Yamli, the company that brought us products like the Smart Arabic Keyboard and Yamli Arabic Search, is releasing an enterprise version of the technology being used for Yamli Arabic Search in the form of a product they’re calling Arabic Query Expander.
The new service provides a solution for search and data normalization problems when it comes to the Arabic language, by analyzing and identifying all the English and Arabic script variations of an Arabic word, increasing the probability to find matches for searches.
Yamli gives examples like the name “Mohamed Mansour” having at least 80 ways of spelling it, “Mohamed Saleh Abdallah” having 7000 variations, “Jumeirah” having 51 spellings, and the song “Habibi Ya Nor El Ein” having 1207 spelling variations; to make the case for the utility of this tool and how much it can help in solving this problem.
With this release, companies with applications for banking, identity management, address management, media management and others can simply integrate Yamli Query Expander to enhance their search results. According to the company, the product plugs in and works directly with the existing database and search engine set up.
As mentioned, this tool used the same technology Yamli does, based on the same algorithms and collected intelligence, and covers the following features:
- Expansion of a query into all its Arabic and English variations. For example عبد الله , عبدالله, Aballah, Abdullah, Abdalla …etc. With the possibility of applying this to people/place names and media content, as well as any generic Arabic word.
- Frequency and ranking bias with each expanded result.
- Data cleansing and deduplication, to eliminiate duplicate entries and make sure the databases are in sync.
The product isn’t being featured on the Yamli site yet, but companies interested should be able to get in touch with the company to inquire more about it and get access to start integrating with their systems.