Abstract
This study examines how to look for a verse of the Quran or a clipped verse in the Qur'an and ranks search results correctly. Because the Qur'an consists of 30 juz, 114 letters and 6236 verses, then if you search for a particular verse then a Quranic verse search system is needed. At present, there is already a verse search system for the Quran, but the search results ranking process still has a number of errors in highlighting the verses that are searched for, the similarity scores that are less precise and have not been good at handling clipped or incomplete verses.
When conducting a search that needs to remember a verse from the Koran, one can remember the whole verse of the Quran but there are also some cases that cannot remember a verse completely or cut off.
For example, a complete verse search query
< ??????? ??????????? ??? ?????? ? ????? >
zalikal kitabu la raiba fihi the result will be different from the clipped verse query like this
< ??????? ???????????? ????? >
zalikal kitabu fihi. Because some truncated characters will be ignored by the system and can reduce the value of similarity. This shows that the existing system does not provide good accuracy in string matching. With this research, it will overcome the LCS search problem which ignores previously unreadable characters. Using the LCS method of query search for all data so as to produce candidate results, then research the neglected character of the candidate results to find similarities between neglected characters and one of the candidates previously obtained. So, improve ranking in the lafzi and increase the ability of the lafzi to search for clauses that are clipped or incomplete.
Keywords -- query search, weighting, ranking