<![CDATA[Proportional Feature Rough Selector (PFRS) is a feature selection method developed based on rough set theory (RST). The development is carried out with detailed version of RST. Beside the definition of lower and upper approximation, PFRS dividing the boundary region into two sections called Member Section (MS) and Non-Member Section (NMS). Howerer, the use of PFRS is still limited to binary text classification such as spam filtering and sentiment analysis. In the other hand, PFRS is developed without consideration of each feature’s correlation with another feature in dataset. This study aims to make an adaptation of PFRS for multi-label classification, not only for text data, but also another type of data like categorical and numeric data. This study using two public dataset, Netflix TV Shows and Ted Talks. PFRS with correlation is tested using four classic classification method such as DT, KNN, NB and SVM. Results showed that using PFRS in multi-label classification is useful to increase classification performance in each method up to 23,76%. This study also showed that using PFRS with correlation consideration can speed up the classification processes.]]>
Copyrights © 2021
