IMPLEMENTATION MODEL OF MACHINE LEARNING ON NEWS CLASSIFICATION INFORMATION SYSTEM
Text classification is a grouping of text data that has not been grouped into groups automatically. The classification of news texts is done by the editor by reading the entire text first, so it takes a long time. For that we need a way to classify news automatically that can cut down the process. This study aims to classify news texts automatically with a text mining approach. This study uses the K-Nearest Neighbor algorithm which has simplicity and efficiency in classifying various types of text. To simplify the research flow, the CRISP-DM (Cross-industry standard process for data mining) method is used, which is the standard in conducting data analysis in industrial applications. The results showed satisfactory results, namely precision, recall, F1-Score and accuracy reached 95% with a value of k = 11. After the text classification application was made and an experiment was carried out by entering several new news texts, only a few seconds the text could be classified correctly. This study shows that the K-Nearest Neighbor algorithm can be used for news text classification and text classification applications can help cut the classification process time.
Text Mining, News Text Classification, K-Nearest Neighbor, CRISP-DM.