Predicting Fitness Centre Dropout.

Pedro Sobreiro,Celina Gonçalves,Pedro Guedes-Carvalho,Paulo Pinheiro,Abel Santos

doi:10.3390/ijerph181910465

Abstract

The phenomenon of dropout is often found among customers of sports services. In this study we intend to evaluate the performance of machine learning algorithms in predicting dropout using available data about their historic use of facilities. The data relating to a sample of 5209 members was taken from a Portuguese fitness centre and included the variables registration data, payments and frequency, age, sex, non-attendance days, amount billed, average weekly visits, total number of visits, visits hired per week, number of registration renewals, number of members referrals, total monthly registrations, and total member enrolment time, which may be indicative of members’ commitment. Whilst the Gradient Boosting Classifier had the best performance in predicting dropout (sensitivity = 0.986), the Random Forest Classifier was the best at predicting non-dropout (specificity = 0.790); the overall performance of the Gradient Boosting Classifier was superior to the Random Forest Classifier (accuracy 0.955 against 0.920). The most relevant variables predicting dropout were “non-attendance days”, “total length of stay”, and “total amount billed”. The use of decision trees provides information that can be readily acted upon to identify member profiles of those at risk of dropout, giving also guidelines for measures and policies to reduce it.

Highlights

Dropout has always been an ongoing concern for fitness centre managers
In this study we investigated and compared the performance of several machine learning classification algorithms: Logistic Regression (LR), Decision Tree Classifier (DTC), Random Forest Classifier (RFC), and Gradient Boosting Classifier (GBC)
The hyperparameter optimisation using grid search on the training data achieved an area under the curve (AUC) score for LR of 0.845, DTC 0.898, RFC 0.947, and GBC 0.965

Summary

Introduction

Dropout has always been an ongoing concern for fitness centre managers. According to the International Health, Racquet and Sports Club Association [1], six causes of dropout have been identified: the high number of members in facilities; dissatisfaction with employees; a lack of interest shown by staff; disappointment with programmes and activities provided, and the inaccessibility or lack of response from individuals in charge.Previous studies have already addressed the difficulties faced by fitness centres in retaining members, in an attempt to reduce high rates of dropout [2,3].The fitness sector is well known for its high incidence of dropout [4], and Portugal is no exception where rates of 65% were experienced in 2018 [5]. Dropout has always been an ongoing concern for fitness centre managers. Previous studies have already addressed the difficulties faced by fitness centres in retaining members, in an attempt to reduce high rates of dropout [2,3]. García-Unanue, Iglesias-Soler, Felipe, and Gallardo [6] showed that only 30–60% of members continue attendance in the second year at fitness centres. Whilst these studies focused on the nature of the fitness services supplied, Sperandei, Vieira and Reis [7] have stated that the most important issue is to identify the profile of those individuals at the highest risk of dropout

Objectives

Methods

Results

Discussion

Conclusion