The Research And Implementation Of Query Suggestion Mechanism In The Absence Of Logs For Specific Search Engine

Search engine queries suggestion for users to make better use of the search engineprovides a great convenience. Most current commercial search engine query suggestionmethod is based on the use of the log analysis. Specific search engine is usually applied tospecific areas, such as paper search engine, campus search engines, forums search enginesand so on., and the number of users is relatively small in terms of general search engines. Sospecific search engines cannot use the traditional suggestion method based on query loganalysis. The traditional query suggestion method also cannot reflect the theme ofprofessionalism in specific search engine. Therefore, this paper presents a method of querysuggestion in specific search engine, to make up for these shortcomings.Research work is mainly in the following three aspects1. Input query term suggestion.This paper proposes two suggestion methods. In the firstmethod, we use Bayesian formula modeled by calculating the probability of Input query termto suggest terms for users; In the second method, we use topic model, and use LDA(LatentDirichlet Allocation) algorithm to preprocess local corpus. The similarity of the input query iscalculated using topic model, and then the system will suggest words belong to differenttopic.2. Related query suggestion.This paper presents an improved modularity algorithm, thealgorithm uses the results of the pseudo relevance feedback to the user query using modularitycommunity mining algorithms, to identify the fuzzy semantic query in different communities.Finally, the algorithm suggests different semantic fuzzy queries to the user.3. Design and implementation of a query Suggestion system. The whole system isdivided into the presentation layer, access layer, model layer, data layer.In this paper, the query suggestion system has been applied to the education network in adistributed search engine SE6. The search engine for China Education and Research Network,provides education and scientific research and other related fields retrieval services.