Perlombongan Data Prestasi Pelajar Siswazah Menggunakan Kaedah Aruhan Berasaskan Atribut

This paper discusses in detail an attribute-oriented induction technique which is one of the data mining techniques. Many types of knowledge can be discovered through this concept such as classification rule, characteristic rule, quantitative classification rule and quantitative characteristic rule. In the implementation, generalization and summarization are the two fundamen- tals of concepts involved. Data generalization is executed into a set of relevant data by using elimination attribute procedure, concept-tree climbing, controlling generalization process by determining the threshold value for the attribute, propagation of counts, and other amount of value functions. Two algorithms are developed, which are attribute-oriented induction and standard query language instruction. The standard query language instruc- tion algorithm acts as a medium for the database to execute standard query language instruction to the database link. Both are related to each other in the execution of attribute-oriented induction which is based on standard query language. A database link design is developed to store all data of the postgraduate students of the Faculty of Information Science and Technology, UKM, which was tested using both algorithms. A set of classification rules containing knowledge on the performance of the postgraduate students is obtained. These rules are then tested with new data of those students in order to determine the accuracy of the classification rule. The experimental results show that the rules obtained can be used in the future: to determine the performance of postgraduate students.