Application of Intelligent Information Processing Based on ALBert-TextCNN in Sentiment Analysis of Catering Delivery Reviews

With the rapid development of information and communication technologies, sentiment analysis of unstructured food delivery reviews plays a crucial role in intelligent decision-making and consumption ecosystem building. This paper proposes an intelligent computing method that integrates ALBert and TextCNN and constructs a sentiment analysis framework. Leveraging ALBert's lightweight bidirectional encoding to extract deep semantics and TextCNN's ability to capture local sentiment features, the framework improves classification accuracy and robustness. Experimental results show that the new model achieves an F1 of 89.97%, an AUC of 90.16%, and a loss of 25.43%, verifying the engineering value of the technical solution and offering a versatile approach for user feedback processing and promoting interdisciplinary innovation.