What type of chart is created when plotting cumulative lift for selection fractions?

A cumulative lift chart is specifically designed to illustrate the effectiveness of a model in predicting responses relative to a baseline. When plotting cumulative lift for selection fractions, this chart allows data analysts to evaluate how well a model performs in terms of capturing true positive responses compared to a random selection.

The cumulative lift can be derived from a model that ranks observations based on their predicted probabilities, allowing stakeholders to observe how much better the model performs compared to random selection. This kind of visualization is essential because it provides insight into the incremental value generated by the model, showing at each selection fraction what percentage of true positives is captured.

In comparison, other types of charts, while useful in their own contexts, do not specifically depict cumulative lift. For instance, a response chart primarily displays the actual responses to a set of predictions without the cumulative perspective. A profit chart focuses on financial metrics, analyzing the profitability of predictions rather than their lift. Lastly, a false positive fraction chart illustrates the proportion of incorrect positive classifications but does not address cumulative success at varied selection fractions. Hence, the cumulative lift chart stands out as the appropriate choice for visualizing cumulative lift based on selection fractions.