Anybody did some research at it? Is it a better choice to be used in order to limit problems like over-fitting? Any insight on how the model will generate to an independent data set? Am I at the right track or are there any more reasons that it is a best choice for the testing?
10 -fold cross validation is a training procedure used to test classification performance.
Data is divided into 10 folds with the same size, then training can be done on 9 folds of the data and testing can be done on the remaining fold.
The process is then repeated 10 times to test on all available data. This why it is better to choose cross-validation rather than having a separate training and testing datastes.
hope this helps,
Hi,
Sometimes you really have no choice. In one of the examples in the CART monograph the authors work with a medical data set that contained data for just 215 patients. Such a small data set will not support a useful division of the data into train and test partitions. So we use cross validation to be able to both train with all the data and then indirectly test with all the data as well. The magic of cross validation is explained in other of our posts and videos.
Cross validation is best method to find out your research how much good it is? any repeated things are there? which one is good method? you have find all these and many from the cross validation.
thanks.
It is used to have better confidence in your prediction accuracy. For example if you split your data into train and test sets them run you classifier and get an accuracy it is only for that arrangement. Performing K-folds cross-validation loops through all the data and gets classification accuracy for each time, they are then averaged to give you a number more representative of overall accuracy. K-folds with k=10 is a good start for cross-validation.
- Badges
- Science topic