Customer Behavior Analysis by Intuitionistic Fuzzy Segmentation: Comparison of Two Major Cities in Turkey

The vast quantity of customer data and its ubiquity, as well as the inabilities of conventional segmentation tools, have diverted researchers in search of powerful segmentation techniques for generating managerially meaningful information. Due to its noteworthy practical use, soft computing-based techniques, especially fuzzy clustering, can be considered one of those contemporary approaches. Although there have been various fuzzy-based clustering applications in segmentation, intuitionistic fuzzy sets that have the complimentary feature have appeared in limited studies, especially in a comparative context. Therefore, this study extends the current body of the pertaining literature by providing a comparative assessment of intuitionistic fuzzy clustering. The comparison was carried out with two other well-known segmentation techniques, [Formula: see text]-means and fuzzy [Formula: see text]-means, based on transaction data that belong to Turkey’s two major cities. Over 10,000 records of customers’ data were processed for segmentation purposes, and the comparative approaches were presented. According to the results, the intuitionistic fuzzy clustering approach outperformed the other methods in terms of the clustering efficiency index being utilized. The validity of the segmentation structure obtained by the superior approach was ensured via nonsegmentation variables. The comparative assessment and the potential managerial implications could be considered as a contribution to the corresponding literature. This study also compares the effects of the different parameter values used in the proposed model.

Type2 IFC with SOA for Topic Detection and Document Clustering Analysis

The automatic document clustering and topic extraction from the corpus provides a very essential requirement in many real time applications. The document clustering and topic detection is utilized to locating data quickly. Hence, in this paper, Type 2 Intuitionistic Fuzzy Clustering and Seagull Optimization Algorithm (Type 2 IFCSOA) is developed for document clustering and topic detection. The Type 2 IFCSOA is utilized to cluster the documents. Additionally, ensemble approach is utilized to identify by the topics from the clustered documents. In the proposed methodology, the pre-processing is utilized to remove unwanted information from the documents such as tokenization, stop word removal and stemming process. After that, the proposed method is utilized to cluster the documents. The clustered documents are labeled with the basis of clusters. After that, to achieve topic detection, the ensemble approach is utilized with feature extraction phases such as Term Frequency- Inverse Document Frequency (TF-IDF), Mutual information (MI), Text Rank Algorithm and analysis of keyword taking out from co-occurrence statistical -Information (CSI). The proposed methodology is implemented in MATLAB and performances were evaluated with the statistical measurements such as precision, recall, accuracy, sensitivity, purity measure and entropy. The proposed method is compared with the conventional methods such as Fuzzy C Means clustering (FCM), FCM-Particle Swarm Optimization (PSO), FCM-Genetic Algorithm (GA) and K means clustering.

Superpixels extraction by an Intuitionistic fuzzy clustering algorithm

A scheme to develop the image over-segmentation task is introduced in this paper, it considers the pixels of an image as intuitive fuzzy sets and develops an intuitionistic clustering process of them. In this regard, the main contribution is to provide a method for extracting superpixels with greater adherence to the edges of the regions. Experimental tests were developed considering biomedical grayscale and natural color images. The robustness and effectiveness of this proposal was verified by quantitative and qualitative results.