| With the continuous development and growth of e-commerce sales model,as an indispensable part of the online sales platform,how to accurately and effectively classify the commodity title text data,help merchants accurately locate the commodity category,and provide convenience for consumers to quickly and accurately find the required commodity information in a large number of commodities has become a practical research topic.At present,there is little research on the classification of commodity title text data in the field of text classification,and the existing models do not aim at the characteristics of commodity title text in online sales platform.Therefore,according to the characteristics of commodity title text,this paper proposes a commodity title text classification model BiTCN based on temporal convolutional network.By using the flexible receptive field of the Causal Dilated Convolution,and using a bi-directional way,we can jointly obtain the deep semantic features of the commodity title text from both positive and negative directions,so as to further capture the text context information.The main work of this paper can be summarized as follows:（1）In view of the shortcomings of traditional machine learning in classifying commodity title text,this paper uses the flexible receptive field of TCN in deep learning and the advantage of capturing long-distance dependence,proposes a commodity title text classification model BiTCN based on time convolution network,which fully captures the deep semantic features and context dependence of text,and realizes the task of commodity title text classification.（2）In order to test the classification performance of the BiTCN model classifier based on time convolution network,the paper sets up a comparative experiment,compares the BiTCN model with the traditional deep learning models TextCNN and FastText respectively,and selects the data sets of various commodity Title texts for model training.According to the experimental results,the performance of BiTCN model classifier is significantly improved,which is better than that of other classification models in commodity classification task to a certain extent. |
