Data Mining For Academic Intervention

Published Date: 02 Nov 2017

ABSTRACT

Internet has made the world a real global village that places educational organizations in a very high competitive environment. The educational organizations should improve quality of their services to stakeholders such as students, faculty and future employers to get more competitive advantages over other competitors. Every educational organization, small or big, has the need to make use of the large scale data available and hopefully turn it into a predictive/analytical model that supports decision making process. Data mining is a young interdisciplinary field, the task of discovering interesting patterns/knowledge from large amounts of data stored in database, data warehouse or other information repositories.

Through Data Mining we can effectively address the challenges for improving the quality to provide new knowledge related to the educational processes and entities to the managerial system.

Keywords: Data Mining, Educational Databases, Performance Indicators

INTRODUCTION

One of the biggest challenges that higher learning institutions face today is to improve the quality of managerial decisions. The managerial decision making process has become more complex as the complexity of educational entities increase. Educational institute seek more efficient technology to better manage and support decision making procedures or assist them to set new strategies and plan for a better management of the current processes. This knowledge can be extracted from historical and operational data that reside in the educational organization’s databases using the techniques of data mining technology. Data mining techniques are analytical tools that can be used to extract meaningful knowledge from large data sets.

This paper presents the capabilities of data mining in the context of higher educational system by proposing the various performance indicators for higher education institutions to enhance their current decision processes.

2. INDIAN HIGHER EDUCATION - CHALLENGES

One of the significant facts in higher learning institution is the explosive growth of educational data. The data are increasing rapidly without any benefit to the management. We believe that to manage this difficult task, new techniques and tools for processing the large amount of generated data in business processes and extracting some useful knowledge and information are required. Data mining techniques are analytical tools that can be used to extract meaningful knowledge from large data sets. This paper addresses the capabilities of data mining in higher learning institution by proposing the various performance indicators and guidelines of data mining application in education. It focuses on how data mining may help to improve decision making processes in higher learning institutions.

In order to propose a new guideline in term of indicators, one must understand the data mining and decision making processes in higher learning institutions. In this regard the literature survey highlights the importance of this technology, what educational system lacks today and how data mining can be applied to the current educational system.

During the recent decades, the education system in the globe particularly the higher and lower systems have become business oriented with the goal of earning money, giving less importance for quality and goodwill. Indian education system plays predominant role in the global competition facing challenges in it. Due to lack of optimum level of investment in education, India is not able to achieve the desired goals of quality. Technology and Internet techniques are shrinking the world into a global village, which increases world- wide competitiveness. There is every need of optimal investment on higher education in India; this will ensure the availability of internationally acceptable highly skilled manpower.

The Indian higher education system has occupied major part in unorganized sector. There is mushroom growth of technical and professional institutions and proportional growth of student strength in private sector. During the past few years, the unaided private higher education institutions are playing vital role in academic activities.

The growth of distance education and vocational training are also playing pre-dominant role in India’s education system. However, Indian higher education system is lacking fundamental challenges of approachable, equality and quality.

3. RELATED RESEARCH

Rapid advances in data collection and storage technology have enabled education organizations to accumulate vast amounts of data. The challenging task is to extract useful information from huge sets of data. Traditional data analysis tools and techniques are not well suited for the massive size of a data set. The following are some of the specific challenges that motivated towards the use of data mining techniques in higher education system. 1) Scalability: Because of advances in data generation and collection, education data sets with sizes of gigabytes, terabytes, or even pent bytes are becoming common. If data mining algorithms are to handle these massive data sets, then they must be scalable. 2) High Dimensionality: In higher education, it is now common to encounter data sets with high dimensionality. Traditional data analysis techniques do not work well for such high dimensional data. 3) Heterogeneous and complex Data: Traditional data analysis methods often deal with data sets containing attributes of the same type, either continuous or categorical. As the role of data mining in business, science, medicine, and other fields has grown, so has the need for techniques that can handle heterogeneous attributes.

Mining techniques can be used in educational environment. In such case, mining is called Educational Data Mining, concern with developing new methods to discover knowledge from educational databases (Galit, 2007)[1] . Data mining techniques and tools can help bridging knowledge gaps in higher education system.

Indicators are agreed measurement scales which identify the quantitative relationships between two variables. They are normally used as numerical values. Indicators are very important in determining the goals and the operational analysis of the educational system (Johnstone, 1976[2]; Johnstone, 1981[3]; Wako, 1988[4]).

The higher learning institution of a country deals with human factors and educating specialists needed by the community, educational promotion, research development, and providing a suitable environment for the country’s growth. Thus, the system essentially requires a principle which can express the qualitative characteristics of the higher learning institution to some quantitative values, and facilitates evaluating the functionalities. This principle is summarized into indicators.

To evaluate the different aspects of the higher learning institution, "performance indicator" is used as one of the main educational system indicators (UNESCO, 2006b) [5]. Educational performance indicators have been known as the base for educational system methodology improvement. There have been many studies (Oakes, 1986 [6]; Scheerens, 1990[7]) which present the importance of performance indicator as a quality improvement tool in an educational domain. They indicate that performance indicator is vital for educational system improvement. The earlier studies (Yang et al., 1999[8]; Fitz- Gibbon and Tymms, 2002[9]; Van Petegem et al., 2004[10]) state that other than performance indicator, an additional step for supporting educational system improvement, which is built on information from performance indicator, is more important. This step is called educational feedback. The feedback should be up-to-date, valid and reliable.

Dr. A. Gopal and Chandrani Singh [11] suggested the use of data mining tools to monitor and improve the performance of the faculty. Dr. K. S. Ramaswami and S. PrakashKumar [12] described an effective analytical predictive model framework using data mining clusters in evaluating the quality of educational institutions in terms of student performance, curriculum, and faculty skill sets. Bharadwaj and Pal [13] applied classification method decision tree to predict the performance of students in end semester examination, which helps in identifying the students who need special attention in advance.

4. DATA MINING FOR ACADEMIC INTERVENTION

Data mining is a technology processes large volumes of data by combining traditional data analysis methods with sophisticated algorithms for processing. DM techniques enable us to analyze old types of data in new ways and also explore and analyze new types of data. Recently, there has been an increasing trend that data mining techniques and tools are used to improve the efficiency of higher educational institutions.

1. Weka: Weka (Waikato Environment for Knowledge Analysis) is a package that offers collection of tools that they can use for data mining. Weka algorithms can be used to classify the available data, used to filter the data contents and use it for data plotting. Weka includes tools for data clustering and regression, association rules and attributes evaluator. New machine learning schemes can be developed by Weka. It is open source and freely available. It is platform-independent and relatively easier to use. It provides flexible facilities for scripting experiments.

2. KEEL (Knowledge Extraction based on Evolutionary Learning): It has been designed for both research and education. It makes use of evolutionary algorithms such as genetic algorithm. It includes algorithms for Classification Discovery, Cluster Discovery, Regression Discovery, Association Discovery, Data Visualization, Discovery Visualization and data pre-processing. It consists statistical analysis library to analyze algorithms. It reduces programming work and easy-to-use. It is open source and freely available.

3. XLMiner - an add-in for MS Excel: XLMiners has the capabilities such as Partitioning, Prediction, Association rules, Classification and clustering. It is also called a Business Intelligence tool. XLMiner also provides machine learning oriented algorithms. It provides several ways to try to solve a problem. But it is the task of miner to select appropriate method would be most appropriate to his problem. It is easy-to-learn and easy-to-use. It is inexpensive.

5. XLSTAT - The Solution for Data Analysis and Data Mining in Microsoft Excel: This tool has capability for multivariate data analysis and data mining. XLSTAT statistical software facilitates the data analysis process by providing you topmost tools for the visualization of data. It facilitates factor Analysis, principal component analysis, regression, and classification. It is easy to use.

6. SPSS (Statistical Package for the Social Sciences): It facilitates algorithms for Correlation, regression, classification, data reduction and clustering. It provides easy-to-use interface and provides in-depth Statistical capabilities. The problem with SPSS is lots of irrelevant output. The routines used not properly documented, and difficult to produce customized analysis.

7. SAS (Statistical Analysis System): This package includes decision trees, neural networks, regression and clustering capabilities, heuristics and the latest statistical algorithms for finding patterns and predicting outcomes. It easier to prepare data for analysis since SAS includes data access, management and cleaning functions. The execution time is less because it supports multithreaded algorithms. It uses hardware resources most efficiently. Enriched data leads to better modeling and more reliable results. It is expensive.

8. KnowledgeSEEKER: It has data mining capabilities include data preparation, data profiling, data visualization, Decision Tree analysis, and strategy design and deployment. It provides graphical, wizard-driven operations, which enables easy-to-use. Data preparation capability of KnowledgeSEEKER allows users to easily extract, manipulate and transform. It supports multiple data sources such as SQL, Microsoft Excel and any database system via ODBC to import data. It is widely used for sales and marketing to improve decision making and performance.

9. KnowledgeSTUDIO: It was designed for market-leading data analysis. It provides features of KnowledgeSEEKER and many advanced modeling and predictive analytics features. It includes broad range of advanced models and algorithms such as neural networks, linear and logistic regression, cluster analysis, principal component analysis, and scorecards. Text analytics capability of KnowledgeSTUDIO combines structured and unstructured data analysis and improves accuracy with greater numbers of predictive variables. It is ease of use. It provides menus and wizard-driven shortcuts, and superior visualization of results.

10. TANAGRA: It is easy to use DATA MINING software for academic and research purposes. It allows researchers to easily add their own data mining methods, and to compare their performances. Free access to source code. It implements various supervised learning algorithms such as an interactive and visual construction of decision trees. It also includes other paradigms such as clustering, factorial analysis and so on.

11. GMDH (Group Method of Data Handling) shell: It is a mathematical tool for predictive modeling and data mining. It can be used for any forecasting or data mining task at faster rare with minimum efforts and accurately. It provides easy-to-use interface. It can be downloadable for free.

12. DBMyne: It is a data mining tool for decision cube analysis. It helps you analyze and present data stored in computer databases with file formats: * DBase, * Paradox, * FoxPro.

Data mining is a powerful analytical tool that enables educational institutions to better allocate resources and staff, proactively manage student outcomes, and improve the effectiveness of alumni development. With the ability to uncover hidden patterns in large databases, community colleges and universities can build models that predict—with a high degree of accuracy—the behavior of population clusters. By acting on these predictive analysis tools, educational institutions can effectively address issues ranging from transfers and retention, to marketing and alumni relations.

5. CONCLUSION AND FUTURE SCOPE

From the above literature survey in higher learning institution it is possible to derive the following conclusions i) Based on the fact that the performance feedback perceived in an educational institution should be accurate, up-to-date, reliable, valid, and toward the goals of educational improvements more effective strategies should be taken into account to improve the feedback from an educational domain. Not only the performance indicator is essential for indicating the actual state of an education system, it is also vital to develop a methodology for educational system performance feedback. ii) Improving the feedback of an educational domain implies further analysis and investigation in the forming components of performance indicator. Data mining is able to improve the educational system in each component of the performance indicator. This improves the feedback from the system.

In this study, the components of performance indicator are based on Vlasceanu etal. (2004) [14] definition. The performance indicators work efficiently only when they are used as part of a coherent set of input indicator (Human resource, financial resources, and sector resources), process indicator (Educating methods, qualitative and quantitative educational improvements, such as registration and dropout rates) and output indicator (alumni and graduates).

Educational Data Mining (EDM) is an emerging field, concerns with developing methods that discover knowledge from data originating from educational environments. In this study, we have explored the problems faced by Indian higher education, also the capabilities of different data mining tools available and background of educational systems based on indicators.

This research still continues with how data mining can help to improve the performance indicators which higher learning institutions can adopt in general. It also includes the mining of educational topics, such as curriculum development, predicting student performance and advising students, evaluating the teacher’s performance, correlations between the student skills and the skills required by the organizations, to realize learning gaps, and also improve teaching methods and educational management processes. The use of the data mining techniques in education provides us with more varied and significant findings, and leads to improve quality of education.