Cluster-Based News Representative Generation with Automatic Incremental Clustering

Nowadays, a large volume of news circulates around the Internet in one day, amounting to more than two thousand news. However, some of these news have the same topic and content, trapping readers among different sources of news that say similar things. This research proposes a new approach to provide a representative news automatically through the Automatic Incremental Clustering method. This method began with the Data Acquisition process, Keyword Extraction, and Metadata Aggregation to produce a news metadata matrix. The news metadata matrix consisted of types of word in the column and news section of each line. Furthermore, the news on the matrix were grouped by the Automatic Incremental Clustering method based on the number of word similarities that arised, calculated using the Euclidean Distance approach, and was done automatically and real-time. Each cluster (topic) determined one representing news as a Representative News based on the location of the news closest to the midpoint/centroid on the cluster. This study used 101 news as experimental data and produced 87 news clusters with 85.14% precision ratio.

Abstract licence: CC BY-NC-SA

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.