![]() The simpler the model, the greater the bias. Bias: error introduced by approximating a real-life problem.The idea is to find estimates which vary little between training datasets. Variance: how the predictions would change if we change the training dataset.Find a smaller tree with fewer splits that lead to lower variance and better interpretation at the cost of little bias ie. The key idea is to minimise the expected test error. This is because the resulting tree might be too complex. The process described above may produce good predictions on the training set, but it is likely to overfit the data, leading to poor test set performance. # create an index variable to identify a training set Create training (70%) and test (30%) sets and use set.seed() for reproducibility.Read the data and create a dummy dataset:. ![]()
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |