A partial leverage plot (also known as a component leverage plot) is a regression diagnostic visualization that decomposes the overall leverage of each observation into contributions from individual predictors. Each data point is assigned a partial leverage value for each predictor variable, allowing analysts to see which specific variables drive a point's influence on the regression model.
The plot is constructed by computing the component leverage h_i(j) = (x_i(j) - x̄(j))² / Σ_k(x_k(j) - x̄(j))² for each observation i and predictor j, then displaying these values graphically — typically as a bar chart or scatter plot ordered by observation. Partial leverage plots are produced by standard statistical software packages (R's car package, Python's statsmodels) and are used in diagnostic regression analysis to pinpoint which predictor(s) cause a given observation to exert disproportionate influence.
[formal: partial_leverage_plot | substrate: mind | horizon: a moment | explicit: yes | epoch: 0.01]