Riya Jain et al. International Journal of Recent Research Aspects ISSN: 2349-7688, Vol. 9, Issue 1, March 2022, pp. 8-12 © 2022 IJRAA All Rights Reserved page- 8- Novel Techniques for Educational Data Mining Riya Jain 1 , Vivek Kumar 1 , Vinay Kumar Mishra 1 , Mayank Singh 1 , Neeraj Kumar Pandey 2 1 Computer Science and Engineering, Teerthanker Mahaveer University, India 2 Computer Science and Engineering, DIT University, India Abstract- Educational data mining (EDM) is a branch of study that focuses on the application of data mining, machine learning, and statistics of data, generated in educational contexts. This research area has been popular and some related terms like academic analytics, institutional analytics, and teaching analytics.Data mining is crucial in the subject of education, especially when assessing behaviour in an online learning environment. Cluster analysis and decision trees were applied as data mining techniques, in this study. The limits of the current study are discussed, as well as suggestions for further research. Keywords— educational data mining, decision tree, cluster analysis, academic analytics, teaching analytics I. INTRODUCTION Data mining is a commonly used way of extracting necessary or relevant information from large data sources. The main purpose of EDM is to evaluate the various types of education-based data to solve research issues in the field of education[1]. Extraction of hidden data patterns and detecting connections between parameters in a huge amount of data are the policies for applying for data mining. “Survey for data in education using data mining techniques is popularly known as educational data mining”[2]. Education data mine is a scientific field that extracts information from educational data. Predicting student performance is one of the things that is done, which helps teachers identify students who need extra help[3].Academic Analytics (AA) and Institutional Analytics (IA) are associated with the gathering, collection, and observation of curriculum activities such as courses and degree programmes, as well as research, student financial income, course testing, resource allocation, and management, in order to build institutional understanding. Teaching Analytics (TA) is the study of teaching activities and performance data, as well as the planning, implementation, and assessment of educational activities. This is focused on the educational challenges as seen through the perspective of educators [4].Learning analytics (LA) is the analysis and representation of student data in order to improve education is known as learning analytics (LA) [5]. LA implementation is often associated with web-based platforms, which provide direct access to student information with minimal effort or optimization [24]. Educational Data Mining (EDM) is concerned with improved methods for evaluating various forms of educational data from different institutions. It can also be explained in terms of the application of data mining techniques (DM) in this specific type of educational data to answer important educational problems[4]. This technology is so beneficial that it can clarify patterns derived from data analysis to understand hidden information and to facilitate decision-making [6]. The cloud computing is used for resource allocation. Cluster analysis is a systematic way of building a collection of patterns by groups based on their similarities of particular property or action because a collection analysis is used for various purposes in educational data mining, one of the most interesting areas in its use separating students to identify common patterns of behaviour [2]. The purpose of decision trees is direct identification object classes. Decision trees (DT) use a variety of attributes to distinguish different low-level items and do not use one attribute or set of prescribed symbols. The appeal of decision- trees is on its way to understanding and interpretation [2, 7]. DT is a branch structure made up of rules. Recursive apportioning is the process of creating a branch structure in a DT [8]. The different techniques used in EDM are: 1. Neural Network: Traditionally, the word "neural network" has been used to describe a network or circuit of biological neurons. The phrase is frequently used in modern usage. Artificial neural networks are made up of artificial neurons. Nodes, or neurons, are the building blocks of the brain. Other types of signaling exist in addition to electrical signaling. Signaling is caused by the diffusion of brain transmitters, which has an impact on electrical signaling. As a result, neural networks are quite useful. The applications are: radial basis function, networks, neural classification, bayesian confidence propagation neural networks. 2. Algorithm Architecture: To calculate a function, algorithm design is described as a finite list of well-defined instructions. Calculation, data processing, and automated reasoning are all done via algorithms, simply describes an algorithm as a computation technique that follows a set of steps. The various application are: gap statistic algorithms, chi-square automated interaction detection, models and algorithms, GRASP, OLAP,