A Novel string grammar relational fuzzy clustering

Abstract

The pattern recognition problems can be divided into two main groups that are syntactic approach and decision-theoretic approach. Normally an object data is represented in a feature vector form. It can be represented as a pairwise relation between them which called relational data. This thesis we focus only on the syntactic approach and using data in a relational form. We propose a new string grammar relational fuzzy clustering method to deal with relation syntactic data. To reveal the performance of the proposed method, we experiment on standard datasets and real-world datasets including a real-world dataset in a recommendation system.