Application of a Deep Learning Algorithm With Integrated Attention Mechanism in Sentiment Analysis of E-Commerce Product Reviews

with the rapid development of e-commerce, product reviews have become an important channel for consumers to express their opinions and share experiences, and also an important reference for consumers and enterprises to make decisions. Therefore, it is very important to analyze the emotional tendency of product reviews. The purpose of this study is to improve the predictive performance of sentiment analysis of product reviews through deep learning algorithm and attention mechanism. The main tasks of this research are as follows: 1) the literature related to sentiment analysis based on sentiment dictionary, machine learning and deep learning algorithm is studied, and the goal and core technology of this research are defined; 2) text preprocessing of product reviews, including data denoising, word segmentation and stopping word removal, is carried out to ensure data quality, and text vectorization technology is used to convert text into numerical features; 3) convolutional neural network (CNN) + Bi-directional Long Short-term Memory Neural Network (Bi-LSTM) model is constructed, and the Attention mechanism is integrated into it to form the CNN + Bi-LSTM + Attention model. By comparing the performance of the two models, it is found that the model with attention mechanism performs better in emotion classification task, with an accuracy of $94.13 \%$, which indicates that the attention mechanism can effectively enhance the ability of the model to capture key emotional information; 4) in order to further verify the superiority of the model, confusion matrix and ROC curve were used as evaluation tools in the study, and the results also proved that the model with the addition of attention mechanism performed better. The problem of data imbalance was encountered in the research process, but it was successfully solved by reasonable text preprocessing operations. This study not only validates the effectiveness of the deep learning model integrating attention mechanism in sentiment analysis of commodity reviews, but also provides new ideas and technical paths for solving specific problems in sentiment analysis.