APLIKASI PERINGKAS OTOMATIS UNTUK MULTI DOCUMENT INDONESIA DENGAN WEIGHT OF FEATURES

Robertus Budihalim, Jeanny Pragantha, Viny Christanti

Abstract


Automatic Text Summarization is a computer-based application that produces text that is shorter than the original text, but still keep the important sentences of the text. This application is designed to perform summarization for Indonesian sociology documents. The purpose of this application is to design a system which can classify documents according to the contents similarities with Complete Linkage Clustering method, to design a system to generate a summary of some sociology documents with Weight Of Features method, helping the reader to get the important sentences of some of the documents, and shorten the time in obtaining the important sentences of several documents. This application was tested by Black Box Testing method, Purity and Question & Answer Evaluation. Based on the testing that has been conducted shows that all of the functions contained in the application program has been going well. The system has been designed to automatically classify documents with the best threshold value on 0.00005 with a 75 804% value of accuracy, and the system can generate a summary of some sociology documents with compression value of 25%, 50% and 75%. The greater value chosen, compression results summary will be more complete. The accuracy of the summary value of compression 25%, 50% and 75% respectively are 50%, 61%, and 81%.

Key words Automatic Text Summarization, Complete Linkage Clustering, Multi Document, Weight of Features


Refbacks

  • There are currently no refbacks.