Search

Background Mobile learning has become an essential instruction platform in many schools, colleges, universities, and various other educational institutions across the globe, as a result of the COVID-19 pandemic crisis. The resulting severe, pandemic-related circumstances have disrupted physical and face-to-face contact teaching practices, thereby requiring many students to actively use mobile technologies for learning. Mobile learning technologies offer viable web-based teaching and learning platforms that are accessible to teachers and learners worldwide. Objective This study investigated the use of mobile learning platforms for instruction purposes in United Arab Emirates higher education institutions. Methods An extended technology acceptance model and theory of planned behavior model were proposed to analyze university students’ adoption of mobile learning platforms for accessing course materials, searching the web for information related to their disciplines, sharing knowledge, and submitting assignments during the COVID-19 pandemic. We collected a total of 1880 questionnaires from different universities in the United Arab Emirates. Partial least squares-structural equation modeling and machine learning algorithms were used to assess the research model, which was based on the data gathered from a student survey. Results Based on our results, each hypothesized relationship within the research model was supported by our data analysis results. It should also be noted that the J48 classifier (89.37% accuracy) typically performed better than the other classifiers when it came to the prediction of the dependent variable. Conclusions Our study revealed that teaching and learning could considerably benefit from adopting remote learning systems as educational tools during the COVID-19 pandemic. However, the value of such systems could be lessened because of the emotions that students experience, including a fear of poor grades, stress resulting from family circumstances, and sadness resulting from a loss of friends. Accordingly, these issues can only be resolved by evaluating the emotions of students during the pandemic.

The Fuhrman nuclear grade is a recognized prognostic factor for patients with clear cell renal cell carcinoma (CCRCC) and its pre-treatment evaluation significantly affects decision-making in terms of management. In this study, we aimed to assess the feasibility of a combined approach of radiomics and machine learning based on MR images for a non-invasive prediction of Fuhrman grade, specifically differentiation of high- from low-grade tumor and grade assessment. Images acquired on a 3-Tesla scanner (T2-weighted and post-contrast) from 32 patients (20 with low-grade and 12 with high-grade tumor) were annotated to generate volumes of interest enclosing CCRCC lesions. After image resampling, normalization, and filtering, 2438 features were extracted. A two-step feature reduction process was used to between 1 and 7 features depending on the algorithm employed. A J48 decision tree alone and in combination with ensemble learning methods were used. In the differentiation between high- and low-grade tumors, all the ensemble methods achieved an accuracy greater than 90%. On the other end, the best results in terms of accuracy (84.4%) in the assessment of tumor grade were achieved by the random forest. These evidences support the hypothesis that a combined radiomic and machine learning approach based on MR images could represent a feasible tool for the prediction of Fuhrman grade in patients affected by CCRCC.

Filariasis is one of the major public health concerns in India. Approximately 600 million people spread across 250 districts of India are at risk of filariasis. To predict this disease, a pilot scale study was carried out in 30 villages of Karimnagar district of Telangana from 2004 to 2007 to collect epidemiological and socio-economic data. The collected data are analysed by employing various machine learning techniques such as Naïve Bayes (NB), logistic model tree, probabilistic neural network, J48 (C4.5), classification and regression tree, JRip and gradient boosting machine. The performances of these algorithms are reported using sensitivity, specificity, accuracy and area under ROC curve (AUC). Among all employed classification methods, NB yielded the best AUC of 64% and was equally statistically significant with the rest of the classifiers. Similarly, the J48 algorithm generated 23 decision rules that help in developing an early warning system to implement better prevention and control efforts in the management of filariasis.

Under the impact of covid-19, the global and domestic manufacturing supply chains, almost suffered from the serious interruption crisis of manpower flow, logistics, information flow and capital flow. The risk of supply chain disruption has become the primary risk of the supply chain. However, some risk inducement of supply chain interruption is complex and diverse, so it is very difficult to grasp and screen the risk data needed for research from the supply chain operation data. To improve the robustness of supply chain for boosting the domestic and international circulation of China's manufacturing, in this paper, according to the characteristics of China's manufacturing supply chain and its risk incentives, the data needed for risk prediction modeling has been sorted out through questionnaire survey, and a regression model of risk prediction for manufacturing supply chain by using empirical method would be put forward. Then, C4.5 decision tree method is used to train and evaluation the risk prediction model. The conclusion shows that the customer satisfaction has great diagnostic value for risk, and the model has a strong sensitivity to market information risk and market order risk. The conclusion is more consistent with general cognition, and the model fits well, indicating that the model proposed in this paper has a certain theoretical significance, and its practical application value is worthy of further testing. © 2021, North Atlantic University Union NAUN. All rights reserved.

Data mining is an interdisciplinary field that uses methods such as machine learning, artificial intelligence, statistics, and deep learning. Classification is an important data mining technique as it is widely used by researchers. Generally, statistical methods or machine learning algorithms such as Decision Trees, Fuzzy Logic, Genetic Programming, Random Forest, Artificial Neural Networks and Logistic Regression have been used in software defect prediction in the literature. Performance measures such as Accuracy, Precision, Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) are used to examine the performance of these classifiers. In this paper, 4 data sets entitled JM1, KC1, CM1, PC1 in the PROMISE repository, which are created within the scope of the publicly available NASA institution's Metric Data Program, are examined as in the other software defect prediction studies in the literature. These datasets include Halstead, McCabe method-level, and some other class-level metrics. Data sets are used with Wakiato Environment for Knowledge Analysis (WEKA) data mining software tool. By this tool, some classification algorithms such as Naive Bayes, SMO, K *, AdaBoost1, J48 and Random Forest were applied on NASA error datasets in PROMISE repository and their accuracy rates were compared. The best value among the accuracy rates was obtained in the Bagging algorithm in the PC1 data set with the values of %94.13. Keywords: Software Defect Prediction, McCabe, Halstead, Data Mining, Accuracy, Random Forest Cite this paper as: GUVEN AYDIN, Z.B., SAMLI, R. (2020). A Comparison of Software Defect Prediction Metrics Using Data Mining Algorithms. Journal of Innovative Science and Engineering. 4(1): 11-21 *Corresponding author: Zeynep Behrin GUVEN AYDIN E-mail: zeynepguven@maltepe.edu.tr Received Date: 24/02/2020 Accepted Date: 05/05/2020 © Copyright 2020 by Bursa Technical University. Available online at http://jise.btu.edu.tr/ The works published in Journal of Innovative Science and Engineering (JISE) are licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

University dropout is a growing problem which, in recent years, is using computer techniques to assist in the detection process. The paper presents the evaluation of some prediction algorithms to detect a student with a high possibility of scholar desertion. The approach uses real data from past scholar periods to create a dataset with different information of the students (i.e., personal, economic, and academic records). The algorithms selected in the experimental phase were: J48 decision tree, K-near neighbors, and support vector machine. We use two similarity metrics to split the dataset with cases with at least 80% of similarity to evaluate each case. We use the data from 2010 to 2016 with real students’ information to predict if there exists the possibility of a real academic dropout in one test for a period. The results show that the J48 algorithm reaches a better performance in both experiments. Besides, the tree generated for each student is taken as a path of attention, reaching around 88% of effectiveness. Finally, the conclusions argue the contributions of the paper and propose a future line of research.

In this paper, we first consider the phenomenon of exponential growth of existing equipment data in substation, introduce the cloud computing operation mode, and build a more scalable MapReduce processing system based on Hadoop platform. Cloud computing is completed by using decision tree C4.5 algorithm, and the algorithm is used to optimize the parallel computing mode on the Hadoop platform as the core of the same data processing. Based on this point, this paper considers the impact of the ecological restoration effect of L County in the northern mountainous area of T city on the environment of the study area, analyzes the climate and hydrological conditions of the study area, provides the basic data of the mountain ecological environment of Y mountain area for the study of mountain landscape planning, and facilitates the practice. The combination of landscape planning, ecological restoration, and environmental reconstruction system theory provides a theoretical and practical basis for ecological planning in the central mountainous area of T city. Finally, based on MOOC components and software architecture, this paper constructs a college English remote system. After comprehensive testing in many aspects, this paper reduces the cost of software development in this field, which not only improves the reliability of the system, but also shortens the development cycle, speeds up the maintenance speed of the system, and increases the scale of the system. About the system testing, this paper tests the function, performance, feasibility, integrity, and security of the system through the network testing method and cookie method. Based on the basic C4.5 decision-making algorithm and college English remote system, this paper studies the impact of mountain ecological environment restoration, provides ideas for promoting the development of the system, and provides reference for the reform and practice methods and processes in other courses.

Background Heart disease (HD) is one of the most common diseases nowadays, and an early diagnosis of such a disease is a crucial task for many health care providers to prevent their patients for such a disease and to save lives. In this paper, a comparative analysis of different classifiers was performed for the classification of the Heart Disease dataset in order to correctly classify and or predict HD cases with minimal attributes. The set contains 76 attributes including the class attribute, for 1025 patients collected from Cleveland, Hungary, Switzerland, and Long Beach, but in this paper, only a subset of 14 attributes are used, and each attribute has a given set value. The algorithms used K- Nearest Neighbor (K-NN), Naive Bayes, Decision tree J48, JRip, SVM, Adaboost, Stochastic Gradient Decent (SGD) and Decision Table (DT) classifiers to show the performance of the selected classifications algorithms to best classify, and or predict, the HD cases. Results It was shown that using different classification algorithms for the classification of the HD dataset gives very promising results in term of the classification accuracy for the K-NN (K = 1), Decision tree J48 and JRip classifiers with accuracy of classification of 99.7073, 98.0488 and 97.2683% respectively. A feature extraction method was performed using Classifier Subset Evaluator on the HD dataset, and results show enhanced performance in term of the classification accuracy for K-NN (N = 1) and Decision Table classifiers to 100 and 93.8537% respectively after using the selected features by only applying a combination of up to 4 attributes instead of 13 attributes for the predication of the HD cases. Conclusion Different classifiers were used and compared to classify the HD dataset, and we concluded the benefit of having a reliable feature selection method for HD disease prediction with using minimal number of attributes instead of having to consider all available ones.

This paper illustrates how to use data mining techniques to help in advising students and predicting their academic performance. Data mining is used to get previously unknown, hidden and perhaps vital knowledge from a large amount of data. It combines domain knowledge, advanced analytical skills, and a vast knowledge base to reveal hidden patterns and trends that are applicable in virtually any sector ranging from engineering to medicine, to business. However, it is possible for educational institutes to use data mining to find useful information from their databases. This is usually called Educational Data Mining (EDM). Advancing the field of EDM with new data analysis techniques and new machine learning algorithms is vital. Classification and clustering techniques will be used in this project to study and analyse student performance. The key importance of this project is that it discusses different data mining techniques in the literature review to study student behaviour depending upon their performance. We tried to identify the most suitable algorithms from the existing research methods to predict the success of students. Various data mining approaches were discussed and their results were evaluated. In this paper, the J48 algorithm was applied to the data set, gathered from Umm Al-Qura University in Makkah.

Several higher educational institutions are adapting the strategy of predicting the student’s performance throughout the academic years. Such a practice ensures not only better academic outcomes but also helps the institutions to reorient their curriculums and teaching pedagogies so as to add to the students’ learning curve. Educational Data Mining (EDM) has risen as a useful technology in this league. EDM techniques are now being used for predicting the enrolment of students in a specific course, detection of any irregular grades, prediction about students’ performance, analyzing and visualizing of data, and providing feedback for overall improvement in the academic spheres. This paper reviews the studies related to EDM, including the approaches, data sets, tools, and techniques that have been used in those studies, and points out the most efficient techniques. This review paper uses true prediction accuracy as a standard for the comparison of different techniques for each EDM applications of the surveyed literature. The results show that the J48 and K-means are the most effective techniques for predicting the students’ performance. Furthermore, the results also cite that Bayesian and Decision Tree Classifiers are the most widely used techniques for predicting the students’ performance. In addition, this paper highlights that the most widely used tool was WEKA, with approximately 75% frequency. The present study’s empirical assessments would be a significant contribution in the domain of EDM. The comparison of different tools and techniques presented in this study are corroborative and conclusive, thus the results will prove to be an effective reference point for the practitioners in this field. As a much needed technological asset in the present day educational context, the study also suggests that additional surveys are recommended to be driven for each of the EDM applications by taking into account more standards to set the best techniques more accurately.

Recognition of emotion-based utterances from speech has been produced in a number of languages and utilized in various applications. This paper makes use of the spoken utterances corpus recorded in Urdu with different emotions of normal and special children. In this paper, the performance of learning classifiers is evaluated with prosodic and spectral features. At the same time, their combinations considering children with autism spectrum disorder (ASD) as noise in terms of classification accuracy has also been discussed. The experimental results reveal that the prosodic features show significant classification accuracy in comparison with the spectral features for ASD children with different classifiers, whereas combinations of prosodic features show substantial accuracy for ASD children with J48 and rotation forest classifiers. Pitch and formant express considerable classification accuracy with MFCC and LPCC for special (ASD) children with different classifiers.

Objectives: The study presents best machine learning models that can be employed on terrorism related data to predict the most accurate terrorist group responsible for an attack based on historic data. Methods: This paper analyses terrorism challenges faced in India by modelling the behaviour of terrorist groups using famous machine learning algorithms like J48, IBK, Naive Bayes and ensemble approach using vote. Findings: The results of the evaluation models show the accuracy percentage of various models employed and their relevance to the dataset. It was found that when the classification models are created on a data that has class imbalance problem the percentage of correctly classified instances will be very less. The paper establishes that sampling plays a main role in determining the accuracy percentage of classifier models and gives better results. The study also shows the accuracy percentage of correctly classified instances for various algorithms and shows the ideal one for the dataset. Application/Improvement: The defined model is a new approach to classify data that has major class imbalance problem by using sampling techniques like oversampling, under sampling and ensemble models. Keywords: Classification, Data Mining, Ensemble, Global Terrorism Database

Objectives: The advent of information technology causes a significant impact on pedagogy and using ICT technologies in school-related activities. With this argument, determining the teacher’s ICT competence is expected to yield a better or positive impact on the learners and the performance of the teachers itself, however, no model was developed in predicting teachers ICT competence. In line with this, it is empirical to study and develop a model that will help the university in predicting teacher’s ICT competence. Methods: Data mining approach utilizing J48 algorithm was applied in this paper to create a model suitable for the actual teachers’ characteristics in the University. Moreover, Cross-validation technique was used to validate the dataset to have an optimum and acceptable model and generating the Receiving Operating Characteristics Curve (ROC) Area under ROC Curve technique. Findings: Decision tree model and decision rule for classification were created. Additionally, there were 92.78% correctly classified with an AUC weighted mean of 92.4%. Also, the model has very high acceptability and accuracy in predicting Teacher’s ICT Competence. However, it also revealed that many teachers still need more exposure in utilizing ICT technologies in pedagogy and any school-related activities. Application/Improvements: The result of this study can be a basis for developing software that will automatically categorize or classify the Teacher’s ICT competence. For more improvement of this paper and the model, it is suggested to add additional parameters to have more factors involved in predicting Teacher’s ICT competence. Keywords: Blended Learning, Decision Tree, ICT Competency, ICT Domain, ICT Pedagogy, J48 Algorithm, Machine Learning

Introduction: This paper is part of a research project that aims to construct a predictive model for students’ academic performance, as result of an iterative process of experimentation and evaluation of the pertinence of some data mining techniques. Methodology: This paper was written in 2016 in the Universidad El Bosque, Bogotá, Colombia, and presents a comparative analysis of the performance and relevance of the J48 and Random Forest algorithms, in order to identify the most influential demographic and icfes score variables, as well as the classification rules, to predict the first year academic performance of the Engineering Faculty students, in Universidad El Bosque, Bogotá, Colombia. Results: The analysis process was carried out on 7,644 students’ records, and it was developed in two phases. Firstly, the data needed to feed the mining process was extracted and prepared. Secondly, the data mining process itself was implemented through preprocessing data and executing the classification algorithms available in Weka. Some significant variables and rules to predict academic performance are found, according to the studied population characteristics. Conclusions: The academic risk seen as the cause of the desertion phenomenon must be studied as a phenomenon itself. Establishing its causes facilitates the creation of preventive strategies for the accompaniment of students through their process, aimed to mitigate the risk of both phenomena.

During recent years, there has been an increase in cyber-crime and cybercriminal activities around the world and as countermeasures, effective attack prevention and detection mechanisms are needed. A popular tool to augment existing attack detection mechanisms is the Honeypot. It serves as a decoy for luring attackers, with the purpose to accumulate essential details about the intruder and techniques used to compromise systems. In this endeavor, such tools need to effectively listen and keep track of ports on hosts such as servers and computers within networks. This paper investigates, analyzes and predicts destination port numbers targeted by attackers in order to improve the effectiveness of honeypots. To achieve the purpose of this paper, the J48 decision tree classifier was applied on a database containing information on cyber-attacks. Results revealed insightful information on key destination port numbers targeted by attackers, in addition to how these targeted ports vary within different regions around the world.

The main objective of any educational organization is to provide quality education and improve the overall performance of an institution by looking at individual performances. One way to analyze learners' performances is to identify the areas of weakness and guide their students to a better future. Although data mining has been successful in many areas, its use in student performance analysis is still relatively new, i.e. the knowledge is hidden in educational data set and it is extracted using data mining techniques. This paper discusses about a learning model for predicting student performance using classification techniques. Also the paper shows the comparative performance analysis of J48, Naive Bayesian classifier and Random forest algorithm.

Objectives: To training classifier based on the features extracted from the poems of Mukkoodar Pallu, authors for various unknown poems can be classified. Methods/Analysis: The classification accuracy by performing classification in the dataset using C4.5 algorithm is illustrated in this paper. Findings: The results of performing classification on dataset that consists of features extracted from the dataset are shown in this paper. Features like number of characters, number of sentences and the classification accuracy when C4.5 algorithm is used is illustrated. Novelty/Improvement: By doing this, authors of various other poems in Tamil language can be identified which will be helpful to the society. Also a generalized authorship identification tool for all regional languages can be achieved.

Accuracy in data classification depends on the dataset used for learning. Now-a-days the most important cause of death for both men and women is due to the Liver Problem. The healthcare industry collects a huge amount of data which is not properly mined and not put to the optimum use. Discovery of these hidden patterns and relationships often goes unexploited. Our research focuses on this aspect of Medical diagnosis by learning pattern through the collected data of Liver disorder to develop intelligent medical decision support systems to help the physicians. In this paper, we propose the use decision trees J48, Naive Bayes, ANN, ZeroR, 1BK and VFI algorithm to classify these diseases and compare the effectiveness, correction rate among them. Detection of Liver disease in its early stage is the key of its cure. It leads to better performance of the classification models in terms of their predictive or descriptive accuracy, diminishing of computing time needed to build models as they learn faster, and better understanding of the models. In this paper, a comparative analysis of data classification accuracy using Liver disorder data in different scenarios is presented. The predictive performances of popular classifiers are compared quantitatively.

Nowadays data mining become one of the technologies that paly major effect on business intelligence. However, to be able to use the data mining outcome the user should go through many processes such as classified data. Classification of data is processing data and organize them in specific categorize to be use in most effective and efficient use. In data mining one technique is not applicable to be applied to all the datasets. Many data users wasting a lot of time trying many classification techniques in order to find the most an appropriate technique to be used. This paper showing the difference result of applying different techniques on the same data. This paper evaluates the performance of different classification techniques using different datasets. In this study four data classification techniques have chosen. They are as follow, BayesNet, NaiveBayes, Multilayer perceptron and J48. The selected data classification techniques performance tested under two parameters, the time taken to build the model of the dataset and the percentage of accuracy to classify the dataset in the correct classification. The experiments are carried out using Weka 3.8 software. The results in the paper demonstrate that the efficiency of Multilayer Perceptron classifier in overall the best accuracy performance to classify the instances, and NaiveBayes classifiers were the worst outcome of accuracy to classifying the instance for each dataset.

Handwritten character recognition is one of the practically important issues in pattern recognition applications. The applications of digit recognition include in postal mail sorting, bank check processing, form data entry, etc. The main problem lies within the ability on developing an efficient algorithm that can recognize hand written digits, which is submitted by users by the way of a scanner, tablet, and other digital devices. This paper presents an approach to off-line handwritten digit recognition based on different machine learning techniques. The main objective of this paper is to ensure the effectiveness and reliability of the approached recognition of handwritten digits. Several machines learning algorithms (i.e. Multilayer Perceptron, Support Vector Machine, Naïve Bayes, Bayes Net, Random Forest, J48, and Random Tree) have been used for the recognition of digits using WEKA. The experimental results showed that the highest accuracy was obtained by Multilayer Perceptron with the value of 90.37%.

In this era of High-Performance High computing systems, Large-scale Data Mining methodologies in the field of education have become a convenience to discover and extract knowledge from Databased of their respective educational archives. Typically, all educational institutions around the world maintain student data repositories. Attributes of students such as the name of the student, gender of student, age group (date of birth), religion, eligibility details, academic assessment details, etc. are kept in it. With this knowledge, in this paper, didactical data mining (DDM) is used to leverage the performance prediction of student and to analyse it proactively. As it is known, Classification and Clustering are the liveliest techniques in mining the required data. Hence, Bound Model of Clustering and Classification (BMCC) have been proposed in this research for most proficient educational data mining. Classification is one of the distinguished options in Data Mining to assign an object under some pre-defined classes according to their attributes, and hence it comes under a supervised learning problem. On the other side, clustering is considered as a non-supervised learning problem that involves in grouping up of objects with respect to some similarities. Moreover, this paper uses the dataset collected from Kerala Technological University-SNG College of Engineering (KTU_SNG) for performing the BMCC. An efficient J48 decision tree algorithm is used for classification and the k-means algorithm is incorporated for clustering here and is optimised with Bootstrap Aggregation (Bagging). The implementation has been done and analysed with a data mining tool called WEKA (Waikato Environment for Knowledge Analysis), and the results are compared with some most used classifications such as Bayes Classifier (NB), Neural Network (Multilayer Perceptron MLP) and J48. It is provable from the results that the model, proposed in this provides high Precision Rate (PR), accuracy and robustness with less computational time, though the sample data set includes some missing values.

Imbalanced dataset is occurred due to uneven distribution of data available in the real world such as disposition of complaints on government offices in Bandung. Consequently, multi-label text categorization algorithms may not produce the best performance because classifiers tend to be weighed down by the majority of the data and ignore the minority. In this paper, Bagging and Adaptive Boosting algorithms are employed to handle the issue and improve the performance of text categorization. The result is evaluated with four evaluation metrics such as hamming loss, subset accuracy, example-based accuracy and micro-averaged f-measure. Bagging ML-LP with SMO weak classifier is the best performer in terms of subset accuracy and example-based accuracy. Bagging ML-BR with SMO weak classifier has the best micro-averaged f-measure among all. In other hand, AdaBoost MH with J48 weak classifier has the lowest hamming loss value. Thus, both algorithms have high potential in boosting the performance of text categorization, but only for certain weak classifiers. However, bagging has more potential than adaptive boosting in increasing the accuracy of minority labels., Comment: Accepted in the 2015 International Conference on Electrical Engineering and Informatics (ICEEI), Best Student Paper Award

Determination of initial ability of examinees is one of the important stages in the adaptive assessment system. The accuracy of the examinee's ability level prediction will influence the appropriateness of choosen item difficulty level for each examinee. This paper discusses the patterns mining of cognitive ability based on cluster using K-Means. The K-means method is utilized to mine the examinees’ ability pattern from examinees’ pretest answers. The patterns are used for developing rules to determine examinee’s ability level in the adaptive assessment system. The addition of this method is proposed to improve the performance of the prediction methods to predict the examinees’ ability level. Extraction of graduation value at each level is done before the pattern excavation process. Patterns found become the basis of making the rules as well as replace the rules from the experts in previous studies. The prediction of participants' ability is done by implementing rule based method classifier. A total of 140 data were used for the experiment. Based on the results of the experiment, it can be concluded that the cluster-based pattern mining using K-means can be utilized to determine the cognitive ability level of examinee. The application of this method to student pretest data shows the performance improvement of all the prediction methods used in this paper, i.e. Naive Bayes, MLP, SMO, Decision Table, JRIP, and J48. This method is suitable for adaptive assessment system where the rules can be adjusted along with the addition of the number of the data as well as the addition of the number of variations in the ability pattern of examinees.

Data mining today is being used widely in diverse areas. For example: fraudulent systems, recommender systems, disease prediction, and numerous other applications. One such application is exploited in this article. This paper presents an approach to detect gender of a person through frontal facial image, using techniques of data mining and Delaunay triangulation. Gender prediction can prove to be a very useful technique in HCI (Human Computer Interaction) Systems. Classification, being a very power technique in data mining to group categorical data, is used here to classify a gender as either male, or female. Various classification algorithms such as Functional Trees, AdaBoost, J48, and few others are used to gauge the maximum accuracy. The model used in this paper is robust and attains accuracy level of 93.8283% along with relative scale invariance. Details of the prediction model and results are reported herein. General Terms Data mining, classification algorithms, WEKA tool, gender detection, Machine learning

Dirty water is the world's biggest health risk. When water from rain roads into rivers, it picks up toxic chemicals, dirt, trash and disease-carrying organisms along the way. Many of our water resources lack basic protections, making them vulnerable to pollution from factory farms and industrial plants. Due to that, a classification model is needed to present the quality of the water environment. In this paper, the data mining techniques are used in this research by applying the classification method for water quality application. Various classifiers were studied in order to find the most accurate classifier for the dataset. This paper presents the comparison of accuracies for the five classifiers (NB, MLP, J48, SMO, and IBk) based on a 10-fold cross validation as a test method with respect to water quality from the datasets of Kinta River, Perak Malaysia. This study also explores which classifier is suitable to classify the dataset. The selected attributes used in this study were: DO Sat, DO Mgl, BOD Mgl, COD Mgl, TS Mgl, DO Index, AN Index, SS Index, Class, and Degree of pollution. The data consisted of 166 instances and obtained from the East Coast Environmental Research Institute (ESERI) of Universiti Sultan Zainal Abidin (UniSZA). The result of MLP and IBk performed better than other classifiers for Kinta River dataset because these classifiers showed the highest accuracy with the same percentage of 91.57%. In the future, we will propose the multiclassifier approach by introducing a fusion at a classification level between these classifiers to get a higher accuracy of classification.