Knowledge graph-based search input system

Abstract

The invention relates to a knowledge graph-based search input system, which comprises a keyword query input module for acquiring keywords input by a user; the knowledge graph generation module comprises an automatic generation unit and is used for automatically generating a knowledge graph according to keywords input by a user; the interaction module comprises a knowledge graph search request interaction unit and is used for visually displaying the current knowledge graph search request; and the search request correction module is used for correcting the current knowledge graph search request. Compared with the prior art, the method has the advantages of being capable of accurately expressing and depicting the search willingness of the user, good in compatibility and the like.

Description

technical field [0001] The invention relates to the technical field of online search, in particular to a search input system based on a knowledge graph. Background technique [0002] Online search is an important bridge connecting users and the Internet. Users frequently obtain massive information resources through online searches every day. However, in the online search process, a large number of users often encounter the problem of inaccurate searches. The reason is that on the one hand, the search needs are complex and changeable, and it is difficult for users to accurately express and describe the search needs; on the other hand, it is difficult to accurately match the search results with the search requests. In terms of expression and input of search needs, users currently use keywords, natural language sentences, voice, pictures, etc. to express and input search requests; in some specific fields, users use gestures and brain waves to input search needs. Looking at th...

AI Technical Summary

Problems solved by technology

Although these methods can clarify search requirements to a certain extent and help improve search satisfaction and accuracy, but because search keywords are grammatical and weak semantic representations, they cannot deeply express some important, relevant, potential, implication, and rich search results. As a result, it is still difficult to truly and accurately characterize and describe the search needs for search keywords, making it difficult to further improve search accuracy

In the study of using natural language sentences as search query expressions, it is very difficult for machines to accurately understand search semantics, resulting in slow progress in research in this field

In the research of using speech as the expression of search requests, on the one hand, the understanding of speech will eventually be converted into the understanding of keywords or natural language sentences, and its advantages and disadvantages are similar to those of keywords and natural language sentences as search requests. ; On the other hand, some literature pointed out that voice search can answer less than half of the search questions, and the search results vary greatly, so the accuracy of the search cannot be guaranteed

In the research of using pictures as the expression of search requests, although the existing picture search can already guarantee a certain search accuracy, most of the Internet information is stored and disseminated in the form of text, and it is difficult to describe and describe comprehensively and universally through pictures. Express user search requests

Images