Consumers' Attitude Toward Cloud Services: Sentiment Mining of Online Consumer Reviews

Automatically generated cloud services users' experiences summaries could aid potential consumers in selecting cloud services. This study proposes a novel methodology for analysing consumer's attitude toward cloud services by applying sentiment mining on online consumer reviews. The cloud services were collected across different web platforms, then analysed using sentiment analysis to identify the attituded of each cloud services review. The analysis conducted using a data mining tool namely RapidMiner and the proposed model is based on fours supervised machine learning algorithms: Nave Bayes, K-Nearest Neighbour (K-NN), Decision Tree and support vector machine. The results show that the prediction accuracy of the SVM-based TF-IDF approach (10-fold cross validation testing) and Naive Bayes TF-IDF approach (10-fold cross validation testing) is 88.29%. This indicates that Naive Bayes and SVM perform better in determining sentiment than in determining other classifiers.