The establishment and application of Cantonese/Mandarin semantic relevant database for search engine

In the existing commercial search engines, the retrieval precision of Cantonese keyword is very low, which cannot meet users' demand. This paper improves machine translation techniques in Cross-Language Information Retrieval. It introduces Cantonese/Mandarin Semantic Relevant Database that's characteristic of Self-learning Artificial Intelligence and then puts forward a search engine model based on this database. Experimental results show that the search engine model based on Cantonese / Mandarin semantic relevant database is much more superior to the traditional search techniques in the Cantonese search. The search engine model greatly improves the retrieval precision of Cantonese keywords while not affecting the performance of existing search engines.