Data Mining and Knowledge Engineering

Vijay Anand Sullare ¹, R. S. Thakur ²

Affiliations
1 M.G.C.G.V., Chitrakoot, Satna, M.P., IN
2 M.A.N.I.T., Bhopal, IN

Abstract

Higher education plays an important role in economy of any nation countries like India need a good higher education to face the challenges of this new era. Manifold growth has been found in the higher education in India in last decade. But there is need to focus more on our education system. The paper aims at the use of data mining techniques for improving the efficiency of higher educational institutions. The association rule mining Techniques can be applied to higher education processes, to help improve student's performance of an institution. This paper contains a methodology to examine the performance of engineering graduate student based on their continuous evaluation and locality. We present an approach based on association rule mining techniques to identify the strategies for improving the performance of students.

Keywords

Association Rule Mining, Recommendation System Educational Data Mining, Knowledge Representation, Higher Education System.

Full Text

     

Yachana Bhawsar ¹, G. S. Thakur ¹, R. S. Thakur ¹

Affiliations
1 Department of Computer Application, Maulana Ajad National Institute of Technology, Bhopal, M.P., IN

Abstract

Social networking sites are increasing their features day by day to gain the attention of users. There are lots of research works in this field. Out of many research areas this paper focuses on link prediction using soft computing technique. We used various features of social network and applied genetic algorithm to predict links. Selection of features to build chromosome is main task in genetic algorithm. Number of runs will get different chromosomes i.e. shown in results. Normalization of features is also done depending upon their priority. Results show that with the increase in dataset size chances of correct prediction increases.

Keywords

Social Network, Link Prediction, Genetic Algorithm.

Full Text

     

Vikrant Sabnis ¹, R. S. Thakur ¹

Affiliations
1 Department of Computer Applications, Maulana Azad National Institute of Technology, Bhopal, M.P., IN

Abstract

With increase in network technologies and number of users working on the network, attempts are being made to discover the useful knowledge from the secondary data. For retrieving knowledge large number of models, techniques and methods are evolving continuously in the area of web content mining. These techniques are becoming very critical for effective management of web sites in the variety of domains such as business, education and e-learning. Based on the prediction approach the user browsing behaviors can be guessed and this information can be utilized for building of proper web sites. This paper proposes star schema for web contents mining from the complex data which is multidimensional in nature. Further the association among web contents is explored using multidimensional ARM approach to know the surfing behavior of web users. At the end Performance computation of proposed work has been discussed, which shows improvement in the gain and implementation explains well the significance of multidimensional association rule in web content data. The paper also compares pros and cons with the traditional state of art approaches.

Keywords

Web Content Mining, Data Mining, Pattern Discovery.

Full Text

     

Vikrant Sabnis ¹, Neelu Khare ¹, R. S. Thakur ¹, K. R. Pardasani ²

Affiliations
1 Department of Computer Applications, Maulana Azad National Institute of Technology, Bhopal, M.P, IN
2 Department of Mathematics, Maulana Azad National Institute of Technology, Bhopal, M.P, IN

Abstract

Web content mining refers to description and discovery of useful information from the web contents/data/documents. Hypertext is one of the most common web content data that has hyperlinks in addition to text. These are modeled with multiple levels of details depending on the application. In this paper Karnaugh map model for multilevel association rule mining has been developed to investigate association relationships among hypertexts of a web site. Karnaugh map model needs single scan of data and stores the information in the form of frequency. Model adopts progressively deepening approach for finding large text sets by utilizing karnaugh map logic for finding frequent text sets at each level of abstraction. Frequent texts sets are generated by the karnaugh map model are used to discover strong association relationships among hypertexts at different levels of abstraction. Further the rules are categorized under three categories and their behavior is studied across the level of abstractions.

Keywords

Karnaugh Map Model, Multilevel Association Rules, Association Relationships, Frequent Text Set.

Full Text

     

Sachin Kamley , Shailesh Jaloree , R. S. Thakur

Abstract

For the last few decades’ education data mining has become one among foremost promising research areas. The only objective of this area is to explore data mining methods in order to analyze the student performance as well as impart the quality education for enhancing the performance of educational institutes. Data mining is the core part of the knowledge discovery process which is used to extract meaningful information from raw data. However, the various data mining techniques are proposed for achieving the most effective quality results. An Association Rule Mining (ARM) one of the well-known and popular data mining techniques which has been used extensively for educational perspective. In this study, higher education institute i.e. Government Girls College (GGC) data are considered and various attributes regarding student performance are analyzed for study purpose. Therefore, the various experiments based on support and confidence measures like 2%, 4%, 10%, 20% and 40% are conducted to generate interesting rules.  The major objective of this research study is to find the weaker students as well as those students who have bright performance in schooling level but could not be performed well on current semester exams due to certain reasons. However, teachers as well as parents can give particular attention to those students, whether they will perform better in the next semester or exams.

Keywords

Apriori Algorithm, Association Rule Mining, Confidence, Education Data Mining, Preprocessing, Prediction, Support, Weka 3.7.

Full Text