Search relevant retrieval on indonesian translation hadith document using query expansion and smoothing probabilistic model

Hadith are words, deeds, decrees and approvals of the Prophet Muhammad SAW which are used as the basis for Islamic Shari'a law after the Qur'an. Currently there are many websites that provide information about hadith to facilitate users in the process of hadith learning or what we usually know as Information Retrievel (IR), such as the Lidwa Pustaka website. Basically, IR provides a search box for users to enter queries that reflect the user's information needs. The hadith search process on Lidwa Pustaka uses exact string matching method, which in the process of searching the hadith to the user's query must be the same as the hadith document in order per word (term), so that for partial matching search (matching the query in each word without sequential) can't be done yet. In addition, the writing of synonyms or variant strings that differ in Indonesian hadith translations, such as (al-khomru) are written as "khamar", "khamer", "khamr" or "minuman keras", making the process of the hadith search less precise. Therefore, this study aims to improve the search system for the hadith, using the approach of query expansion and smoothing probability models, namely Jelinerk-Mercer Smoothing, Dirichlet Smoothing and Absolute Discounting Smoothing. The use of query expansion and smoothing probability models in this study resulted in Mean Average Precision ALL (MAP ALL) values in all hadith documents, Mean Average Precision@30 and the highest recall value of 30 compared to other methods, such as exact matching method for Lidwa Pustaka, Latent Semantic Indexing, Probability Model and Cosine Similarity.