An Optimization of Fuzzy Rough Set Nearest Neighbor Classification Model Using Krill Herd Algorithm for Sentiment Text Analytics

Document Type


Publication Date



In recent years, social networking sites like Twitter and Facebook are more popular through which emotions and thoughts of its users are shared. Detection and classification of emotions, expressed in social media content, is beneficial for several applications in the domains like e-commerce, politics, social welfare and so on. Several works have been conducted earlier focusing sentiment and emotion analysis. These works primarily concentrated on single-label classification, leaving beside the correlation among different emotions expressed by an individual. In this view, the current research article presents a new sentiment and emotion classification model using Fuzzy Rough Set Nearest Neighbor (FRSNN) with Krill Herd (KH) algorithm i.e., FRSNN-KH. The proposed FRSNN-KH algorithm involves preprocessing, feature extraction and classification. Initially, preprocessing is executed to remove the unwanted words present in the tweet. Next, the features are extracted from the pre-processed tweet by following Bag of Words (BoW) model. Afterwards, FRSNN-based classification process is carried out to segregate the instances under different class labels. Finally, soft computing-based KH algorithm is applied to optimize the rule generation sets by FRSNN model. The presented FRSNN-KH algorithm was experimentally validated using benchmark dataset. The simulation outcome inferred the goodness of the proposed FRSNN-KH algorithm over compared methods under several dimensions.

Publication Title

Studies in Fuzziness and Soft Computing

First Page Number


Last Page Number




This document is currently not available here.