<![CDATA[A thesis is a scientific paper that discusses a particular topic or field, written and designed by a student with the guidance of a supervisor as a condition for obtaining a bachelor's degree. In the last five years, various titles and thesis topics have been designed by students of the Faculty of Computer Science, Universitas Brawijaya. The research data used is the title of the thesis with labeling based on the thesis topic of FILKOM UB students on the UB repository website. The topic class of the title of this classification thesis consists of areas of interest in FILKOM, namely Image Processing, Software Engineering, Decision Support Systems. The data used are 225 data with each topic totaling 75 data. This study aims to determine the classification of the thesis title data that has been labeled with the topic with the Support Vector Machine Algorithm using the Radial Basis Function kernel and the selection of Chi-Square features. The results of the system evaluation obtained by testing applying 10-fold cross-validation and 50% of the features used with Chi-Square are accuracy = 0.88, recall = 0.9063, precision = 0.881, f1-score = 0.8789 with the best parameter values ​​from the Support Vector Machine namely parameter (sigma) RBF kernel = 2.5, parameter (lambda) = 0.3, parameter (gamma) = 0.001, parameter C (complexity) = 0.1, parameter (epsilon) = 0.00001, and iteration = 75.]]>
Copyrights © 2022
