StatReasonRegression
Regression models relate an outcome variable to one or more predictors, producing a formula you can inspect rather than a black-box prediction. Every model here reports its coefficients, a measure of overall fit, and residual behavior alongside the prediction itself, since a model that fits poorly can still produce a confident-looking number if you only look at the prediction.
These tools calculate sample size, parameter count, and residual patterns directly from your data, but whether your predictor and outcome were measured in a valid temporal order, whether important variables were omitted, and whether prediction beyond your data's range is appropriate are left for you to verify.
Engines in this category
- Linear Regression Calculator
Fit and interpret a simple linear predictive relationship between two variables.
- Adjusted R-Squared Calculator
Compare model fit accounting for the number of predictors used.
- Durbin-Watson Calculator
Summarize first-order autocorrelation in ordered regression residuals.
- Regression Prediction Interval Calculator
Give prediction uncertainty for one new individual outcome.
- Regression Confidence Interval Calculator
Give mean-response uncertainty at a specific x value.
- Multiple Linear Regression Calculator
Fit adjusted linear associations across two or more predictors at once.
- Logistic Regression Calculator
Model a binary outcome from one or more predictors via maximum likelihood.
- Polynomial Regression Calculator
Fit a declared curved (quadratic or cubic) relationship between two variables.
- Residual Diagnostics Calculator
Inspect fitted-model residual patterns, leverage, and influence flags.
- Multicollinearity VIF Calculator
Quantify how much predictor collinearity inflates coefficient variance.
- Cook's Distance Calculator
Flag observations disproportionately influential to an OLS fit.
- Regression Model Comparison Calculator
Compare two properly nested OLS models with a partial F-test.