Division ratio 1 training and test data sets
Division ratio data set to train: val: test = 7: 2: 1;
2 randomly selected data (upset)
Sklearn used in train_test_split () function;
Tested train_test_split () function (i.e. without explicitly set default parameters, random_state case parameter),
Initialize the random seed is a true random initialization,
Testing process is as follows:
The first test, open Jupyter Lab, using the following code was as follows:
After that, close the browser, close Jupyter Lab process in the terminal;
The second test, open Jupyter Lab, test results are as follows:
Two tests were conducted at different Jupyter Lab of the process and the results generated random division is different,
Thus, train_test_split () function random_state case where the default value is a true random initialization;