What Is Concept Mining?

Opinion Mining, also known as text opinion mining or sentiment analysis, is the mining and analysis of sentiment information such as the subject, opinion holder, subjectivity, and emotional attitude of text information to identify the emotion of subjective text. tend.

Opinion mining, also known as text opinion mining or sentiment analysis, is the mining and analysis of emotional information such as the subject, opinion holder, subjectivity, and emotional attitude of text information, and then identifies the emotional trend of subjective text. Its research object is mainly text on the Web, especially comment text published by users. According to the different granularity of the processed text, opinion mining can be divided into three levels of research: words, sentences and texts. According to the different categories of processed text, it can be divided into opinion mining based on product reviews and news reviews. Different tasks can be mainly divided into three progressive tasks, namely emotional element extraction, emotional classification and opinion retrieval. ^[1]

Emotional information extraction, as a basic task of opinion mining, is a fine-grained analysis technique designed to extract valuable emotional information in texts for further perspective mining research later. Kim and Hovy ^[2]

Emotion classification includes subjective text recognition and sentiment classification of subjective text. General texts are used to describe objective facts and to express subjective views. Therefore, the task of text sentiment classification is generally subjective text recognition, that is, subjective and objective text classification, and then subjective text sentiment classification.

Opinion retrieval, also known as opinion retrieval, is designed to find text that contains a user's opinion, opinion, or opinion based on a query subject. Opinion retrieval is different from the traditional information retrieval technology oriented to objective factual text. Through opinion retrieval, users can search for people's opinions on this topic by entering the keyword of the topic.

Generally speaking, the retrieval of opinion information requires that the documents retrieved must meet two basic criteria at the same time: first, the relevance of the subject, that is, the retrieved documents must be highly relevant to the subject (query term); The document clearly reflects the user's perspective.