基于深度学习的商品评论情感分类研究

Purpose/significance] The existing text representation and classification algorithms are combined,and a combination mode of low complexity and less training time is selected to construct an optimizationmodel for the classification of emotional texts of commodity reviews. [Method/process] Firstly, this papertook the Keras API as an application environment, input Word2vec word vector into Embedding embeddedlayer. Then, based on sentence word index sequence, three kinds of commodity comment text representationwere realized by controlling the trainable parameter. Finallyin this paper, different text representationswere matched with different classification algorithms, differences in classification effects were analyzed, andthe better combination of algorithms was established. [Result/conclusion] The text representation methodwhich is continued training by Inputting Word2vec Word Vector into Embedding embedded Layer, combinedwith the TextCNN algorithm establishes the classification model. It performs better on the product reviewtest set. Its accuracy and ROC curve area AUC values are 94.02% and 0.9827, respectively. The applicationshows that the classification model can better realize the emotional classification of commodity reviews andhas better classification generalization ability.