Sentiment Classification with PSO Based Weighted K-NN

Abstract

Millions of data are collected daily in the social media environment. This data mostly reflects people's ideas in a certain topic. Emotion analysis is a new data mining subfield and is usually concerned with information extraction and identification from emotions in social media. Twitter is a platform where users share their ideas through messages. These messages can be classified to analyze the feelings of users in a specific context and to find out the points of view. In this study, a K-nearest neighbor classification algorithm, which is based on particle swarm optimization, has been proposed to classify twitter data. The proposed study is compared with different methods on specific benchmark data sets and the accuracy of the method has been proven.