A Machine Learning Framework for Evaluating MSME Incubation Dynamics

Abstract

The Business Incubation is an important way to assist entrepreneurs in transferring their innovative ideas into viable businesses, offering resources, expertise, and connecting them with available resources for assistance. In India, the MSME Hackathons represent one type of competition to generate new and exciting ideas by identifying the best solutions across different technology and industry sectors. This study will develop a data-based framework for understanding innovation patterns from the MSME business incubation framework by using the officially approved ideas generated in multiple MSME Idea Hackathons. This research will also use an innovation analytics perspective as the basis for developing results, rather than using traditional descriptive approaches for understanding MSME incubator policy development; therefore, the results will identify how many institutions, states, and knowledge bases are in the overall data set and relate to the overall structure of the innovation network. A machine learning–based analytical approach is employed to examine innovation domains and structural relationships among different innovation ecosystems. Supervised ensemble modeling is used to evaluate the influence of institutional affiliation and geographical location on thematic specialization, while unsupervised clustering and dimensionality reduction techniques are applied to explore structural patterns within the dataset. The classification model achieved an accuracy of 0.672, precision of 0.599, recall of 0.672, and an F1-score of 0.624, indicating moderate predictive capability while confirming that institutional and geographical attributes provide meaningful signals for innovation domain classification. The results further indicate that the structure of institutional networks and regional environments significantly influences innovation-specific focus areas, while also revealing the coexistence of both organized and distributed forms of innovation activity reflected in hackathon participation patterns. The study contributes a scalable and interpretable analytical framework for the macro-level evaluation of incubation initiatives and provides empirical insights to support evidence-based policy formulation aimed at strengthening the MSME innovation system.