Bayesian network analysis for the dynamic prediction of early stage entrepreneurial activity index
Highlights
► A dynamic relationship among entrepreneurial attitudes, activities, and aspirations is identified. ► Bayesian network analysis is used to predict the early stage entrepreneurial activity index. ► The proposed index turns out to have very high prediction accuracy. ► This index can be used to predict various aspects of entrepreneurial aspiration.
Introduction
The Global Entrepreneurship Monitor (GEM) project is “an annual assessment of the entrepreneurial activity, aspirations and attitudes of individuals across a wide range of countries (Bosma, Wennekers, & Amoros, 2011)”. The GEM focuses on three main objectives: “(1) to measure differences in entrepreneurial attitudes, activity and aspirations among economies, (2) to uncover factors determining the nature and level of national entrepreneurial activity, and (3) to identify policy implications for enhancing entrepreneurship in an economy (Bosma et al., 2011).”
To identify the flow of entrepreneurial activities related to national economic growth from a society, culture, and politics, the GEM study constructed a conceptual framework of the original GEM model (Bygrave, Hay, Ng, & Reynolds, 2003). GEM considers that national economic growth is the result of such set of entrepreneurship activities. The initial GEM model was to develop to integrate advances in understanding the entrepreneurial process and to allow for further exploration of patterns detection (Herrington, Kew, & Kew, 2010). On the other hand, recently a GEM model was reworked based on the combination of three main components: entrepreneurial attitudes, entrepreneurial activity, and entrepreneurial aspirations (Bosma et al., 2011). According to Hessels, van Gelderen, and Thurik (2008), entrepreneurial motivations of perceived opportunity for entrepreneurship have relevance to entrepreneurial aspirations. In addition, if an individual exhibits positive attitudes toward entrepreneurship, the tendency to be a potential entrepreneur or get involved in entrepreneurial activity will increase (Krueger, 2007, Krueger and Brazeal, 1994).
Entrepreneurial attitudes, entrepreneurial activity, and entrepreneurial aspirations are latent variables and are comprised of many observable variables. However, none of the existing studies consider the dynamic causal relationship among the observable variables of entrepreneurial attitudes, activity, and aspirations, although Sohn and Ju (submitted for publication) attempted to model the relationship among these three factors. In addition, no research was conducted to reflect potential time-lagged causal relationship among them. If such a relationship is identified, it can be utilized to predict the degree of future entrepreneurial activity levels of various nations and corresponding entrepreneurial policies can be set to take necessary actions to vitalize entrepreneurial activities.
The main purpose of this study is to identify a dynamic causal relationship among measurement variables of entrepreneurial attitudes, early stage entrepreneurial activity, and aspirations at a national level. We utilize a Bayesian network (BN) to identify such a relationship. The BN is a powerful formalism for representing the joint probability distribution of a set of related variables. It can represent an area of knowledge and uncertainties of related variables, enabling reasoning with uncertainties.
Moreover, we propose an index for early stage entrepreneurial activity that can be used to predict the level of future early stage entrepreneurial activity reflecting causal relationships found through BN. We compare this index to the Total early-stage Entrepreneurial Activity (TEA) index (Bosma, Acs, Eutio, Coduras, & Levie, 2009). The TEA index is a composite measure of the current year’s degree of early stage entrepreneurial activity representing the percentage of both nascent entrepreneur and new business owner and therefore cannot be used for prediction.
We expect that our findings will contribute to the understanding of the dynamic relationship among national entrepreneurial attitude, activity, and aspirations. This can be used as feedback information to boost entrepreneurship.
This paper is organized as follows. In Section 2, we review the existing literature of GEM. In Section 3, we introduce a proposed BN as well as the data used in this study. In Section 4, we discuss the results of our analysis. Finally, we conclude our study and suggest areas for future research in Section 5.
Section snippets
Literature review and research hypotheses
GEM employs a comprehensive socio-economic approach and takes into account the degree of involvement in entrepreneurial activity within a country, illustrating different types and phases of entrepreneurship. Especially, the approach and the focus on the individual as an entrepreneur differentiate GEM measures from other data sets that measure new business registrations. GEM project uses several measurement variables for three factors (attitudes, activity and aspirations), respectively, as
Bayesian network
BNs are directed acyclic graphs (DAGs) used to denote uncertain knowledge in artificial intelligence (Huang and Bian, 2009, Jensen, 1996). A BN is defined as follows: BN = (S, θ), where S = (N, A) represents the network structure and N represents the set of all nodes in a BN. Each node describes a discrete variable with a finite number of mutually exclusive states; A is the set of all edges in a BN. Each edge represents the relationship of a father and child by linking two nodes; θ represents the set
The results of BN
Fig. 3 shows the results of trained BN model 1 based on hypothesis 1.
According to fitted BN model 1, nascent entrepreneurship activity in the present year is related to all entrepreneurial attitudes in the previous year. Such activity is relevant to high-growth expectation early-stage entrepreneurial activity (HEA) in the next year. Based on this BN, when the high level of entrepreneurial intention variable increased to 100% using ‘what–if’ analysis, nascent entrepreneurship activity sharply
Conclusion
This study examined the time-lagged dynamic relationship between entrepreneurial attitudes, activity, and aspirations, which are factors newly included in the GEM conceptual model. Bayesian network (BN) analysis was used to investigate the dynamic causal relationship among measurement variables of these three factors.
With respect to perceived capabilities and entrepreneurial intention the previous year, entrepreneurial attitudes positively affected both nascent entrepreneurship and new business
Acknowledgment
This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-2011-327-B00324). We also appreciate Yong Han Ju for his comments on our work. Permission to use excerpts/tables from Global Entrepreneurship Monitor: 2008 and 2009 Executive Reports, which appear here, has been granted by the copyright holders. The GEM is an international consortium and this report was produced from data collected in, and received from, 22 countries in 2008 and 18
References (46)
- Acs, Z. J., & Szerb, L. (2009). The global entrepreneurship index (GEINDEX), Jena Economic Research Papers, No. 2009,...
- et al.
Opportunity costs and entrepreneurial activity
Journal of Business Venturing
(1995) - et al.
Entrepreneur’s exit experience and growth aspirations
Frontiers of Entrepreneurship Research
(2010) - et al.
Effects of age at entry, knowledge intensity, and imitability on international growth
The Academy of Management Journal
(2000) - Bacciu, D., Etchells, T. A., Lisboa, P. J. G., & Whittaker, J. (in press). Efficient identification of independence...
- et al.
A framework for the analysis of dynamic processes based on Bayesian networks and case-based reasoning
Expert Systems with Applications
(1998) - Bosma, N. (2012). The global entrepreneurship monitor (GEM) and its impact on entrepreneurship research. GEM Working...
- Bosma, N., Acs, Z. J., Eutio, E., Coduras, A., & Levie, J. (2009). Global entrepreneurship monitor: 2008 Executive...
- Bosma, N., & Levie, J. (2010). Global entrepreneurship monitor: 2009 Executive Report. Babson College. Babson...
- Bosma, N., Wennekers, S., & Amoros, J. (2011). Global Entrepreneurship Monitor 2011 extended report: Entrepreneurs and...
Improving algorithms for structure learning in Bayesian networks using a new implicit score
Expert Systems with Applications
Executive forum: A study of informal investing in 29 nations composing the global entrepreneurship monitor
Venture Capital: An International Journal of Entrepreneurial Finance
Identifying product failure rate based on a conditional Bayesian network classifier
Expert Systems with Applications
The influence of variables of attitude on collective entrepreneurship
International Entrepreneurship and Management Journal
Software maintenance project delays prediction using Bayesian networks
Expert Systems with Applications
The behavioral impact of entrepreneur identity aspiration and prior entrepreneurial experience
Entrepreneurship Theory and Practice
Bayesian network classifiers
Machine Learning
The efficient estimation of missing information in causal inverted multiway trees
Knowledge-Based Systems
Social interactions and entrepreneurial activity
Journal of Economic & Management Strategy
Entrepreneurial aspirations, motivations, and their drivers
Small Business Economics
A Bayesian network and analytic hierarchy process based personalized recommendations for tourist attractions over the internet
Expert Systems with Applications
Cited by (18)
The role of institutions in early-stage entrepreneurship: An explainable artificial intelligence approach
2024, Journal of Business ResearchOne size fits all? Using machine learning to study heterogeneity and dominance in the determinants of early-stage entrepreneurship
2022, Journal of Business ResearchCitation Excerpt :Woodside (2013) suggests moving beyond the traditional statistical methods which focus on linear regression. Although the use of analytics techniques have been limited in the entrepreneurship research, recent work in the entrepreneurship field has begun to adopt a wider variety of methods such as fuzzy-set qualitative comparative analysis (FSQCA) (Beynon et al., 2016, 2018, 2020; Coduras et al., 2016), and Bayesian networks (Sohn & Lee, 2013). Decision trees and other analytics approaches have been applied in the wider business literature (Duchessi & Lauría, 2013; Schivinski, 2021), but have not been applied to the study of heterogeneity in the entrepreneurship literature.
Modelling Entrepreneurial Intentions and Attitudes towards Business Creation among Emirati Students Using Bayesian Networks
2022, World Journal of Entrepreneurship, Management and Sustainable DevelopmentAnalysis of Influencing Factors of Social Mental Health Based on Big Data
2021, Mobile Information Systems