve (AUC) value

KEYWORDS

Financial Fraud Detection, Feature Engineering, Feature Creation, Feature Selection, Machine Learning

For More Details :
https://aircconline.com/csit/papers/vol10/csit101517.pdf

Volume Link :
http://airccse.org/csit/V10N15.html

Byoung Jik Lee
School of Computer Sciences, Western Illinois University Macomb, IL, U.S.A.

ABSTRACT

We propose a valid approach to find the degree of important attributes in unlabeled dataset to improve the clustering performance. The significant degrees of attributes are extracted through the training of unsupervised simple competitive learning with the raw unlabeled data. These significant degrees are applied to the original dataset and generate the weighted dataset reflected by the degrees of influentialvalues for the set ofattributes. This work is simulated on the UCI Machine Learning repository dataset. The Scikit-learn K-Means clustering with raw data, scaled data, and the weighted data are tested. The result shows that the proposed approach improves the performance

KEYWORDS

Unsupervised MachineLearning, Simple Competitive Learning, SignificantDegree of Attributes, Scikit-learn K-Means Clustering, Weighted Data, UCI Machine Learning Data

For More Details :
https://aircconline.com/csit/papers/vol10/csit101608.pdf

Volume Link :
http://airccse.org/csit/V10N16.html

Fatimah Alshamari¹, ² and Abdou Youssef^1
1Department of Computer Science, The George Washington University, Washington D.C, USA ²Department of Computer Science, Taibah University, Medina, KSA

ABSTRACT

Document classification is a fundamental task for many applications, including document annotation, document understanding, and knowledge discovery. This is especially true in STEM fields where the growth rate of scientific publications is exponential, and where the need for document processing and understanding is essential to technological advancement. Classifying a new publication into a specific domain based on the content of the document is an expensive process in terms of cost and time. Therefore, there is a high demand for a reliable document classification system. In this paper, we focus on classification of mathematics documents, which consist of English text and mathematics formulas and symbols. The paper addresses two key questions. The first question is whether math-document classification performance is impacted by math expressions and symbols, either alone or in conjunction with the text contents of documents. Our investigations show that Text-Only embedding produces better classification results. The second question we address is the optimization of a deep learning (DL) model, the LSTM combined with one dimension CNN, for math document classification. We examine the model with several input representations, key design parameters and decision choices, and choices of the best input representation for math documents classification.

KEYWORDS

Math, document, classification, deep learning, LSTM

For More Details :
https://aircconline.com/csit/papers/vol10/csit101702.pdf

Volume Link :
http://airccse.org/csit/V10N17.html

John Hawkins
Transitional AI Research Group, Sydney, Australia

ABSTRACT

Prioritization of machine learning projects requires estimates of both the potential ROI of the business case and the technical difficulty of building a model with the required characteristics. In this work we present a technique for estimating the minimum required performance characteristics of a predictive model given a set of information about how it will be used. This technique will result in robust, objective comparisons between potential projects. The resulting estimates will allow data scientists and managers to evaluate whether a proposed machine learning project is likely to succeed before any modelling needs to be done. The technique has been implemented into the open source application MinViME (Minimum Viable Model Estimator) which can be installed via the PyPI python package management system, or downloaded directly from the GitHub repository. Available at https://github.com/john-hawkins/MinViME

KEYWORDS

Machine Learning, ROI Estimation, Machine Learning Metrics, Cost Sensitive Learning

For More Details :
https://aircconline.com/csit/papers/vol10/csit101803.pdf

Volume Link :
http://airccse.org/csit/V10N18.html

Yazhou Li¹ and Yahong Rosa Zheng^2
1School of Computer Science and Engineering, Beihang University, Beijing, China ²Department of Electrical and Computer Engineering, Lehigh University, Bethlehem, PA, 18015, USA

ABSTRACT

This paper presents two methods, tegrastats GUI version jtop and Nsight Systems, to profile NVIDIA Jetson embedded GPU devices on a model race car which is a great platform for prototyping and field testing autonomous driving algorithms. The two profilers analyze the power consumption, CPU/GPU utilization, and the run time of CUDA C threads of Jetson TX2 in five different working modes. The performance differences among the five modes are demonstrated using three example programs: vector add in C and CUDA C, a simple ROS (Robot Operating System) package of the wall follow algorithm in Python, and a complex ROS package of the particle filter algorithm for SLAM (Simultaneous Localization and Mapping). The results show that the tools are effective means for selecting operating mode of the embedded GPU devices.

KEYWORDS

Nvidia Jetson, embedded GPU, CUDA, Automous Driving. Robotic Operating Systems (ROS).

For More Details :
https://aircconline.com/csit/papers/vol10/csit101811.pdf

Volume Link :
http://airccse.org/csit/V10N18.html

Ali M. Alagrami¹ and Maged M. Eljazzar^2
1Department of Computer Science, University of Venice, Italy ²Faculty of Engineering, Cairo University, Egypt

ABSTRACT

Tajweed is a set of rules to read the Quran in a correct Pronunciation of the letters with all its Qualities, while Reciting the Quran. which means you have to give every letter in the Quran its due of characteristics and apply it to this particular letter in this specific situation while reading, which may differ in other times. These characteristics include melodic rules, like where to stop and for how long, when to merge two letters in pronunciation or when to stretch some, or even when to put more strength on some letters over other. Most of the papers focus mainly on the main recitation rules and the pronunciation but not (Ahkam AL Tajweed) which give different rhythm and different melody to the pronunciation with every different rule of (Tajweed). Which is also considered very important and essential in Reading the Quran as it can give different meanings to the words. In this paper we discuss in detail full system for automatic recognition of Quran Recitation Rules (Tajweed) by using support vector machine and threshold scoring system.

KEYWORDS

SVM, Machine learning , speech recognition , Quran Recitation, Tajweed

For More Details :
https://aircconline.com/csit/papers/vol10/csit101812.pdf

Volume Link :
http://airccse.org/csit/V10N18.html