Providing functions for automate selection and fitting of scoring & weighted regression based models
The objective of the codes is to automate scoring model selection and fitting. The codes optimize fitting of a weighted regresion model on intervals constructed with a trend wize approach to every variable. The codes also provide the prediction procedure for new data.
Depending on the type of variable, different performance metrics and statistical significance tests are available.
- trameado: Estimates the optimum weighted interval partition for each explanatory variable in the set. This partition is optimum on the chosen performance metric. It also provide the prediction procedure for new data.
- modelado: Fits a weighted regression model and makes an optimum variable backward selection conidering minimum count of variables, collinearity, variable weight thresholds and significance. It also provide the prediction procedure for new data.
- transformaciones: Apply transformations to re shape each variable or shifts to consider previous information.