Predictive Analytics in Education Context

Predictive Analytics in Education Context

• 1.
  27 International Journalfor Modern Trends in Science and Technology Volume: 2 | Issue: 10 | October 2016 | ISSN: 2455-3778IJMTST Predictive Analytics in Education Context Neha Kawchale1 | Prof. Rachana Satao2 1,2Department of Computer Engineering, Smt.Kashibai Navale College of Engineering, Pune, Maharashtra, India. Now-a-days data volumes are growing rapidly in several domains. Many factors have contributed to this growth, including inter alia proliferation of observational devices, miniaturization of various sensors ,improved logging and tracking of systems, and improvements in the quality and capacity of both disk storage and networks .Analyzing such data provides insights that can be used to guide decision making. To be effective, analysis must be timely and cope with data scales. The scale of the data and the rates at which they arrive make manual inspection infeasible. As an educational management tool, predictive analytics can help and improve the quality of education by letting decision makers address critical issues such as enrollment management and curriculum Development. This paper presents an analytical study of this approach’s prospects for education planning. The goals of predictive analytics are to produce relevant information, actionable insight, better outcomes, and smarter decisions, and to predict future events by analyzing the volume, veracity, velocity, variety, value of large amounts of data and interactive exploration. KEYWORDS: Descriptive analytics, Learning analytics, Action analytics and Predictive analytics. Copyright © 2016 International Journal for Modern Trends in Science and Technology All rights reserved. I. INTRODUCTION Analytics is the process of discovering [1],[7], analyzing, and interpreting meaningful patterns from large amount of data. The importance of predictive analytics is its ability to suggest the most favorable planning for the future [8] by combining data about who, what, where, and when (the four W‘s) to analyze why and how. Organizations from business to education have adopted analytics for several reasons, which includes increasing operational or financial efficiency, expanding global local impact, establishing new funding models during a changing economic climate, and respond to demand for greater accountability. Using predictive analytics has several key benefits, including better future business performance, more insight into business dynamics, and optimum use of available data. Organizations adopt predictive analytics for several key reasons—in particular [3], to compete, grow, enforce, improve, satisfy, learn, and act[9].As in business, analytics in higher education defines goal-directed practices for ensuring organizational success at all levels[10]. A. Research Challenges IBM reports that ―90% of the data available today was created just in the past two years.‖ In the era of big data, the challenges of predictive analytics include the quality of the data, because the prediction model‘s quality depends on it; the quantity of the data, because limited data provided during the training phase can make the analysis incapable of generalizing the derived knowledge when fed the new data; and the ability to satisfy analytical performance criteria—that is, results must be accurate and make statistical sense, and outcomes must be actionable—so that the analytics can identify the actual necessity for predicting an educational goal[4]. Analytics in education can address the following questions: •What will be the cutoff for the particular college with course-wise? • Which students will enroll in particular course programs? ABSTRACT
• 2.
  28 International Journalfor Modern Trends in Science and Technology Predictive Analytics in Education Context • Which students will need assistance to graduate? • What types of courses will attract more students as per industry needs? • What is current demand, which programs are trending, and which are becoming obsolete? • Which students are likely to drop out? • What is the level of student satisfaction in the current education system? • What will be the fees structure in the upcoming years? • What are recruiter requirements, and how can institutions and curriculum be designed to fulfill. The market‘s exact needs and also rapidly changing needs? B. Analytics Terminology and Goals Being an interdisciplinary area of research, analytics uses a variety of terminologies. All analytics can be broadly divided into the following categories with respect to the intent of the activity. Descriptive analytics (what happened) is the first step in the successful application of predictive or prescriptive analytics, and involves preparing and analyzing historical data. Predictive analytics (what will happen) predicts future trends and probabilities. Finally, prescriptive analytics (how you can make it happen) optimizes descriptive and predictive analytics. The following terminology is used in higher education analytics: 1. Academic analytics: ―Process for providing the data necessary to respond to the reportage and Decision-making challenges facing contemporary universities.‖ 2. Learning analytics (academia) [3]: ―Interpretation of a wide range of data produced by and gathered on behalf of students in order to assess academic progress, predict future performance, and spot potential issues.‖ 3. Action analytics: ―Fusion of new analytic tools with the increasing expectations for higher education accountability.‖ Although predictive analytics originated from artificial intelligence as a way to make predictions based on discovered and recognized patterns in a dataset, we can treat it as an interdisciplinary field that inherits properties from database management systems, data mining, educational data mining, machine learning, learning analytics, and so on. It‘s also known as ―one-click data mining‖ because it simplifies and automates the data mining process. The goals of predictive analytics are to produce relevant information, actionable insight, better outcomes, and smarter decisions, and to predict future events by analyzing the volume, veracity, velocity, variety, and value of large amounts of data. Formally, we can define predictive analytics as follows: C. Definition 1. ― It is process of extensive use of data, statistical and quantitative analysis, explanatory and predictive models, and fact-based management to drive business decisions and actions‖ (www.accenture.com/sitecollectiondocuments /pdf/accenture-business-intelligence-and- predictiveanalytics.pdf). 2.―It is a process that serves all levels of higher education and business and acts as a connector between the data collected, intelligent action that can be taken as a result of the analysis, and, ultimately, informed decision making.‖ Finally, we can break predictive analytics into the following categories: data mining, predictive Modeling, pattern recognition and alerts, Monte-Carlo simulation, forecasting, and root cause analysis. D. Role of Predictive Analytics in Education Education and research play a vital role in a nation‘s overall development process. For many countries, ―Quality Education‖ remains the basis for a sustainable and prosperous future. Several factors contribute to a quality education, including actuality and goal-based information; Relevance, such as learning- and discipline-specific environments; and context—for example, the delivery of an appropriate syllabus. Moreover, quality education will be value-oriented and will provide an understanding of industry needs. Acquiring all these factors in the same place to effectively develop successful and goal-oriented education systems is a difficult task. Several issues are crucial in education systems. 1. Resource utilization within an institute - This requires gathering real-life data through a communications channel and having the facility to store this data in various forms and types. Identifying which data belong to a particular category and maintaining it in the relevant repository or database is another challenging aspect.
• 3.
  29 International Journalfor Modern Trends in Science and Technology Volume: 2 | Issue: 10 | October 2016 | ISSN: 2455-3778IJMTST 2. Enrollment management and Decision prediction - University applicants have various options when it comes to the branch of study, courses, and programs in which they can enroll; this access to options creates confusion. Perfect prediction thus helps society by producing well- trained and skilled professionals to better serve societal needs. 3. Curriculum development - Companies and recruiters still find it difficult to find students equipped with the skills these organizations require. So, there is a need for a quality education system that can help planners design a curriculum focused on the demands of the future- workforce. 4. Adoption of standards - It‘s difficult to obtain best practices in online courses; hence, to promote quality in online education, standardization is unavoidable. Predictive analysis is one way to effectively address these issues. By analyzing current and historical facts to make predictions about the future, decision makers can take action and make decisions today to attain tomorrow‘s goals. Predictive analytics in educational systems comprises of following steps, as shown in Figure 1. Fig 1: The main steps for predictive analytics in educational systems. Predictive analytics helps education stakeholders in the decision-making process. II. LITERATURE SURVEY Predictive analytics as a valuable tool [2] with which to engineer positive change throughout the student life cycle. As the cost to recruit a student rises, it becomes ever more important to retain students until they graduate, which will: 1. Improve student learning outcomes. 2. Improve retention and graduation rates. 3. Improve the institutional return on investment (ROI) on recruitment costs. 4. Increase operational efficiency. 5. Help the institution demonstrate success in a key area of focus for accrediting agencies and the Federal government. 6. Demonstrate positive efforts to other important entities (e.g., state legislatures that allocate funding to public colleges and universities). III. PROPOSED WORK Use Case: University Experience Now we next examine the following use case, which we conducted in the any Department of Engineering. We wanted to answer the following questions: What will be the cutoff for the particular college with course-wise? • Which students will enroll in particular course programs? • What is current demand, which programs are trending, [4] which are becoming obsolete? •Which students are likely to drop out? • What will be the fees structure in the upcoming years? • What is the level of satisfaction of students in the current education system? • Which students will need assistance to graduate (future grade prediction)? To answer these questions, we applied predictive analytics. We based this work primarily on the first two categories: data mining and predictive modeling. We executed predictive analytics according to the four steps outlined in the below section. STEP 1: COLLECTION AND PREPROCESSING OF DATA Data collection is the first and foremost step in the predictive analytics process. To plan for quality education, it is necessary that any analysis navigate an ever-expanding sea of educational data and focus on gathering knowledge that can improve future prospects. Educational data combine offline data, online interaction data, and uncertain data. Offline data includes learner/educator information, students‘ attendance records. Online and interaction data would be distance and Web-based education, computer-supported collaborative learning, social networking sites and online group forums, email, chat transcripts, and so on. Uncertain data include scientific measurement techniques and
• 4.
  30 International Journalfor Modern Trends in Science and Technology Predictive Analytics in Education Context heterogeneity in designing data warehouses, sensor-generated data, privacy preservation process data, and data summarizations. Collected data must be preprocessed—that is, cleaned, transformed, and integrated. STEP 2: BUILD THE PREDICTIVE MODEL To gain insight and perceive knowledge from educational data, it‘s necessary to build a predictive model. Different modeling techniques are used by distinct systems and vendors. The following are the most-used commercial and open source model building environments (see www.slideshare.net/idigdata/practical-predictive -analytics-with): • decision tree—predicts categories of item class (supervised learning); • clustering—discovers data clusters (unsupervised learning); • Association and prediction—determines what occurs together and what will happen; • Divergence detection—finds changes or deviations; • Estimation and time series [5]—predicts a continuous value; • link analysis—discovers relationships; • Neural network—learns by example (supervised learning); STEP 3: MODEL VALIDATION Validating the model is necessary to tune it with new data—that is, to measure its accuracy. For a Nonbiased measurement of accuracy, a previously unused (that is, not included in training) sample should be used. A high rate of accuracy in the model depicts the maximum prediction accuracy that is necessary to truly affect the organization‘s baseline. STEP 4: ANALYTICS-ENABLED DECISION MAKING From the student viewpoint, our analysis lets students predict the most suitable program and engineering branch from the variety offered. It increases employability [6] and enables them to pursue higher study or research due to satisfaction with the chosen branch and ability to perform well during their course tenure. Finally, students can identify the level of risk in pursuing particular branch of study by predicting their grades. From an education planner or administrator viewpoint, branch prediction helps an institution estimate the number of students seeking admission in a particular branch. It also helps predict the approximate strength of students in a particular branch, and identifies popular branches of study. Through this use case, we observe that predictive analytics plays a vital role as an educational management tool to improve the quality of education. IV. TRENDS IN DATA ANALYTICS Due to the exponential penetration of the Internet, existing business analytics practices aren‘t sufficient to harness the potential of the millions of users who visit organizations‘ websites. This change in requirements has made Web-based business analytics and cloud-based analytics as a service (AaaS) the most preferred, cost-effective business analytics solutions. In this crowded market, establishing and maintaining positive brand awareness and client loyalty through a regularly updated, well-written website with rich content is an indispensable marketing tool. With social networking sites such as Facebook, Twitter, and LinkedIn, the value of positive word-of-mouth exposure and trust is more important than conventional marketing in terms of reduced costs because it replaces labor-intensive business analysis activity with software- supported activity. V. CONCLUSION The predictive analytics ‗prospects for enhancing the quality of education to sustain a nation is discussed. I have also examined the critical issues and challenges of higher education— specifically, technical education and the role that predictive analytics can play in addressing these issues. These theoretical discussions are supported by the use case, which demonstrates how predictive analytics are useful in higher (engineering) education. The exponential deployment of cloud computing in education is likely to facilitate cloud-based predictive analytics in a cost-effective and efficient manner. ACKNOWLEDGMENT This survey would not have been possible without the kind support and help of many individuals. I would like to extent my sincere thanks to all of them. I am highly indebted to Prof. Rachana Satao, Department of Computer Engineering of Smt. Kashibai Navale College of Engineering affiliated to Savitribai Phule Pune
• 5.
  31 International Journalfor Modern Trends in Science and Technology Volume: 2 | Issue: 10 | October 2016 | ISSN: 2455-3778IJMTST University for her guidance and constant supervision as well as all the other staff members for providing important information regarding the survey. REFERENCES [1] Jindal Rajni and Dutta Borah Malaya, ―PredictiveAnalytics in a Higher EducationContext―IT Professional ,2015IEEE Journals & MagazinesVolume: 17, Issue: 4,Pages: 24 - 33, DOI: 10.1109/MITP.2015.68 [2] Shankar M. Patil ―Predictive Analytics in Higher Education‖, International Journal of Advanced Research in Computer and Communication Engineering Vol. 4, Issue 12, December 2015 [3] A.V. Barneveld et al., ―Analytics in Higher Education: Establishing a Common Language,‖ Educause, Jan.2012; [4] KarthikeyanNatesan Ramamurthy, Moninder Singh, Michael Davis, J. Alex Kevern,Uri Klein, and Michael Peran ,‖ Identifying Employees for Re-Skilling using an Analytics-Based Approach ―,2015 IEEE 15th International Conference on Data Mining Workshops, pp. 345-354,DOI 10.1109/ICDMW.2015.206 [5] Matthew Malensek; SangmiPallickara; ShrideepPallickara, ―Analytic Queries over Geospatial Time-Series Data Using Distributed Hash Tables ―,2016 IEEE Transactions on Knowledge and Data Engineering,DOI: 10.1109/TKDE.2016.2520475 [6] JosepBerral;Nicolas Poggi; David Carrera; Aaron Call; Rob Reinauer; Daron Green ,―ALOJA: A Framework for Benchmarking and Predictive Analytics in Hadoop Deployments‖ ,Computing Year: 2015, Volume: PP, Issue: 99Pages: 1 - 1, DOI: 10.1109/TETC.2015.2496504 [7] P. Raj and G.C. Deka, ―Big Data Predictive and Prescriptive Analytics,‖ A Handbook of Research on Cloud Infrastructure for Big Data Analytics, IGI Global, 2014,pp. 370–391; doi:10.4018/978-1-4666-5864-6.ch015. [8] Predicting the Future of Predictive Analytics, SAP report, Dec. 2013; http://tinyurl.com/ooygmtq. [9] E. Sigel, ―Seven Reasons You Need Predictive Analytics, ―Prediction Impact, 2010; http://www-01.ibm.com/common/ssi/cgi- bin/ssialias?infotype=SA&subtype=WH&htmlfid=Y TW03080USEN. [10] R. Jindal and M. Dutta Borah, ―A Survey on Educational Data Mining and Research Trends,‖ Int‘lJ. Database Management Systems, vol. 5, no. 3, 2013, pp. 53–73.