Scopus Indexed Publications

Paper Details


Title
A Sentiment Analysis Based Approach for Understanding the User Satisfaction on Android Application
Author
Md. Mahfuzur Rahman, Md. Tahsir Ahmed Munna, Shaikh Muhammad Allayear, Sheikh Shah Mohammad Motiur Rahman,
Email
mahfuzur.web@daffodilvarsity.edu.bd
Abstract

The consistency of user satisfaction on mobile application has been more competitive because of the rapid growth of multi-featured applications. The analysis of user reviews or opinions can play a major role to understand the user’s emotions or demands. Several approaches in different areas of sentiment analysis have been proposed recently. The main objective of this work is to assist the developers in identifying the user’s opinion on their apps whether positive or negative. A sentiment analysis based approach has been proposed in this paper. NLP-based techniques Bags-of-Words, N-Gram, and TF-IDF along with Machine Learning Classifiers, namely, KNN, Random Forest (RF), SVM, Decision Tree, Naive Byes have been used to determine and generate a well-fitted model. It’s been found that RF provides 87.1% accuracy, 91.4% precision, 81.8% recall, 86.3% F1-Score. 88.9% of accuracy, 90.8% of precision, 86.4% of recall, and 88.5% of F1-Score are obtained from SVM.

Keywords
NLP, TF-IDF, Sentiment analysis, Machine learning, Mobile apps review >
Journal or Conference Name
Advances in Intelligent Systems and Computing
Publication Year
2020
Indexing
scopus