Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Explainability of Highly Associated Fuzzy Churn Patterns in Binary Classification

D. Y. C. Wang, Lars Arne Jordanger, Jerry Chun-Wei Lin

TL;DR

This study emphasises not only the explainability of customer churn through machine learning models, but also the importance of identifying multivariate patterns and setting soft bounds for intuitive interpretation, to identify highly associated patterns of customer churn with intuitive identification, referred to as Highly Associated Fuzzy Churn Patterns (HAFCP).

Abstract

Customer churn, particularly in the telecommunications sector, influences both costs and profits. As the explainability of models becomes increasingly important, this study emphasizes not only the explainability of customer churn through machine learning models, but also the importance of identifying multivariate patterns and setting soft bounds for intuitive interpretation. The main objective is to use a machine learning model and fuzzy-set theory with top-\textit{k} HUIM to identify highly associated patterns of customer churn with intuitive identification, referred to as Highly Associated Fuzzy Churn Patterns (HAFCP). Moreover, this method aids in uncovering association rules among multiple features across low, medium, and high distributions. Such discoveries are instrumental in enhancing the explainability of findings. Experiments show that when the top-5 HAFCPs are included in five datasets, a mixture of performance results is observed, with some showing notable improvements. It becomes clear that high importance features enhance explanatory power through their distribution and patterns associated with other features. As a result, the study introduces an innovative approach that improves the explainability and effectiveness of customer churn prediction models.

Explainability of Highly Associated Fuzzy Churn Patterns in Binary Classification

TL;DR

This study emphasises not only the explainability of customer churn through machine learning models, but also the importance of identifying multivariate patterns and setting soft bounds for intuitive interpretation, to identify highly associated patterns of customer churn with intuitive identification, referred to as Highly Associated Fuzzy Churn Patterns (HAFCP).

Abstract

Customer churn, particularly in the telecommunications sector, influences both costs and profits. As the explainability of models becomes increasingly important, this study emphasizes not only the explainability of customer churn through machine learning models, but also the importance of identifying multivariate patterns and setting soft bounds for intuitive interpretation. The main objective is to use a machine learning model and fuzzy-set theory with top-\textit{k} HUIM to identify highly associated patterns of customer churn with intuitive identification, referred to as Highly Associated Fuzzy Churn Patterns (HAFCP). Moreover, this method aids in uncovering association rules among multiple features across low, medium, and high distributions. Such discoveries are instrumental in enhancing the explainability of findings. Experiments show that when the top-5 HAFCPs are included in five datasets, a mixture of performance results is observed, with some showing notable improvements. It becomes clear that high importance features enhance explanatory power through their distribution and patterns associated with other features. As a result, the study introduces an innovative approach that improves the explainability and effectiveness of customer churn prediction models.

Paper Structure

This paper contains 15 sections, 3 equations, 4 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Explainability of SHAP Value
  4. Fuzzy-Set Theory and Fuzzy Pattern Mining
  5. Top-K High Utility Itemset Mining
  6. The Developed HAFCP Model
  7. Problem Statement
  8. Analytical Framework for HAFCP Mining
  9. A. Feature Importance Calculation
  10. B. Data Transformation
  11. C. Top-K High Utility Itemset Mining
  12. D. Pattern Evaluation and Model Enhancement
  13. E. A Simple Example
  14. Case Study and Discussion
  15. Conclusion

Figures (4)

  • Figure 1: An illustrative example of highly associated fuzzy churn patterns. The figure simply illustrates the concept of HAFCP. In the diagram, 'L', 'M', and 'H' represent low, medium, and high respectively, which are calculated by the membership function in fuzzy-set theory.
  • Figure 2: The proposed framework for mining HAFCPs: This comprehensive framework outlines the process of extracting HAFCPs, detailing the methodology from initiation to conclusion. To facilitate understanding, the extension work will present a straightforward algorithmic example. This example will elaborate each step of the process, providing a clear explanation of how HAFCPs are identified and derived within our framework.
  • Figure 3: An example of Gaussian membership functions representing linguistic terms (Low, Medium, and High) for Age and Spending Variables in fuzzy-set theory
  • Figure 4: Top-10 features of explainability in five datasets

Theorems & Definitions (1)

  • Definition 1