Unsupervised Modelling of E-Customers’ Profiles: Multiple Correspondence Analysis with Hierarchical Clustering of Principal Components and Machine Learning Classifiers DOI Creative Commons
Vijoleta Vrhovac, Marko Orošnjak, Kristina Ristić

et al.

Mathematics, Journal Year: 2024, Volume and Issue: 12(23), P. 3794 - 3794

Published: Nov. 30, 2024

The rapid growth of e-commerce has transformed customer behaviors, demanding deeper insights into how demographic factors shape online user preferences. This study performed a threefold analysis to understand the impact these changes. Firstly, this investigated (e.g., age, gender, education) influence e-customer preferences in Serbia. From sample n = 906 respondents, conditional dependencies between demographics and were tested. hypothetical framework 24 tested hypotheses, successfully rejected 8/24 (with p < 0.05), suggesting high association with purchase frequency reasons for quitting purchase. However, although reported test statistics suggested an association, understanding interactions categories profiles was still required. Therefore, second part considers MCA-HCPC (Multiple Correspondence Analysis Hierarchical Clustering on Principal Components) identify profiles. revealed three main clusters: (1) young, female, unemployed e-customers driven mainly by reviews; (2) retirees older adults infrequent purchases, hesitant buy without experiencing product person; (3) employed, highly educated, male, middle-aged who prioritize fast accurate delivery over price. In third stage, clusters are used as labels Machine Learning (ML) classification tasks. Particularly, Gradient Boosting (GBM), Decision Tree (DT), k-Nearest Neighbors (kNN), Gaussian Naïve Bayes (GNB), Random Forest (RF), Support Vector (SVM) used. results that GBM, RF, SVM had performance identifying Lastly, after performing Permutation Feature Importance (PFI), findings work status, education, income determinants shaping developing marketing strategies.

Language: Английский

Optimal Service Strategies of Online Platform Based on Purchase Behavior DOI Open Access
Xudong Lin,

Tingyi Shi,

Hanyang Luo

et al.

Sustainability, Journal Year: 2024, Volume and Issue: 16(19), P. 8545 - 8545

Published: Sept. 30, 2024

In the rapidly evolving platform economy, online platforms have emerged as pivotal providers of digital services to sellers. The paper investigates how optimize service strategies based on consumers’ purchase behavior, influencing sellers’ pricing and social welfare. Using a two-period Hotelling model cooperative game framework, we discover that optimal with data collecting capabilities are collaborating two sellers offer extend new consumers in second period, maximizing profits for all platform. Applying Shapley value analysis, determine platform’s equitable charge strategies. When adopt behavior-based (BBP), escalates first also enhance However, BBP intensifies competition, leading generally lower pricing. Our findings suggest period should increase enhanced quality perception, which is provided by heightened privacy concerns, while decreasing regular consumers. Lastly, policy recommendations, exploring regulatory scenarios—limiting or not limiting collection—to maximize welfare consumer surplus, Mathematica software used identify distinct intervals.

Language: Английский

Citations

0
Unsupervised Modelling of E-Customers’ Profiles: Multiple Correspondence Analysis with Hierarchical Clustering of Principal Components and Machine Learning Classifiers DOI Creative Commons
Vijoleta Vrhovac, Marko Orošnjak, Kristina Ristić

et al.

Mathematics, Journal Year: 2024, Volume and Issue: 12(23), P. 3794 - 3794

Published: Nov. 30, 2024

The rapid growth of e-commerce has transformed customer behaviors, demanding deeper insights into how demographic factors shape online user preferences. This study performed a threefold analysis to understand the impact these changes. Firstly, this investigated (e.g., age, gender, education) influence e-customer preferences in Serbia. From sample n = 906 respondents, conditional dependencies between demographics and were tested. hypothetical framework 24 tested hypotheses, successfully rejected 8/24 (with p < 0.05), suggesting high association with purchase frequency reasons for quitting purchase. However, although reported test statistics suggested an association, understanding interactions categories profiles was still required. Therefore, second part considers MCA-HCPC (Multiple Correspondence Analysis Hierarchical Clustering on Principal Components) identify profiles. revealed three main clusters: (1) young, female, unemployed e-customers driven mainly by reviews; (2) retirees older adults infrequent purchases, hesitant buy without experiencing product person; (3) employed, highly educated, male, middle-aged who prioritize fast accurate delivery over price. In third stage, clusters are used as labels Machine Learning (ML) classification tasks. Particularly, Gradient Boosting (GBM), Decision Tree (DT), k-Nearest Neighbors (kNN), Gaussian Naïve Bayes (GNB), Random Forest (RF), Support Vector (SVM) used. results that GBM, RF, SVM had performance identifying Lastly, after performing Permutation Feature Importance (PFI), findings work status, education, income determinants shaping developing marketing strategies.

Language: Английский

Citations

0