ML Insights
Data mining models applied to startup funding data
Four machine learning approaches are applied to the Crunchbase dataset: classification to predict startup outcomes, K-Means clustering to segment markets, DBSCAN for density-based clustering with outlier detection and association rule mining to discover hidden patterns.
Success Prediction
Seven classifiers trained on 5,500 labeled startup outcomes. Input a startup profile and see the predicted probability of acquisition vs. closure.
Market Clusters
K-Means clustering groups 190+ markets by funding volume, deal size, success rate and round progression. Interactive scatter with detail panel.
Outliers & DBSCAN
DBSCAN finds density-based clusters and noise points. Isolation Forest scores individual companies by anomaly level. Complements K-Means.
Funding Patterns
Apriori algorithm discovers co-occurrence rules, e.g., startups with trait X are Y× more likely to get acquired.