Intelligent Investigation Mechanism based on Fuzzy Markup Language for social media application
The experimental results on 2016 Taiwan President Election show that the proposed IIM with the ability of FML and machine learning is feasible to apply to social media domain.
Abstract
The Intelligent Investigation Mechanism (IIM) based on Fuzzy Markup Language (FML) and genetic learning ability for social media application is presented in this paper. We take the social media contents of presidential election domain on Facebook as an example in this paper, and construct the presidential election ontology for defining the knowledge base and rule base of FML. Seven categories, including Presidential Candidate, Municipality Mayor Supporter, Beneficial Event, Common Student Supporter, Famous Student Supporter, Common People Supporter, and Famous People Supporter, are adopted to present the president election domain ontology. With the defined ontology, we apply the FML to describe the fuzzy variables and fuzzy rules of the presidential election for IIM. The proposed IIM infers the possibility of Candidate Support Degree for each presidential candidate based on the fuzzy variables and the posts on Facebook (FB). The experimental results on 2016 Taiwan President Election show that the proposed IIM with the ability of FML and machine learning is feasible to apply to social media domain. In the future, we will compare our proposed approach with the existing ones and further analyze the experimental results with the actual data.