在实际训练模型时，我们需要将样本数据随机分割为训练集和测试集，其中训练集用来训练模型，测试集用来验证模型的实际效果。
下面介绍使用sklearn开发时如何划分数据集。
为了演示分割效果，我们需要生成一些数据集，生成数据集的代码如下：
import numpy as np sample_number = 10 feature_number = 6 X = np.random.rand(sample_number, feature_number) y = np.random.randint(0, 2, sample_number) print(X) print(y)
输出结果如下：
[[ 0.66130409 0.97676746 0.71803508 0.99730168 0.08229832 0.86526473] [ 0.30021108 0.06706006 0.98098814 0.68709696 0.48207104 0.71118232] [ 0.65834414 0.28948205 0.26636723 0.10397637 0.30376824 0.08447934] [ 0.61166203 0.76414791 0.53293265 0.72645187 0.83650127 0.80324535] [ 0.3168741 0.20964209 0.1050409 0.96667613 0.76779378 0.11599263] [ 0.72638776 0.26922676 0.34266498 0.24431192 0.79987585 0.9393169 ] [ 0.15487034 0.35483264 0.83128334 0.48502717 0.12895195 0.82433822] [ 0.72285669 0.54464814 0.75142994 0.96835823 0.87040141 0.97489327] [ 0.81949806 0.39009266 0.57812486 0.96307663 0.39236456 0.20011817] [ 0.93098734 0.0461967 0.27630984 0.34199953 0.21883858 0.07402595]] [1 1 0 1 0 1 1 1 1 1]
下面介绍如何分割数据集，分割数据集的代码如下，其中random_state是随机种子。
from sklearn.model_selection import train_test_split X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.3, random_state=42) print(X_train) print(X_test) print(y_train) print(y_test)
输出结果如下：
[[ 0.66130409 0.97676746 0.71803508 0.99730168 0.08229832 0.86526473] [ 0.72285669 0.54464814 0.75142994 0.96835823 0.87040141 0.97489327] [ 0.65834414 0.28948205 0.26636723 0.10397637 0.30376824 0.08447934] [ 0.93098734 0.0461967 0.27630984 0.34199953 0.21883858 0.07402595] [ 0.3168741 0.20964209 0.1050409 0.96667613 0.76779378 0.11599263] [ 0.61166203 0.76414791 0.53293265 0.72645187 0.83650127 0.80324535] [ 0.15487034 0.35483264 0.83128334 0.48502717 0.12895195 0.82433822]] [[ 0.81949806 0.39009266 0.57812486 0.96307663 0.39236456 0.20011817] [ 0.30021108 0.06706006 0.98098814 0.68709696 0.48207104 0.71118232] [ 0.72638776 0.26922676 0.34266498 0.24431192 0.79987585 0.9393169 ]] [1 1 0 1 0 1 1] [1 1 1]