lift chart

A lift chart is a visual plot used in predictive modeling and data science that compares the performance of a classification model against a baseline random predictor. It plots cumulative gain (the percentage of positive cases captured) against the percentage of the dataset examined at various classification thresholds. The chart is constructed by sorting predictions in descending order of confidence, dividing the sorted data into equal-sized deciles (or other bins), and calculating the hit rate in each bin. The lift value at each point is the ratio of the model's hit rate to the overall positive rate in the dataset — a lift of 2.0 at the top decile means the model captures positives twice as efficiently as random sampling would. The chart enables practitioners to visualize how much better the model is than chance and to select an operating threshold that balances capture rate against the cost of examining cases. It is widely used in marketing response modeling, credit risk assessment, and medical screening optimization. The curve begins at the origin, rises above the diagonal baseline (which represents random guessing), and may flatten as the tail of low-confidence predictions is reached. [formal: charta elevatio | substrate: mind | horizon: a moment | explicit: yes | epoch: 0.01]