com JashanUppalUG StudentDepartment of CSESRM University,[email protected] ABSTRACTIn this era of computerization, educationhas also revitalized itself and is no more limited to the old methods.
Thequest to find new and advanced ways to make educational system more efficientand to make students intellect have begun. These days, a lot of data iscollected in educational databases, but it remains unutilized in the databases.In order to make legitimate use of such a large amount of data, powerful toolsand algorithms are required. It is very important to study and analyze educationaldata to help & improvise the students. Educational Data Mining (EDM) is anemerging field, exploring data in educational context by applying different DataMining (DM) techniques/tools. It provides intrinsic knowledge of teaching andlearning process for effective educational planning. This paper presents acomprehensive survey, a travelogue towards educational data mining & its scopein future. I.
INTRODUCTIONInthe span of last 10-20 years, the number of education institutions have procreatedrapidly. Large number of graduates are produced by them every year. Institutesmay follow best of the inculcation methods; but still they face the problem ofdropout students, low achievers and unemployed students. EducationalData Mining (EDM) is an emerging field exploring data in educational context byapplying different Data Mining (DM) techniques.
EDM inherits properties fromareas like Learning Analytics, Artificial Intelligence, Information Technology,Machine learning, Statics, Database Management System, Computing and DataMining. It can be considered as interdisciplinary research field which providesintrinsic knowledge of teaching and learning process for effective education. EducationalData Mining is a new trend in the data mining and Knowledge Discovery inDatabases (KDD) field which focuses in mining useful patterns and discoveringuseful knowledge from the educational information systems, such as, admissionssystems, registration systems, course management systems and any other systemsdealing with students from schools, to colleges and universities. Researchersin this field focus on discovering useful knowledge either to help theeducational institutes manage their students in a better fashion, or to helpstudents to improvise their education and enhance their performance. Understandingand analyzing the factors for poor performance is a complex and ceaseless processbased on the past and present information besieged from academic performanceand students’ behavior. Powerful techniques and algorithms are required toanalyze and predict the performance of students scientifically.
Although,institutions collect a humongous number of students’ data, but this dataremains unutilized and does not help in any way to improve the performance ofstudents. If, Institutionscould identify the factors for low performance earlier and is able to predictstudents’ behavior, this knowledge can help them in taking pro-active actions,so as to improve the performance of such students. It will be a win-winsituation for all of them involved i.e. management, teachers, students andparents. Students will be able to identify their weaknesses beforehand and canimprove themselves.
Teachers will be able to plan their lectures as per theneed of students and can provide better guidance to such students. Parents willbe reassured of their ward performance in such institutes. Eventually, thiswill help in the proper growth of the nation. II.RELATED WORKBaradwajand Pal conducted a research on a group of 50 students enrolled in a specificcourse program across a period of 4 years, with multiple performanceindicators, which includes q PreviousSemester Marksq ClassTest Gradesq SeminarPerformanceq Assignmentsq GeneralProficiencyq Attendanceq LabWorkq EndSemester MarksTheyused ID3 decision tree algorithm to construct a decision tree, if-then rules. Thisapplication is supposed to help the instructors as well as the students tobetter understand and predict students’ performance at the end of the semester.
They defined their objective of this study as: “This study will also work toidentify those students which needed special attention to reduce fail rationand taking appropriate action for the next semester examination”. Abeerand Elaraby conducted a research that mainly focuses on generatingclassification rules and predicting students’ performance in a selected courseprogram based on previously recorded students’ behavior and activities. Theyprocessed and analyzed previously enrolled students’ data in a specific courseprogram across 6 years, with multiple attributes collected from the university.As a result, they were able to predict, the students’ final grades in theselected course program. They defined their objective of study as: “Help thestudents to improve the student’s performance, to identify those students whichneeded special attention to reduce failing ration and taking appropriate actionat right time”.
Bhardwajand Pal conducted a significant data mining research using the Naïve Bayesclassification method, on a group of BCA students. A questionnaire was conductedwith the help of each and every student before the final examination, which hadmultiple personal, social questions which was used in the study to identifyrelation between these factors and the student’s performance and grades. They statedtheir main objectives of this study as: q Generationof a data source of predictive variablesq Identificationof different factors, which effects a student’s learning behavior andperformance during academic careerq Constructionof a prediction model using classification data mining techniques on the basisof identified predictive variablesq Validationof the developed model for higher education students studying in IndianUniversities or Institutions. Theyfound that the most influencing factor for student’s performance is his gradein senior secondary school, i.e.
those students who performed well in theirsecondary school, will definitely perform well in their bachelors. It was alsofound that the living location, medium of teaching, mother’s qualification,student other habits, family annual income, and student family status, highlycontribute in the students’ educational performance. Bakerand Yacef describes the following to be the four goals of EDM:q Predictingstudent’s future learning behaviorq Discoveringor improving domain modelsq Studyingthe effects of educational supportq Advancingscientific knowledge about learning and learners Predictingstudent’s future learning behavior – Thisgoal can be achieved by creating student models that incorporate the learner’scharacteristics, including detailed information such as their knowledge,behaviors and motivation to learn. Discovering orimproving domain models – Through thevarious methods and applications of EDM, discovery of new and improvements toexisting models is possible. Studying theeffects of educational support – It canbe achieved through learning systems. Advancingscientific knowledge about learning and learners- By building and incorporating student models, the field of EDM research andthe technology can be improvised to a lot extent. III.
DATAMINING DEFINITION AND TECHNIQUESData mining refers to extracting or”mining” knowledge from large amounts of data. Data mining techniques areused to operate on large amount of data to find new and hidden patterns, relationshipswhich can be helpful in decision making.The various techniques used in Data Miningare: qAssociation analysisAssociation analysis is the discovery ofassociation rules showing attribute-value conditions that occur frequentlytogether in a given set of data.
Association analysis is widely used fortransaction data analysis. q PredictionIn prediction, the goal is to develop amodel which can infer a single aspect of data from some combination of otheraspects of data. If we study prediction extensively then we get three types ofprediction: classification, regression and density estimation. In any categoryof prediction, the input variables will be either categorical or continuous. Classification is the processing offinding a set of models (or functions) which describe and distinguish dataclasses or concepts, for the purposes of being able to use the model to predictthe class of objects whose class label is unknown.
q Clustering AnalysisUnlike classification and predication,which analyze class labeled data objects, clustering analyzes data objectswithout consulting a known class label. In general, the class labels are notpresent in the training data simply because they are not known to begin with.Clustering can be used to generate such labels. The objects are clustered orgrouped based on the principle of maximizing the intraclass similarity andminimizing the interclass similarity. That is, clusters of objects are formed sothat objects within a cluster have high similarity in comparison to oneanother, but are very dissimilar to objects in other clusters.
Each clusterthat is formed can be viewed as a class of objects, from which rules can bederived. VI.ALGORITHM USEDNaive Bayes: classifiers are a collection of classification algorithms based on Bayes’ Theorem.
It is not a single algorithm but a family of algorithmswhere all of them share a common principle – “every pair of features beingclassified is independent of each other”.Naïvemodel is the default model that predicts the classes of all examples in adataset as the class of its mode (highest frequency). For example, let’sconsider a dataset of 100 records and 2 classes (Yes & No), the “Yes”occurs 70 times and “No” occurs 30 times, the default model for this datasetwill classify all objects as “Yes”, hence, its accuracy will be 70%. Eventhough it is useless, but equally important, it allows to evaluate theaccuracies produced by other classification models.
This concept can begeneralized to all classes/labels in the data to produce an expectation of theclass recall as well. V. CONCLUSION& FUTURE WORKData mining is a tremendously vast areathat includes employing different techniques and algorithms for patternfinding. The algorithms discussed in this paper are the ones used in educationmining. These algorithms have shown a remarkable improvement in strategies likecourse outline formation, teacher student understanding and high output andturn out ratio. ICDM conference encourages employment and development ofalgorithms helpful in data mining. An appreciable research is still being doneon various algorithms.Prediction with data mining has reapedbenefits; such as finding set of weak students, determining student’ssatisfaction for a particular course, Faculty Evaluation, Comprehensive studentevaluation, Class room teaching language selection, predicting students’dropout, course registration planning, predicting the enrollment headcount,evaluation of collaborative activities etc.
Oneof the most recent and biggest challenge that higher education faces today ismaking students skillfully employable. Many universities/institutes are not inposition to guide their students because of lack of information and assistancefrom their teaching-learning systems. To better administer and serve studentpopulation, the universities/institutions need better assessment, analysis, andprediction tools. REFERENCESq Nat’l Research Council,Building a Workforce for the Information Economy, Nat’l Academies Press, 2001.q C. Romero, S. Ventura, and E. Garca,”DataMining in Course Management Systems: Moodle Case Study and Tutorial,” Computers& Education, vol.
51, no. 1, 2008, pp. 368–384.q .
L. Pappano, “The Yearof the MOOC,”The New York Times, 2 Nov. 2012;www.nytimes.com/2012/11/04/education/edlife/massive-open-online-courses-are-multiplying-at-a-rapid-pace.html?_r=0.q Z.
Pardos et al., “Adapting Bayesian KnowledgeTracing to a Massive Open Online Course in edX,” Proc. 6th Int’l Conf.Educational Data Mining (EDM 13), 2013; www.educationaldatamining.org/EDM2013/papers /rn_paper_21.pdf.
q A. Elbadrawy, R.S. Studham, and G. Karypis,”Collaborative Multiregression Models for Predicting Students’ Performance inCourse Activities,” Proc.
5th Int’l Conf. Learning Analytics and Knowledge (LAK15), 2015, pp. 103–107.