Book Data Content Similarity Detector With Cosine Similarity (Case study on digilib.uad.ac.id)

Abstract

A library always tries to complete its book collection, yet visitors often find it difficult to search for an alternative book of an unfound book they search. Based on the questionnaire result, 83.3% respondents of library visitors take a long time in finding the exact data content of a book as an alternative of an unfound book. There are some books having similar title but the content is different. As a result, the visitors get it more difficult. For now digilib.uad.ac.id has not been able to give an alternative book when book searching done by visitors. The aim of this research is to ease the library visitors to find the exact reference book as an alternative book of an unfound book in the library by implementing Cosine Similarity method in detecting similarity.The research subject was the similarity detector system of the book data content in the library by using Cosine Similarity method. The data collection method was literature study, interview, and questionnaire. The system development phases included database design, interface design, flowchart design of book data content similarity and system evaluation with Blackbox Test, Alpha Test, and Accuracy Testing.This research developed a web-based application to detect book data content similarity by comparing the resume of a book that visitor searches with other books by using Cosine Similarity method based on the highest to the lowest score rank. System evaluation result with Blackbox test to librarians shows that the system runs well, the input can be received properly, and the output meets the expectation. On the Alpha Test evaluation, the system is also accepted by the users with 92.5% of respondent evaluation shows a highly positive response towards it. This application is reasonably used with the result of Accuracy Testing of system comparison of 82.14% relevant.