Title :
Analysis and Predictions on Students´ Behavior Using Decision Trees in Weka Environment
Author :
Bresfelean, V.P.
Author_Institution :
Babes-Bolyai Univ., Claj-Napoca
Abstract :
Decision trees classifiers are simple and prompt data classifiers as supervised learning means with the potential of generating comprehensible output, usually used in data mining to study the data and generate the tree and its rules that will he used to formulate predictions. One of the major challenges for knowledge discovery and data mining systems stands in developing their data analysis capability to discover out of the ordinary models in data. The excellence of a university is specified among other concerns by its adapting competence to the constant changing needs of the socio-economic background, the quality of the managerial system based on a high level of professionalism and on applying the latest technologies. This article represents an implementation of a J48 algorithm analysis tool on data collected from surveys on different specialization students of my faculty, with the purpose of differentiating and predicting their choice in continuing their education with post university studies (master degree, Ph.D. studies) through decision trees.
Keywords :
computer aided instruction; computer science education; data analysis; data mining; decision trees; further education; human factors; learning (artificial intelligence); J48 algorithm analysis tool; data analysis; data mining; decision trees classifier; higher education; knowledge discovery; socio-economic background; student behavior; supervised learning; Algorithm design and analysis; Biomedical imaging; Business; Classification tree analysis; Continuing education; Data analysis; Data mining; Decision trees; Economic forecasting; Object oriented modeling; C4.5; Decision tree; J48; algorithm; data mining; performance measures;
Conference_Titel :
Information Technology Interfaces, 2007. ITI 2007. 29th International Conference on
Conference_Location :
Cavtat
Print_ISBN :
953-7138-10-0
Electronic_ISBN :
1330-1012
DOI :
10.1109/ITI.2007.4283743