Why are there any methods that can prevent or overcome overfitting?
What is overfitting:
The so-called overfitting (Overfit) is a phenomenon in which a hypothesis can be better fitted to the training data than other assumptions, but not well fitted to data sets outside of the training data. At this point we call this hypothesis that there is a overfit phenomenon.
Reasons for overfitting:
The main reason for this is that there is noise in the training data or too little training data.
Measures to prevent or overcome:1 . Increase the amount of data
2, reduce the number of feature (manual definition of how many feature or algorithm to choose these feature)
3, regularization (leaving all the feature, but for the partial feature definition of its parameter very small)
4. Cross-validation
Why is it possible to have a fit, and what methods can be used to prevent or overcome overfitting