# Quantitative Regression Analysis

Jump to navigation
Jump to search

## Contents

These methods develop regression models which attempt to predict a quantity based on measurements of responses (x-block) and corresponding quantities (y-block) on known samples.

The y-block may contain a physical quantity which is directly related to the measurements in the x-block, or it may be a value which is indirectly related to the measured x-block values. In the latter case, the resulting model is considered an "inferential" model.

## Standard Linear Modeling Methods

- analysis - Graphical user interface for data analysis.
- cls - Classical Least Squares regression for multivariate Y.
- crossval - Cross-validation for decomposition and linear regression.
- mlr - Multiple Linear Regression for multivariate Y.
- pcr - Principal components regression for multivariate Y.
- pls - Partial least squares regression for multivariate Y.
- stepwise_regrcls - Step-wise regression for CLS models.

## Multiway Models

- npls - Multilinear-PLS (N-PLS) for true multi-way regression.
- modelviewer - Visualization tool for multi-way models.

## Local, Non-linear, and Other Methods

- ann - Artificial Neural Network regression models.
- cr - Continuum Regression for multivariate y.
- frpcr - Full-ratio PCR calibration and prediction.
- lwr - Locally weighted regression for univariate Y.
- polypls - PLS regression with polynomial inner-relation.
- ridge - Ridge regression by Hoerl-Kennard-Baldwin.
- svm - SVM Support Vector Machine for regression.
- svmda - SVM Support Vector Machine for classification.
- xgb - Gradient Boosted Tree Ensemble for regression using XGBoost.
- xgbda - Gradient Boosted Tree Ensemble for classification (Discriminant Analysis) using XGBoost.

## Other Topics

- Application of Models to New Data
- Model Analysis and Calculation Utilities
- Plotting Utilities
- Related Tools

(Sub topic of PLS_Toolbox_Topics)