CURE AANN An Effective Hierarcical Clasification In Education Datamining Techniques To Valuate Students Performamce

Abstract

Educational data mining (EDM) is relatively latest research area. It still has many open newlineissues. EDM searches for an enhanced understanding of the learning process and involvement newlineof students on it.Current issues identified in EDM especially engineering student s newlineperformance arises mainly due to the huge dimension of records in educational databases. newlineOther issues identified in EDM mainly incorporate identification or prediction of weak newlineengineering student s performance. Several researches have been done so far for predicting newlinethe performance of weak students to make improvements in their performance. But in most newlineresearch works, very few attributes are considered in order to predict the students newlineperformance. It is always penetrating for quality progress and cost effectiveness of the newlineeducation system. So a better mining algorithm has to be implemented to effectively identify newlinethe behavior of weak students. In most of the works, the attributes identified are irrelevant newlineand are neglected as missing attributes leading to inconsistent results. The succeeding newlinedrawback identified with the existing clustering algorithm, is the similarity achieved between newlinethe cluster partitions. Also with the better dataset, the hierarchical clustering algorithm breaks newlinedown due to non-linear time complexity. In order to reduce this drawback an efficient newlineAdaptive Artificial Neural Network Algorithm (AANN) will be proposed in this work. newline