Sentiment Analysis on Platform X (Twitter) Towards The 2024 General Election Using The Probabilistic Neural Network Algorithm

Abstract

Sentiment analysis is the process of analyzing an opinion or public opinion regarding a phenomenon that has occurred, is currently occurring or will occur. Sentiments that are commonly discussed are the public's assessment of an object positively, negatively or neutrally. Twitter is the most popular media used to express all forms of public emotions and opinions in the form of tweets or text. The issues discussed in it include many events, such as the 2024 general election which will be held in Indonesia. This media is easily accessible to many people to show other people's opinions regarding existing phenomena. This research discusses the topic after the 2024 general election with opinions based on three sentiment classes, namely positive, neutral and negative. The aim of this research is to build a sentiment analysis system by applying the Probabilistic Neural Network algorithm as a classification model. The method used is data collection, cleaning, preprocessing, TF-IDF word weighting, labeling, classification models, and evaluation of results. The data used amounted to 2002 data with a division of 1035 positive tweets, 693 neutral tweets and 274 negative tweets. The program was built using Google Colaboratory and the Python programming language. Testing was carried out with 3 (three) comparisons, namely 90:10, 70:30, and 50:50. By comparing 90% training data and 10% testing data, the greatest model accuracy was obtained with a value of 88.42% and taking into account the evaluation using the confusion matrix and precision parameters of 89%, recall of 88%, and f1-score of 88%. Evaluation was also carried out for website-based applications on new data with an accuracy value of 66%.