Linear Regression Models

From Open Risk Manual

Definition

Linear Regression denotes a popular class of models that establish linear relationships between input variables (explanatory variables or characteristics) and output variables (dependent variables).

Can be estimated by Ordinary Least Square Methods.

Usage

A linear regression is the typical first attempt at modelling a large variety of financial / risk variables.

Generalized linear models

Example

Issues and Challenges

While linear regression models are simple, intuitive and fairly well understood there are many circumstances in which the underlying behavior is more complex and cannot be adequately captured by linear relationships except in a very approximate way.

External Links