Deep learning method of attribute emotion word vector

Abstract

The invention discloses a deep learning method of an attribute emotion word vector. The deep learning method comprises the following steps: a) using a label of 'common impression' in comments on e-commerce website products as an attribute emotion label to automatically mark a product comment data set, acquiring pre-training word vector dictionary through language model training; b) putting forward a deep learning model ASWV to infuse attribute emotion information contained in the comments into a word vector training process, obtaining an attribute emotion word vector; c) if iteration stop conditions are not satisfied, propagating back an attribute emotion error to update the word vector; and d) outputting the attribute emotion word vector. The deep learning method provided by the invention infuses attribute and emotion information of products into automatic learning of word vector characteristics by adopting a deep learning technology, automatically obtains attribute emotion word vector characteristics of the products without using a method for artificially designing characteristics in traditional natural language processing, avoids error propagation and superposition caused by stepwise strategy, and also overcomes a difficulty that emotion words can only have single emotion polarity constraint.

Description

technical field [0001] The present invention relates to sentiment analysis of Internet product reviews, to using deep learning technology to obtain product attribute emotion word vector expression, and to a method for integrating attribute and emotion information into traditional semantic word vector expression. Background technique [0002] With the rapid development of the Internet and e-commerce, online shopping has a huge impact on people's consumption patterns. Various e-commerce platforms, such as Taobao, JD.com, Yihaodian, etc., cover thousands of products. When purchasing a product without touching the actual product, the user's product review information has a high reference value, but it is time-consuming and laborious to browse thousands of reviews one by one or summarize through artificial rules. Applying natural language processing In-depth automatic mining of product reviews with machine learning methods, presenting a concise and intuitive "everyone's impressi...

AI Technical Summary

Problems solved by technology

This step-by-step strategy faces three main problems: (1) The number of product attributes must be determined in advance, but there are thousands of product types, and the attribute descriptions are ever-changing, so the migration of the attribute extraction system is poor; (2) The sentiment dictionary must be manually set in advance, but the polarity of some sentiment words is variable, and the fixed sentiment dictionary cannot correctly judge the correct polarity of the sentiment words based on different attribute contexts; (3) There are only positive and negative sentiments in traditional sentiment analysis Polarity, lack of fine-grained sentiment analysis for product attributes

Images