crossval.regmodel function

Cross-validation of a regression model

Does cross-validation of a regression model

crossval.regmodel(obj, x, y, cv, cal.fun, pred.fun, cv.scope = "local")

Arguments

• obj: a regression model (object of class regmodel)
• x: a matrix with x values (predictors from calibration set)
• y: a matrix with y values (responses from calibration set)
• cv: number of segments (if cv = 1, full cross-validation will be used)
• cal.fun: reference to function for model calibration
• pred.fun: reference to function for getting predicted y-values (see description)
• cv.scope: scope for center/scale operations inside CV loop: 'global' — using globally computed mean and std or 'local' — recompute new for each local calibration set.

Returns

object of class plsres with results of cross-validation

Function pred.fun must take four agruments: autoscaled x-values, array with regression coefficients, vectors for centring and scaling of y-values (if used). The function must return predicted y-values in original units (unscaled and uncentered).