Classification of Positive and Negative Sentiments Using the K-Nearest Neighbor Algorithm on iQIYI Aplication

Abstract

In the current state of the Covid pandemic, the government has implemented restrictions on community activities or PPKM, which has an impact on the number of cinemas in the country temporarily closed to reduce the spread of the virus. The number of films that have been postponed for release due to this outbreak and also the decreasing use of VCDs / DVDs have made movie streaming applications begin to be favored by the public, one of which is the iQIYI movie streaming application. iQIYI is a movie streaming app launched in April 2010, so that users can know that the iQIYI application is considered good is to do a sentiment classification on the application. Therefore, this study aims to implement sentiment classification in review data using the K-Nearest Neighbor (K-NN) algorithm. K-NN itself is an algorithm that functions to classify data based on its learning data (train data sets). The data used is iQIYI user reviews as many as 400 review data, the first stage carried out is the data cleaning process or Pre-Processing, the next step is to design a K-NN algorithm model in RapidMiner Studio software to process sentiment classification. The test results using 400 review data using the K-NN algorithm obtained an Accuracy value of 99.50% then a Precision value of 100% and a Recall value of 99.44%. Which means that this study managed to get the best and best algortima in classifying positive reviews and negative reviews against the iQIYI application.