FEATURE SELECTION BERBASIS ABC-SVM DAN PSO-SVM DALAM MASALAH KLASIFIKASI

FEATURE SELECTION BERBASIS ABC-SVM DAN PSO-SVM DALAM MASALAH KLASIFIKASI

• By Dharmawan, Weiskhy Steven;
• Tesis
• 286

Dimensional reductional is necessary in the data analysis for large dimensional dataset. In this study proposes a method for reducing dimensions based on the Artificial Bee Colony algorithm. The proposed method serves as a feature selection method to select an optimal subset or feature that meets the set objectives for evaluating fitness of a feature set. Support Vector Machine is used to provide classification results from feature selection performed by Artificial Bee Colony. It is also conducted the model of Particle Swarm Optimization as a comparison method in the selection of other features. The results of the UCI Repository 4 is obtained show an increase of the comparison method in the calculated AUC value. The Vowel UCI Repository dataset has a substantial increase with 12.6% of the comparison method, Sonar dataset with 0.1% increase, while Wavefrom 0.5% dataset, and 2.2% Heart Statlog dataset in Area Under Curve (AUC) value calculation

• Katakunci
• URL http://
• Penerbit ,

REFERENSI