Cluster Text Random Opinion Tweet In Yogyakarta Using Automatic Clustering

Abstract

Tweet Besides making computations difficult, the data obtained is also inefficient and complicated to interpret. Therefore, it is necessary to explore how to overcome these problems. This study proposes an approach to find the global optimum and make automatic grouping by analyzing moving averages, namely K-Means Automatic Clustering. So the purpose of this study was to explore and evaluate high-dimensional data from a collection of tweets, namely random opinion text tweets in Yogyakarta. The K-means Automatic Clustering algorithm is used for clusters based on the data attributes that have been obtained. Pre-processing experiments were carried out among others. Cleansing, Case folding, Tokenizing, Filtering, Stemming. Then look for the variance cluster to find the global optimum as an ideal cluster by identifying the moving variance by placing λ as the threshold (Global Optimum). So that the ideal cluster value is 0.332975. That is, the closer the cluster value obtained to number 1, the more the cluster search finds the optimum point. This research can be utilized in exploring and evaluating high-dimensional data, so that it becomes a consideration in providing approximate patterns from unstructured data sets with Visualization.