Article directory
1. In the Scikit-learn library, you can use the train_test_split function to divide the training set and test set
InScikit-learn (remember) library, you can use the train_test_split function to divide the training set and test set. This function receives four parameters: 数据集，测试集大小，随机种子和随机状态.
from sklearn.model_selection import train_test_split
# 假设 X 是特征数据，y 是标签数据
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
In this example, X和y是原始的数据和对应的标签. test_size=0.2 means20%的数据将被用作测试集. random_state=42 is to ensure that every time it runs, 数据分割的方式相同. If you want to get a different split every time you run it, you can omit this parameter.
Note: The division of the training set and the test set should be at数据预处理之前进行的 to ensure that the data distribution of the two sets is similar to the original data set.
Summarize
The train_test_split function is a very commonly used function in the scikit-learn library, which is used to split the original data set into a training set and a test set. this函数的主要作用如下：
数据集分割: During the training process of a machine learning model, it is usually necessary to divide the original data set into a training set and a test set. The training set is used to train the model, and the test set is used to evaluate the performance of the model. The train_test_split function can easily implement this data set splitting operation.
保护数据：By using a part of the original data set as a test set, you can protect the original data set from being used entirely for training, so that the integrity of the data can be maintained when the original data is further analyzed or used for other purposes.
模型评估：The existence of the test set allows us to evaluate the performance of the trained model to understand the performance of the model on new data. This helps to identify potential problems with the model, such as overfitting or underfitting, and make adjustments accordingly.
随机性：The train_test_split function is stochastic, which means that every time you run the function, you may get slightly different results. This provides randomness to the data partitioning and helps improve the model's generalization ability.