Penerapan Model OKAPI BM25 Pada Sistem Temu Kembali Informasi

Abstract

The rapid growth of information and digital documents has caused the process to retrieve desire documents becomes more difficult when performed manually. Therefore, an information retrieval system is needed as the search engine to find relevant documents as user desire. This research is to develop a information retrieval system based on OKAPI BM25 model. Two main steps would do in build an information retrieval system with Okapi BM25 model are preprocessing and applied the model to it. The preprocessing steps are collection of document (corpus), tokenization, linguistic preprocessing, and indexing by creating an inverted index. After preprocessing, applied Okapi BM25 model to the system to compute the relevant value between each document to user’s query. The system retrieve a list of documents which in descending sort based on relevant value. System performance is measured by calculating precision and recall. Experiment on several query key, shows a good performance where precision value about 74%-92% and 100% recall. This system is portable so it can be applied to the local server to meet the needs of the organization in finding a document in accordance with its documents collection.
Keywords : Information Retrieval, Okapi BM25, Precision, Recall.