Embedding evaluation method and embeding evaluation system for social network blogger
The embodiment of the invention provides an embed evaluation method and system for social network bloggers, and the method comprises the steps: obtaining the feature information of all to-be-classified bloggers, carrying out the training of a plurality of embedding vectors, and generating a pluralit...
Saved in:
Main Authors | , |
---|---|
Format | Patent |
Language | Chinese English |
Published |
22.12.2020
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The embodiment of the invention provides an embed evaluation method and system for social network bloggers, and the method comprises the steps: obtaining the feature information of all to-be-classified bloggers, carrying out the training of a plurality of embedding vectors, and generating a plurality of embedding vectors for each to-be-classified blogger; for each embedding vector, taking the distances between the embedding vectors of the other bloggers and the embedding vector of the blogger in the set center of each domain as a clustering measurement standard, and obtaining a plurality of embedding vector clustering results, and according to the plurality of clustering results of each kind of embedding vector, forming an evaluation result of the kind of embedding vector, comparing the evaluation result, and judging the training quality of each kind of embedding vector. According to the method, blogger embedding vectors are trained in different modes, and the blogger is recommended byadopting the embedding v |
---|---|
Bibliography: | Application Number: CN202010873558 |