A Comparison of Text Classification Methods: Naïve Bayes and Support Vector Machine for E-Commerce Item Classification

Currently, buying products from the e-commerce industry has already become the norm for many people. It is important to categorize the item that is sold in e-commerce so that the customer can easily find the item they want to buy and have a better experience browsing through the application. We have found a dataset to use for our experiment that we will preprocess so we can use it without any problem. The experiment will consist of looking for the accuracy, precision, recall, and f1-score of each algorithm with the dataset we use and compare it so we can find which algorithm is better. Out of 6 experiments we have done, we found out that SVM outperforms Naive Bayes four times. From that we can conclude that SVM is overall better than Naive Bayes with 96.6% accuracy and 0.967 F1-Score.