A machine learning approach to predict classification of fans’ attitudes toward sponsors

Abstract

Purpose

This study aims to develop and validate an accurate machine learning model to categorize NBA fans into meaningful clusters based on their perceptions of sport sponsorship. Additionally, by predicting the intensity of NBA fans’ attitudes toward sponsors, the authors intend to identify the specific features that influence prediction, discuss these findings and offer implications for academics and practitioners in sport sponsorship.

Design/methodology/approach

This study used a sample of 1,142 NBA fans who were recruited through Amazon Mechanical Turk (MTurk). Fans identification, sponsorship fit, behavioral intentions, sponsor altruistic motive, sponsor normative motive, sponsor egoistic motive were surveyed as predictors, whereas fans’ attitudes toward sponsors was collected as the dependent variable. The LASSO regression, SVM, KNN, RF and XGboost were used to develop and validate the prediction model after verifying the measurement model by the Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Findings

The RF model had the best accurate in predicting the intensity of fans’ attitudes toward sponsors, achieving an AUC of 0.919 with a sensitivity of 0.872, a specificity of 0.828, a PPV of 0.873, a NPV of 0.828 and an accuracy of 0.848. The most influential feature in the model was “the fit of 0.301”. “Fans’ perceptions of sponsor’s normative motive”, “behavioral intentions supporting sponsors”, “fans’ identification with their favorite team”, “fans’ perceptions of sponsor’s altruistic motive” and “fans’ perceptions of sponsor’s egoistic motive” were exhibited in descending order.

Originality/value

This study is the first in sport sponsorship to accurately classify the intensity of fans’ attitudes toward sponsors as either high or low using machine learning models, and to formulate how fans’ attitudes formed toward sponsors from their perceptions of sponsorship process.