A Three Layer Chinese Sentiment Polarity Detection Framework with Case Study

In terms of Chinese sentiment analysis or polarity detection, despite numerous proposed methods, they often exhibit significant language or knowledge dependencies, limiting their generality and scalability in real engineering projects. To address this issue, this paper introduces a Three Layer Chinese Sentiment Polarity Detection Framework (TLCSPDF). This framework decouples the final polarity detection from language processing by dividing the task into three interactive layers. This modular approach allows for the incorporation of various modules at different layers, facilitating reusability and configuration to meet specific project requirements. Additionally, this paper presents a specialized TLCSPDF designed for polarity detection in Chinese product comments, along with an in-depth exploration of its inner modules.