Search Engine Error Correction Algorithm And Error Correction Bad Case Mining

Search engine error correction function is of vital importance to improve search efficiency, and a good error correction function can provide better human-computer interaction for the user experience. According to the features of the Chinese language itself, an error correction function based on the establishment of N-gram Statistics Model is studied in this paper. Then the N-gram Statistics Model is smooth optimized. To further improve the search engine error correction function and raise the precision rate of error correction results, an efficient machine mining analysis method is presented here. The method analyses the user click log and some characteristic properties of the error term to mine the error correction Bad Case. According to statistical theory, statistical analysis of user click behavior on error correction term situation, quantitative modeling determine whether the system gives the wrong error correction term; statistical analysis the attribute relationships between previous word and next word entered by the user to determine whether the word entered by the user before the system is not corrected by the system. By smoothing optimization model N-gram statistical language as well as Error Correction Bad Case mining, we can further improve the search engine’s automatic error correction function, and improve the accuracy of automatic error correction results rate. It is verified by experiment that eventually smoothed optimized N-gram statistical language models and error correction Bad Case mining have a nice effect for the search engine. A nice auto correction function of Chinese input keywords which significantly improves the precision and recall rate of search engine is achieved.