Unsupervised text sentiment analysis method based on quantum theory

The invention relates to an unsupervised text sentiment analysis method based on a quantum theory. The unsupervised text sentiment analysis method comprises the following steps: creating two sentiment dictionaries, namely a positive sentiment dictionary PSD and a negative sentiment dictionary NSD; preprocessing the positive sentiment dictionary PSD, the negative sentiment dictionary NSD and texts in the corpus set; constructing a quantum text representation model, extracting features from the preprocessed positive emotion dictionary PSD, the preprocessed negative emotion dictionary NSD and the preprocessed text, and constructing a positive emotion dictionary density matrix rho PSD, a negative emotion dictionary density matrix rho NSD and a text density matrix rho text; and applying a quantum relative entropy algorithm to obtain an emotion classification result of each text. 本发明涉及一种基于量子理论的无监督文本情感分析方法：包括如下步骤：创建两个情感词典，分别是积极情感词典PSD与消极情感词典NSD；对积极情感词典PSD、消极情感词典NSD以及语料集中的文本进行预处理；构建量子文本表示模型，分别对预处理后的积极情感词典PSD、消极情感词典NS