Twitter Sentiment Analysis on Digital Payment in Indonesia Using Artificial Neural Network

In the rapid development of technology, the need for big data processing is increasingly important, especially in the context of digital transactions such as e- wallets in Indonesia. On the other hand, sentiment analysis of digital payment platforms via Twitter requires fast and accurate data processing, but often faces challenges in managing big data and optimal classification quality. This study uses the Term TF-IDF method for text preprocessing and Artificial Neural Network (ANN) for sentiment classification. The preprocessing process includes case folding, removing numbers and punctuation, tokenization, filtering, and stemming. For classification, ANN is used which is optimized with the Backpropagation and K-fold Cross Validation algorithms to improve the accuracy of the model in grouping positive and negative sentiments from tweets about digital payment platforms. Through this approach, the study produces a sentiment classification model in analyzing big data. The results in this study are Gopay gets a positive value and gets the first value in sentiment assessment with an accuracy rate of 72% using ANN. Of the 5 digital payments that received a negative value and ranked last, namely Link Aja with an achievement rate of 43%. Based on these results, it shows that this approach contributes to identifying consumer sentiment towards e-wallet platforms, which is useful for developing digital marketing strategies. The contribution given is in improving sentiment analysis of digital payment platforms by utilizing Big Data processing technology and machine learning, so that it can be used to improve services and marketing strategies based on user data.