Interest point type identification method and device, computer equipment and storage medium

Abstract

The invention provides an interest point type identification method and device, computer equipment and a storage medium. The method comprises the steps that an interest point text is obtained; the interest point text is processed to obtain a corresponding word sequence, and the word sequence comprises a plurality of segmented words in the interest point text and a plurality of pieces of position information corresponding to the segmented words; a text representation vector corresponding to the word sequence is generated, the text representation vector is composed of a plurality of segmented word representation vectors, and the segmented word representation vectors are vector representations of segmented words; a position representation vector corresponding to the word sequence is generated according to the multiple pieces of position information; and a target type of the point-of-interest text is determined according to the text representation vector and the position representation vector. According to the method and the device, the accuracy and the recognition efficiency of the POI type recognition can be effectively improved, and the POI type recognition effect is effectively improved.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, and in particular to a method, device, computer equipment, and storage medium for identifying types of points of interest. Background technique [0002] In some applications, it will involve point of interest (POI) to add type tags. POI can be, for example, a community, a store, an office building, a shopping mall, an organization company, etc., for example, 'also known as a science and technology innovation community' , 'A Building', 'B Plaza', 'C Mathematics'. The so-called type label is a summary of POI attributes in the category dimension. For example, the type label of "D Group" is a conventional enterprise, while the type label of "A Building" where "C Digital Technology" is located is an office building. [0003] The POI type identification method in the related art is usually based on manual annotation, or based on the text classification model of the existing "...

AI Technical Summary

Problems solved by technology

[0004] In these ways, the cost of manual labeling is high and the efficiency is low, while the text classification model based on the "Chinese thesaurus" word segmentation, general thesaurus and word structure have fewer feature dimensions and lack words when representing the features of POI texts. Or the position information of the phrase in the POI, which leads to the low efficiency and accuracy of the type recognition of the POI, and the poor effect of the type recognition of the POI

Images