Hybrid Corrective Critic Neural Network for Sentiment Classification in Community Media
Jayashree Jagdale, M. Emmanuel · 2019 3rd International conference on Electronics, Communication and Aerospace Technology (ICECA) · 2019
The issues faced by sentiment classification in sentiment analysis are a growing research field. The main goal of the sentiment classification is to classify the text into various sentiment classes, like negative class or positive class. Here, a technique for the sentiment analysis of the movie reviews from community media is introduced. Hybrid Corrective Critic Neural Network is used for the purpose. The movie reviews collected from the community media are fed to pre-processing for removing unnecessary and redundant words from the data. After pre-processing, the feature extraction is carried out. Sentiment related features, such as Hashtags, All - caps, Emoticons, sentiment lexicon and elongated units from the reviews are used. Based on the extracted features, the sentiment classification is performed by Hybrid Corrective Critic Neural Network. Evaluation is carried out on the output generated by the developed technique based on metrics sensitivity, specificity and accuracy. The reviews from the movie review database are used for training the approach. The proposed Hybrid Corrective Critic Neural Network based sentiment classification is found to be effective with accuracy of 74.42%.