Investor IQ: A Data-Driven Machine Learning Framework for Startup Success Prediction Using Competition Percentage and Innovation Percentage

Startups drive economic growth and innovation, yet investors face challenges in manually assessing their success potential. The Proposed Model “Investor IQ” is an AI-driven platform designed to bridge this gap by providing a data-driven success prediction model. Our model evaluates startups based on Problem Statement, Solution, Sector, and Funding Amount. Unlike traditional models, Investor IQ employs dynamic thresholding calculated from sector-specific growth potential and competition saturation for more accurate predictions. Using startup datasets, cosine similarity is applied, and dynamic thresholding is to assess innovation, competition, and success probability. The model undergoes K-Fold CrossValidation to ensure robustness and is further evaluated using classification reports, AUC-ROC scores, and confusion matrices. It leverages a stacking-based ensemble approach, combining Logistic Regression, Decision Trees, Random Forest, XGBoost, LightGBM, and CatBoost as base models with a meta-model for final predictions. Through this method, proposed model Investor IQ achieves 78% accuracy, empowering investors and entrepreneurs with intelligent insights to minimize risk and enhance decision-making in the dynamic startup ecosystem.