<strong>Getting more data</strong> would help to avoid overfitting.<br>
And some machine learning algorithms have hyper parameters that is effective for avoiding overfitting (<font color="#1E90FF"><strong>reguralization parameters</strong></font>).<br>
<br>
So, how we can check whether the model is overfitted or not?<br>
We want to estimate the performance of the model for unknown data.<br>
To estimate that, <font color="#1E90FF"><strong>cross validation</strong></font> method is commonly used.<br>
In the cross validation, the data is partitioned into k equal sized subset of the data.<br>
Of the k subsets, a single subset is retained as the validation set for testing the model, and the remaining k-1 subsets are used for training the model.<br>
This process is repeated k times, with each of the k subsets used exactly once as the validation data.<br>
Finally, the averaged value of the k results is used as the estimated score for unknown data.<br>
This is called <strong>k-fold cross validation</strong>.