Social Group Discovery Method for Uncertain Action Semantics in Multi-Scale Space

Abstract

The invention relates to a social group discovery method with uncertain behavioral semantics in multi-scale space, belonging to the field of data mining and knowledge discovery. The invention is based on the behavior trajectory of the user social network Twitter, and publishes tweets based on the similarity of the geographical location and the tweets The semantic similarity of the uncertain activities expressed by the entries is used to discover whether users have similar behavioral relationships, so as to find corresponding user groups with similar behaviors. Experiments have proven that the present invention is better than the current invention in terms of accuracy in discovering similar behavioral groups of users. The existing judgment methods have high practical application value. If they can be greatly promoted, they will definitely help industrial innovation, promote cross-border integration, benefit society and people's livelihood, and promote the innovative development of our country's economy and society.

Description

technical field [0001] The invention belongs to the fields of data mining and knowledge discovery, and in particular relates to a social group discovery method with uncertain behavior semantics in multi-scale space. Background technique [0002] With the rapid popularization of social network applications, more and more users are integrated into social networks. Typical applications include domestic Sina Weibo, foreign Twitter (Twitter), etc. These social applications allow users to share their latest news And ideas are posted to mobile phones or websites in the form of text messages, and if users wish, they can also post Weibo-tweeting physical location information. Weibo-Tweet content is short, but it contains certain semantics, which can be used to deduce user behavior to a certain extent; and allowing the disclosure of physical location information can facilitate tracking the latest updates of users. If the above two aspects can be effectively By using it, we can better...

AI Technical Summary

Problems solved by technology

[0003] Regrettably, so far, in the research of behavioral semantics, almost all research results believe that behavioral semantics is deterministic, but in fact, behavioral semantics itself often has a certain degree of uncertainty. When the "term set (a set of terms)" corresponding to the text expresses the "activity" semantics implied by the behavior, there is an uncertain semantic mapping relationship between the "term" and "activity". For example, a "word Articles" can belong to multiple "activities", and an "activity" can contain multiple entries. It is the uncertainty of this semantic mapping that affects the accuracy of user groups with similar behaviors to a certain extent. However, currently However, this issue has not attracted the attention of relevant personnel.

On the other hand, when using social data such as Weibo and Twitter to serve various applications, the influence of different geospatial scales on the clustering of social groups is not fully considered.

Images