cross validation error meaning?

I need help deciphering what a cross validation error (kfoldloss) of 536 means. I've developed a program that creates a decision tree. The program runs with 2,286 data points for several different variables. Does this mean that the model gets 536 predictions incorrect?

Answer 1

[Mukul Rao]

Hello,

To begin with I would recommend reading a little about what cross-validation stands for from here:

Cross Validation

The documentation for kfoldLoss

https://www.mathworks.com/help/stats/regressionpartitionedmodel.kfoldloss.html

says that L is "The loss (mean squared error) between the observations in a fold when compared against predictions made with a tree trained on the out-of-fold data. If mode is 'individual', L is a vector of the losses. If mode is 'average', L is the average loss."

This post has a mathematical representation of the aforementioned statement:

https://stats.stackexchange.com/questions/17431/a-mathematical-formula-for-k-fold-cross-validation-prediction-error

Therefore, the way I see it, the value of 536 indicates that the average mean squared error over the cross validation sets for your model is 536.

If you would like to see the individual loss values corresponding to each of the partitioned data sets, you can set the 'mode' property for kfoldLoss to be 'individual'.