Using Intuitionistic Fuzzy Set to Classify Uncertain and Linearly Non-Separable Data

Abstract

The problem of non-linearly separable data points requires more efforts to classify the data sample with high accuracy. This paper proposes a new classification approach that employs intuitionistic fuzzy sets to accurately classify non-separable datasets and to efficiently deal with uncertain labelled datasets. The dataset used contains 124 students with 9 features and 1 class for each student. First, the dataset is normalized to train and test the proposed approach. Second, the intuitionistic fuzzy sets were constructed using three features and the fuzzy model was created by calculating the equation of the straight line passing through the intuitionistic fuzzy sets of dataset classes. Finally, the classification is performed by calculating the distance between each class and the unseen sample that is subject to classification. Experimental results show that the classification performance of the proposed approach is competitive and superior to that of other state-of-the-art algorithms on the aforementioned dataset.