Session: Government Agency Student Posters

Paper Number: 173215

Agent-Based Model Enhanced With Machine Learning to Predict Entrepreneurial Success Through Network Analysis 

The network success theory suggests that there is a strong link between how actively entrepreneurs engage in networking and the overall success of their startups. As a result, having a robust entrepreneurial network is often seen as a key factor in achieving entrepreneurial success. This idea is grounded in the belief that entrepreneurs benefit from socially embedded connections, which help them access resources at lower costs than through traditional market avenues. While this concept makes sense in theory, it becomes much more challenging for entrepreneurs to interpret its practical implications in quantitative terms. Entrepreneurial networks are expansive and inherently difficult to quantify, and the process of collecting, analyzing, and interpreting network data is both methodologically complex and data intensive. This study seeks to address this gap in understanding and modelling the quantitative relationship between professional networks and entrepreneurial success by pursuing three key objectives. First, it examines how structural and relational characteristics of professional networks predict entrepreneurial learning outcomes within an entrepreneurial program called NSF I-Corps. The research team worked with 35 entrepreneurs to map their professional networks, a process that ranged from 4 to 20 hours per participant. These entrepreneurs were then categorized as high, medium, or low-quality learners based on their performance and learning during the I-Corps program. Various network metrics were quantified and analyzed to determine which indicators most strongly correlated with learning success. Four key metrics were identified, each with defined thresholds corresponding to learner categories. Three of the most promising metrics were rooted in social network analyses (clustering, connectance, and the proportion of weak ties), while the fourth was nestedness inspired by bio-inspired design principles. These four stood out among 20 metrics evaluated by the research team. The second component of the study involved training a machine learning model using the professional networks of the 35 entrepreneurs. This model was used to design a predictive tool that can generate an entrepreneurial network profile with limited input, significantly reducing the time and data required. The resulting tool, developed as a graphical user interface (GUI), enables entrepreneurs to rapidly visualize their network and their position relative to the key network metrics and learning classifications. Insights from the first two phases informed the development of an agent-based model used to simulate and evaluate potential networking interventions. This model enabled the testing of multiple intervention strategies to identify best practices that can support and enhance entrepreneurial education. These interventions were integrated into the GUI, allowing entrepreneurs not only to map their current networks but also to explore how targeted strategies reshape their networks in real time. The tool shows how these changes move users closer to the structure of a successful entrepreneurial network. This interactive platform allows entrepreneurs with actionable recommendations to improve discovery, collaboration, and overall entrepreneurial performance.

Presenting Author: Hadear Hassan Texas A&M University

Presenting Author Biography: Hadear Hassan is a Ph.D. Candidate in Mechanical Engineering at Texas A&M University, where she is pursuing research at the intersection of innovation and manufacturing. Her work focuses on advancing smart and sustainable manufacturing systems, particularly through the use of bio-inspiration to enhance energy efficiency at a systems level. In addition to her research, Hadear is deeply committed to engineering education. She has been recognized with several prestigious honors, including the J. George H. Thompson Fellowship (2022), the Association of Former Students Distinguished Graduate Student Award for Excellence in Teaching (2023), the Walker Impact Award (2023), and the Cain Impact Award (2024). She is also an Associate Fellow in the Center for the Integration of Research, Teaching, and Learning (CIRTL) Academy for Future Faculty. In 2025, she was selected to attend the Global Young Scientists Summit in Singapore, a competitive international event for promising early-career researchers.

Authors:

Hadear Hassan Texas A&M University
Astrid Layton Texas A&M University
M. Cynthia Hipwell Texas A&M University