Text key information extraction method and device and medium

Abstract

The embodiment of the invention discloses a text key information extraction method and device and a medium. The extraction method comprises the following steps: obtaining a to-be-extracted text, wherein the to-be-extracted text comprises a title and a main body; generating a first list, wherein the first list comprises at least one candidate key unit, and the candidate key units are composition units extracted from the main body by using the similarity weight of the composition units in the main body; selecting a title similarity unit from the main body, the title similarity unit being a composition unit having the highest similarity with the title; if the title similarity unit is different from any candidate key unit, adding the title similarity unit into the first list; and generating key information by using the first list. By means of the extraction method, the situation that the extracted key information is missed can be avoided, the accuracy of the extracted key information is improved, and meanwhile the extracted key information is relatively simpler and more comprehensive.

Description

technical field [0001] The invention relates to the fields of information extraction and text mining, in particular to a method, device and medium for extracting key text information. Background technique [0002] With the continuous development of information technology, massive data has become the most valuable wealth. How to quickly and accurately grasp information and make accurate and reasonable decisions has become the only way for enterprises to survive and develop. This requires mining effective key information from massive texts, such as text summaries, keywords, etc. [0003] Many texts, such as news texts on the Internet, are unstructured texts. To dig out effective and structured key information from these unstructured texts, the TextRank algorithm can be used. The TextRank algorithm is a graph-based ranking algorithm for text. The basic idea is to divide the text into several components (such as sentences) and establish a graph model, use the voting mechanis...

AI Technical Summary

Problems solved by technology

[0004] The TextRank algorithm is used to extract key information from the text. Due to the limitations of the TextRank algorithm itself, the extracted key information is prone to high similarity or repeated content.

For example, the TextRank algorithm is used to extract 5 sentences from a text A including 100 sentences to form a text summary of text A, but there may be 3 sentences with high similarity in the text summary, and these 3 sentences express The semantic information output is very similar, resulting in the text summary is not concise enough

Images