Sentiment Analysis Based on Review of Puncak B29 Lumajang using Backpropagation Neural Network

Abstract

Sentiment analysis is a method that applies the concept of text mining to provide a classification that has positive, negative or neutral polarity for each sentence or document. The problem formulation carried out in this research is the role of sentiment analysis in analyzing reviews of the Puncak B29 Lumajang tourist attraction based on user comments on Google Maps. This research was carried out in 3 stages, starting with data collection in the form of a review of the Google Maps application which was carried out by scrapping data, carrying out text preprocessing, including case folding, tokenizing, stopwords, and stemming and categorizing each review according to sentiment using the Backpropagatin Neural Network (BNN) classification method. Sentiment classification based on Puncak B29 reviews on Google Maps using Backpropagation Neural Network has the best accuracy, recall and F1-score evaluation results for a total of 50 iterations, each with an average value of 97.33%, 100.00% and 98.47%. Meanwhile, the best precision value for the number of iterations is 10 iterations, which has an average value of 99.72%. From this description, it can be concluded that the evaluation value will get better along with the number of iterations carried out throughout the classification process.