ABSTRACT
In the business world, understanding the level of customer satisfaction with services is very important. As time goes by and there is a lot of competition, companies must be able to provide services and understand the level of customer satisfaction. Service has been proven to be one of the influences on customer satisfaction, but most companies think too much about quantity compared to quality of service. To get data in the form of reviews or comments using web scraping techniques on Twitter. Therefore, we need a system that can classify the level of customer satisfaction efficiently from reviews or comments on Twitter. This research also classifies customer satisfaction from reviews by implementing a customer satisfaction classification system for IndiHome services using the TF-IDF and K-Nearest Neighbor methods. The TF-IDF method is used to convert text with the frequency of occurrence of each word based on customer reviews, while the K-Nearest Neighbor method is used to classify the level of customer satisfaction. Testing the level of accuracy using the Confusion Matrix, from the results of experiments that have been carried out, it produces an accuracy of up to 98%.

Keywords:
Customer Satisfaction, K-nearest neighbor, TF-IDF reviews, Twitter, Confusion matrix