Prediction of Future New Entrepreneurs Using Population Survey Data
Presentation Type
Poster
Presentation Type
Submission
Keywords
Entrepreneurship, Data Science, Population Survey, Machine Learning
Department
Nutritional Science
Major
Computer Science / Math
Abstract
Entrepreneurship is essential to the U.S. economy and understanding the characteristics of entrepreneurs can help organizations like the Small Business Administration (SBA) assist with growth of startup businesses. Understanding what makes an entrepreneur can be difficult and predicting future entrepreneurs even more so. Using data from the U.S. Census Bureau’s Current Population Survey (CPS), we are able to track annual survey respondents across a 16 month period from non-self-employment to self-employment. With our own technique, we successfully matched records from nearly 1.2 million respondents over a 10 year period from 2009 to 2019. Using machine learning techniques trained with an optimized set of features, we created one model to classify entrepreneurs and another to predict future entrepreneurs or the transition to entrepreneurship. The first model can be used to help understand defining factors of entrepreneurship and the second can be used to assist potential new or first time business owners. Other research conducted on the subject relies on niche datasets, which are expensive to collect and hard to come by; through machine learning and data mining approaches our technique shows the untapped potential of population data sets, such as the CPS, in business and economic research such as this.
Faculty Mentor
Alfonso Berumen
Funding Source or Research Program
Keck Scholars Program
Location
Waves Cafeteria
Start Date
22-3-2024 1:30 PM
End Date
22-3-2024 2:30 PM
Prediction of Future New Entrepreneurs Using Population Survey Data
Waves Cafeteria
Entrepreneurship is essential to the U.S. economy and understanding the characteristics of entrepreneurs can help organizations like the Small Business Administration (SBA) assist with growth of startup businesses. Understanding what makes an entrepreneur can be difficult and predicting future entrepreneurs even more so. Using data from the U.S. Census Bureau’s Current Population Survey (CPS), we are able to track annual survey respondents across a 16 month period from non-self-employment to self-employment. With our own technique, we successfully matched records from nearly 1.2 million respondents over a 10 year period from 2009 to 2019. Using machine learning techniques trained with an optimized set of features, we created one model to classify entrepreneurs and another to predict future entrepreneurs or the transition to entrepreneurship. The first model can be used to help understand defining factors of entrepreneurship and the second can be used to assist potential new or first time business owners. Other research conducted on the subject relies on niche datasets, which are expensive to collect and hard to come by; through machine learning and data mining approaches our technique shows the untapped potential of population data sets, such as the CPS, in business and economic research such as this.