Those formulae allows computation of Gain, Bias and Error from just five accumulation variables. This eliminates the need of storing and processing all samples of the set (2N variabiles)
>>>Linear Regression<<<
Bias and Gain
>>>Linear Regression Error<<<
Error Estimation
>>>Correlation<<<
Pearson Correlation and my custom Linear Correlation metric for use in neural networks
No comments:
Post a Comment