A dfits measure is a regression diagnostic that quantifies the standardized difference between the fitted values from the full model and the fitted values from the model fit after deleting a single observation. For observation i, dfits_i = (y_hat_i - y_hat_i^(i)) / (s_i^(i) * sqrt(h_ii)), where y_hat_i is the fitted value, y_hat_i^(i) is the leave-one-out fitted value, s_i^(i) is the delete-i standard error estimate, and h_ii is the leverage of observation i from the hat matrix. It is computed algorithmically from regression output and persists through statistical software implementations and published diagnostic tables. [formal: dfits | substrate: behavior | horizon: a moment | explicit: yes | epoch: 0.01]
Accepted ontology entry
dfits
A dfits measure is a regression diagnostic that quantifies the standardized difference between the fitted values from the full model and the fitted values from the model fit after deleting a single observation. For observation i, dfits_i =…
Definition
Why it is in scope
A human-made regression diagnostic measure that quantifies the standardized change in a model's fitted values when a single observation is deleted from the fitting data.
Names and aliases
- dfitsen · CANONICAL
Relations from this entry
- cmsesl4wm05uq3vv3rv0svp7mDERIVED_FROM →
dfits is mathematically derived from residuals — it measures the influence of each observation on the fitted values, computed from the residuals of deleted-case fits. Residuals existed first and fed into this metric.
- cmskb43xn048dnobp35duoxb0DERIVED_FROM →
dfits is computed using the hat matrix (diagonal leverage values) and deleted-case residuals. The hat matrix existed first and fed into the construction of dfits as an influence measure. The which-came-first test: the hat matrix concept predates the dfits statistic.
- cmsdai2d503n23vv3e00xn5baSERVES →
dfits measures the influence of each observation on fitted regression values. It is built and maintained as a diagnostic tool specifically for regression analysis — its designed purpose is to serve regression by quantifying which data points disproportionately affect model fit.
Relations to this entry
No accepted relations in this direction.
Record identity
- Created
- Aug 8, 2026, 4:03 PM UTC
- Content hash
- 549c804f32029ffef5eb8b1730ec3657930dd67071f75085c1c263d8b75d2f34