Text summary helps in understanding the content of a text without having to read the contents of the text as a whole. Automatic text summarization can be used to summarize the text easier. In this paper a frequent term based text summarization for Bahasa Indonesia is designed and implemented in java. The proposed system generates a summary for a given input document based on identification and extraction of important sentences in the document. The system counts nouns and verbs term frequency because they are considered as the most representative to the content of the text. The system also integrated to statistical approach with two underlying concepts such as title of the news article and location of the sentence. The generated summaries were compared with human generated summaries. Precision, recall and f-measure ratio are used to evaluate the accuracy of the generated summary. Assessment of the system summary result quality by respondents is also done by giving a value from 1 to 100. Based on the experimental results, the system is able to produce an effective summary with the average f-measure of 78%, at the compression rate of 30%. The average value of the quality of system summary result provided by respondents is 83,3
展开▼
