Enhancing Fuzzy Topsis to Improve Prediction Student On Selection Properly Majors at Vocational School
Agus Nursikuwagus, Lusi Melian, Tono Hartono
Universitas Komputer Indonesia
Abstract
Speed in the fuzzy topsis process, to get the expected results, can be applied using mathematical functions that have been formulated. This research has aimed to predict ability student on every test as prerequisite to enter the major. fuzzy topsis, with the criteria and alternative approaches, can be determined according to the problems applied. Problem in fuzzy topsis is not provided classification in the last step when we obtain many predictions classification. fuzzy topsis was only executed to get rank in a case. In order to solve that problem, we have added a function in the last step fuzzy topsis like rule base. Rule base is divided into four majors such as software engineering, animation, networking, and multimedia. To completed the prediction, we are introduced some criterion that deployed some assessment such as final examine, competency test, report, physical test, interview, and psychological tests. The result obtained for the process precision was 59.2%, and recall was acquired 60%. The reason why the precision and recall were not got a high value because the dataset is very short (over fit) only 270 to process in extended fuzzy topsis. Another reason is preferences of function was not proper for dataset and imbalance data, and dataset has centred in one favourite major is network and S/W engineering
Keywords: Fuzzy Topsis, Enhancing, Prediction, Membership Function, Majors
Topic: Informatic and Information System