Text Sentiment Analysis Model based on Feature Fusion of TextCNN and Transformer Networks

As a major social media platform for many users, Weibo hosts a vast amount of user-generated information, including opinions on events, products, and more. Therefore, sentiment analysis of Weibo comments is of significant practical importance. Traditional sentiment analysis methods face limitations such as inadequate feature representation, inability to capture deep meanings, and low classification accuracy. This study aims to address these issues by combining Text Convolutional Neural Networks (TextCNN) and Transformer networks to perform sentiment analysis on a large volume of comment text data from the Weibo platform. The comments are first subjected to a series of preprocessing steps (including data cleaning, stop word removal, Jieba tokenization, etc.), followed by vectorization using the Word2vec method. The word vectors are then input into the model, and sentiment classification is achieved through a Softmax activation function. Results indicate that the proposed model effectively improves text classification accuracy, thereby validating the effectiveness of both TextCNN and Transformer models.