A label embedding method and device for cross-domain knowledge transfer

Abstract

The invention relates to a cross-domain knowledge transfer tag embedding method and apparatus. The method comprises the steps of obtaining text data of a source domain and a target domain, performing model representation, solving word vector parameters of keywords in the source domain and the target domain, and performing transfer of keyword tags from the source domain to the target domain; obtaining nearest neighbors of labeled keywords in the source domain and the target domain, performing weight assignment on keywords of the nearest neighbors by keyword tags of the labeled keywords to obtain extended keyword tags; performing user-level keyword tag labeling according to extracted user-level text data; dynamically optimizing parameters of user-level keyword tag parts according to click and / or access data information of a user based on the word vector parameters of the keywords and the user-level keyword tags; and obtaining new user-level text data from the target domain, performing user-level keyword tag labeling prediction and sorting, and outputting a result. According to the method and the apparatus, the accuracy and high efficiency of tag labeling can be taken into account and business demands of business personnel are met.

Description

technical field [0001] The invention relates to the application field of Bayesian network and text mining, in particular to a label embedding method and device for cross-domain knowledge transfer. Background technique [0002] In recent years, with the rapid development of big data technology, all walks of life pay more and more attention to the value of data, and the data sources and data structures accumulated by each company show a variety of characteristics, and the generation of text data is also increasing. , For example, various online media, e-commerce reviews, Weibo, online advertisements, etc. will generate a large amount of text data. It is very important for various enterprises to identify the interests of users by mining the information in these historical behavior data of users. Due to the high-dimensional and sparse feature expression of text data, and the complex semantics of Chinese, semantic analysis and classification of these texts has always been a major...

AI Technical Summary

Problems solved by technology

[0003] However, we found in the actual use and research process that the existing technology has at least the following problems: the existing technology is based on the cluster analysis method of unsupervised learning or relies on a large amount of labeled data, which cannot meet the individual needs of business personnel

In actual use, business personnel often design a top-down label system according to their own business development. In this case, the results output by unsupervised learning cluster analysis technology are often quite different from the business system itself; if Choose to use a supervisory method, such as text classification, to label the expected sentences according to the system designed by the business personnel, and generate sample data, while the labeling of sentences or articles is time-consuming and costly for text data as a whole, and the accuracy of labeling It also depends on the business experience of the personnel involved in the labeling

[0004] In short, the algorithms for labeling text data in the prior art cannot take into account the accuracy and high efficiency of labeling to meet the business needs of business personnel

Images