An Improvised Ideology based K-Means Clustering Approach for Classification of Customer Reviews

Background/Objectives: To provide a framework for improving the classification of customer reviews on products. Methods/Statistical Analysis: We propose an integrated framework for classifying the customer reviews based on the textual analysis with constraint-based association rules using ontology. It involves preprocessing the customer reviews including symbols and handling feature extraction. An improved K-Means algorithm with ontology is proposed to consolidate the reviews based on textual analysis method to handle reviews that represent at least one feature of the product. Findings: The empirical results reveal that the accuracy of the system increases with the use of ontology and modified K-Means algorithm, improving overall performance of the recommendation system. Combining preprocessing and ontology considerably improves the accuracy of classification of customer reviews.Applications/Improvements: The proposed approach can be used to recommend product based on users' review.