The Application of Selected Supervised Machine Learning Methods in the Classification of Family Businesses in the Context of Cluster Formation

Daria Wotzka, Pawel Fracz, Jolanta Staszewska, Joachim Foltys, Malgorzata Smolarek, Krzysztof Orzechowski
European Research Studies Journal, Volume XXVIΙ, Issue 4, 248-272, 2024
DOI: 10.35808/ersj/3516

Abstract:

Purpose: The article focuses on the application of selected supervised machine learning methods for the classification of family businesses in the context of cluster formation. The research aim was to evaluate various learning algorithms to develop a tool for classifying entrepreneurs, intended for use in an online application. Design/Methodology/Approach: Through a comprehensive survey, 448 responses were gathered, addressing various aspects of clusters and related experiences. Based on the collected data, classification methods for respondents were developed in the context of cluster formation. The classifier categorizes entrepreneurs based on their under-standing of cluster concepts, managers' perceptions of clusters, companies' experiences with clusters, the operational status of clusters, and experience in business networks. The article conducts a comparative analysis of the classification outcomes derived from the application of decision trees and neural networks across diverse configurations. This analysis, based on distinct evaluation metrics, culminates in the identification of the most optimal algorithm suited for the task at hand. Findings: As a result of the conducted research, a supervised machine learning algorithm in the form of an ensemble bagged tree was selected. This algorithm achieves an average effectiveness of 82%, measured as the arithmetic mean of accuracy, specificity, precision, sensitivity, F1 score, and the Matthews correlation coefficient. The median value was 96%. Practical Implications: The presented results have been implemented in the form of a computer application that allows for the simulation and classification of entrepreneurs based on their business experiences. The developed tool is being deployed as a web-based application, serving as a platform to showcase the numerous possibilities and benefits of cluster formation. Originality/Value: This study represents a novel approach, as there are no available articles specifically applying machine learning techniques to classify entrepreneurs, particularly family-owned businesses, in the context of cluster formation.


Cite Article (APA Style)