A Novel Method for a COVID-19 Classification of Countries Based on an Intelligent Fuzzy Fractal Approach.

Oscar Castillo,Patricia Melin

doi:10.3390/healthcare9020196

Abstract

We outline in this article a hybrid intelligent fuzzy fractal approach for classification of countries based on a mixture of fractal theoretical concepts and fuzzy logic mathematical constructs. The mathematical definition of the fractal dimension provides a way to estimate the complexity of the non-linear dynamic behavior exhibited by the time series of the countries. Fuzzy logic offers a way to represent and handle the inherent uncertainty of the classification problem. The hybrid intelligent approach is composed of a fuzzy system formed by a set of fuzzy rules that uses the fractal dimensions of the data as inputs and produce as a final output the classification of countries. The hybrid approach calculations are based on the COVID-19 data of confirmed and death cases. The main contribution is the proposed hybrid approach composed of the fractal dimension definition and fuzzy logic concepts for achieving an accurate classification of countries based on the complexity of the COVID-19 time series data. Publicly available datasets of 11 countries have been the basis to construct the fuzzy system and 15 different countries were considered in the validation of the proposed classification approach. Simulation results show that a classification accuracy over 93% can be achieved, which can be considered good for this complex problem.

Highlights

We describe in this article a novel hybrid approach for the classification of countries based on the COVID-19 data complexity
We have to say that we decided to use the specific time window mentioned above to validate the method, but the method can be used for other time windows and with the evolution of the pandemics the classification of a specific country can change depending on what is happening in the new time window
The method that uses the fractal dimension in the fuzzy system to build a classification of countries in the world is tested

Summary

Introduction

We describe in this article a novel hybrid approach for the classification of countries based on the COVID-19 data complexity. The proposed approach consists of a hybridization of fractal theoretical constructs and fuzzy logic concepts to achieve the goal of classifying the countries based on the complexity of their time series data. There is a wide variety of algorithms for calculating the fractal dimension producing a crisp value by using as data the time series for a particular problem. This value offers an estimation of the complexity of a specific time series. The numeric values of the fractal dimension are used as inputs to the set of fuzzy rules to perform the classification

Methods

Results

Conclusion