Sentiment Analysis about E-Commerce from Tweets Using Decision Tree, K-Nearest Neighbor, and Naïve Bayes
Achmad Bayhaqy, Sfenrianto Sfenrianto, Kaman Nainggolan, Emil Robert Kaburuan · 2018
Data mining can be used for data analysis of social media users who visit E-Commerce. This study uses data mining techniques aimed at comparing the classification in sentiment analysis from the views of E-Commerce customers who have been written on Twitter. The data set is derived from tweets about E-Commerce in Tokopedia and Bukalapak. Text mining techniques, transform, tokenize, stem, classification, etc. are used to build classification and analysis of sentiment analysis. Rapidminer is also used to assist in making analysis sentiments for comparison by using three different classifications in the dataset with the Decision Tree, K-NN, and Naïve Bayes Classifier approaches to find the best accuracy. The highest result of this study is the Naïve Bayes approach with an accuracy of 77%, precision 88.50% and recall of 64%.