Model Comparision


Thinking of pattern matching further, I came across splines. Here is my query, suppose I fit linear spline on two similar data streams ( stock prices) which have similar patterns. Now my question is how to compare these two models for significance ( or say similarity) . What tests can be used to compare two model outputs on similar looking data ?



@ambergupta This problem can be thought as a regression problem and treated as such. Then you can use metrics such as

  • Mean Squared Error
  • Mean Absolute Error
  • R square, etc

Refer this guide, for more evaluation techniques,


Adjusted R-square , AIC , BIC criterion works better