METHODS FOR CLASSIFYING COMMENT DATA OF ONLINE CUSTOMER IN VIETNAM BASED ON SUPERVISED MACHINE LEARNING

Abstract

The study aims to apply a supervised machine learning method to classify
comments data as customer product comments in online shopping. The study
conducted automatic data collection with 2,530 customer comments about
products on the top of e-commerce sites in Vietnam, then trained with
supervised machine learning models. to find the model that the best fits the
training dataset and apply this model to predict the comment content for the
entire dataset. The results show that the Machine Learning methods Support
Vector Machines (SVM), Decision Tree (DT) and Neural Network (NN) have the
best performance with classifying customer comments in Vietnamese. The
research results have reference value for comment mining applications in the
field of online business.