Predictive Analytic for Investment Target Selection in Startup Companies Using Logistic Regression and K-Means Algorithms

This study aims to evaluate the impact of early-stage funding (first-time funding, number of funding rounds, amount of funding) on the future success of startups using a predictive analytics model. The data used in this research was obtained from Crunchbase, which includes over 3 million startup records as of the end of 2023. The study focuses on startups operating in Southeast Asia that were founded between 2008 and 2023. The logistic regression algorithm is used to analyze whether early-stage funding influences the future success of startups, measured through IPO, acquisition, unicorn status, or other late-stage criteria. The test results show that the logistic regression model achieved an CA value of 0.905, F1 score of 0.899, and Precision of 0.901. Compared to k-NN and gradient boosting algorithms, logistic regression demonstrates the most consistent and strong performance across almost all metrics. Additionally, the K-means algorithm is used to determine the clusters of startups to be selected for funding. This study concludes that early-stage funding significantly impacts the future success of startups, and the logistic regression model shows good classification performance in predicting this success.