Implementation of Information Retrieval Using Tf-Idf Weighting Method On Detik.Com’s Website

Abstract

Information Retrieval is a process to find back the information that is needed by system. News is not only communicated via the print media, but also through online media. The rapid technology makes people more up to date to on news or current information. Detik.com is one of the online news website that serves a variety of the latest information. Based on the results of questionnaires taken from 30 respondents, the results obtained percentage of 100% which states that online news is important But in detik.com website visitors often get articles that are not in accordance with what is referred to, is evidenced by the results of the percentage is 66.7%. It is claimed that the keywords entered are not relevant to the search results. This research was conducted by applying a weighting method TF-IDF (Term Frequency Inverse Document Frequency). There are several preprocessing stages that conducted in the search for relevance weighting value starting from tokenizing process, Sitering process, stemming process followed by a TF-IDF weighting method. The weighting of the results obtained weight value relevance of each article from highest to lowest weight. This research resulted a web applications Information Retrieval on the site detik.com using TF-IDF weighting method. The test results showed recall value of 1 indicating that the relevant articles can be found by the system and the precision value of 0:50 indicates there are relevant articles that are not found in the system. Recall and precision resulted in a value of 1 if the query (keyword) which included having one term (word). Precision low value indicates that the average accuracy of the keywords entered by the article irrelevant search results.