Search word optimizing method based on graph data structure

Abstract

The invention discloses a search word optimizing method based on a graph data structure. Multiple word sets and the relations among the word sets are extracted from the rules of graphs; the word sets and relations are settled; every word set is named; the word sets and the relations are simplified into multi-line And / Or expressions; every expression is analyzed; a weight value is endowed to every word count set; the parts of speeches are identified through word segmentation and inverse document rates; the association degrees of the word sets and themes are accurately analyzed; the minimum search word sets with the highest demand association degrees can be rapidly extracted from tens of thousands of rules. According the optimizing method, the relatively high recall rate is obtained in the indexing process; the logic expressions are covered most completely; the word count sets are minimum; the finally generated collection task amount is reduced; and the production efficiency of an enterprise is improved.

Description

Technical field [0001] The invention relates to the technical field of Internet information collection, in particular to a search term optimization method based on graph data structure. Background technique [0002] With the development of Internet technology, all kinds of data are vast, such as news, microblogs, forums, e-commerce, etc. Some customers are concerned about the dynamics of a certain event, and some customers are concerned about a certain brand. Word of mouth, some customers are concerned about the reputation of a certain company, how to accurately and quickly obtain the data that customers really care about from the Internet, you need to filter the data through the search engine, but whether the search term is appropriate or not will directly affect the search effect. Too many useless words in the search terms will reduce the search effect, and there will be less relevant data or even zero results. If the search terms are too limited, the amount of data retrieved ...

AI Technical Summary

Problems solved by technology

[0002] With the development of Internet technology, all kinds of data are vast, such as news, Weibo, forums, e-commerce, etc. Some customers pay attention to the dynamics of a certain event, and some customers pay attention to the news of a certain brand. Word of mouth, some customers are concerned about the reputation of a certain company. How to accurately and quickly obtain the data that customers really care about from the Internet requires the help of search engines to filter and filter the data. However, whether the search terms are appropriate or not will directly affect the search. Effect

If there are too many useless words in the search terms, the search effect will be reduced, and there will be less relevant data or even zero results. If the search terms are too limited, the amount of retrieved data will be too large, and further data screening is required to expand the Retrieval workload