Predicting Success Rate of Startups using Machine Learning Algorithms

It is essential for startups to assess if they are on the path to success. The failure rate of startups was around 90% in the year 2019 and hence it is necessary to know the success rate of a startup. Success for startups can be twofold: launching an IPO (Initial Public Offering) or getting merged/acquired by another company. This paper attempts to determine the success of a startup in terms of getting merged or acquired. With the help of historical data available on startups, five models have been built and compared to predict if a startup would get acquired or not. The models that have been used are Decision Trees, Random Forest, Gradient Boost, Logistic Regression, and MLP Neural networks. The data used to train these models includes key features such as valuations, funding rounds, investments, etc. By using said models, one would be able to get an estimate of the trajectory of the company. After applying the models, we were able to get an accuracy of around 92% in all of them. This information would be vital to various stakeholders of the company as well as potential investors.