1、根据比例划定测试集索引和训练集索引
n_rating = 100
ratio = 0.2
test_index_list = np.random.choice(n_rating, size=int(n_rating * ratio), replace=False)
train_index_list = list(set(range(n_rating))-set(test_index_list))
print(test_index_list)
print(train_index_list)
2、引入真实数据,按照索引划分
data = np.random.rand(100, 5)
label = np.arange(1, 101)
label = np.reshape(label, (-1, 1))
train_data, test_data = data[train_index_list], data[test_index_list]
train_label, test_label = label[train_index_list], label[test_index_list]
print(train_data.shape)
print(test_data.shape)
print(train_label.shape)
print(test_label.shape)
全部代码
import numpy as np
n_rating = 100
ratio = 0.2
test_index_list = np.random.choice(n_rating, size=int(n_rating * ratio), replace=False)
train_index_list = list(set(range(n_rating))-set(test_index_list))
print(test_index_list)
print(train_index_list)
data = np.random.rand(100, 5)
label = np.arange(1, 101)
label = np.reshape(label, (-1, 1))
train_data, test_data = data[train_index_list], data[test_index_list]
train_label, test_label = label[train_index_list], label[test_index_list]
print(train_data.shape)
print(test_data.shape)
print(train_label.shape)
print(test_label.shape)