Python -- Pandas & Sklearn：选取训练集&测试集

1. Pandas 中的 sample

n，设置直接取原数据中的多少数据。frac，设置取原数据的多少比例。

df = pd.DataFrame(np.random.randn(400).reshape(100,4))
df.head()
	0	1	2	3
0	0.014483	0.044465	0.850342	0.764259
1	0.967565	0.174887	0.519950	0.450956
2	-0.669433	-0.629411	-1.288723	0.119729
3	0.995928	0.530986	0.398481	2.161532
4	-1.581833	0.758884	-0.018676	-0.205574

X_train1 = df.sample(n = 20,random_state = 0)   
X_train1.head()
	0	1	2	3
26	-0.034815	-0.427423	0.380800	0.365956
86	-1.556810	-2.761708	-0.290238	0.699914
2	-0.669433	-0.629411	-1.288723	0.119729
55	1.223293	0.506993	-0.466281	-0.375706
75	0.802527	-1.057052	0.749458	-0.191170


X_train2 = df.sample(frac = 0.2,random_state = 0)  #random_state一样时，结果一样
X_train2.head()
0	1	2	3
26	-0.034815	-0.427423	0.380800	0.365956
86	-1.556810	-2.761708	-0.290238	0.699914
2	-0.669433	-0.629411	-1.288723	0.119729
55	1.223293	0.506993	-0.466281	-0.375706
75	0.802527	-1.057052	0.749458	-0.191170

X_test = df.drop(index = X_train1.index)
X_test.head()
0	1	2	3
0	0.014483	0.044465	0.850342	0.764259
1	0.967565	0.174887	0.519950	0.450956
3	0.995928	0.530986	0.398481	2.161532
4	-1.581833	0.758884	-0.018676	-0.205574
5	2.639475	-0.835314	-0.252094	-1.401924

2. Sklearn 中的 train_test_split

from sklearn.model_selection import train_test_split
X_train, X_test ,y_train, y_test= train_test_split(X, y,test_size=0.2, rando
m_state = 20, shuffle=True)

X_train.head()
	0	1	2
37	0.709817	0.343003	-0.211449
51	0.109385	0.479072	0.817219
35	0.206000	-0.783487	-0.694645
65	1.697172	-0.728110	0.445701
31	-1.622200	-1.202743	-0.035591

y_train.head()
37    0.069042
51    1.961325
35   -0.379156
65   -1.098712
31   -0.142196
Name: 3, dtype: float64