CODEWITHSUNDEEP
pythonnumpymachine-learning
user5487286
user5487286

Reputation:

Train/Test Split Python

There are 250 randomly generated data points that are obtained as follows:

[X, y] = getDataSet()  # getDataSet() randomly generates 250 data points

X looks like:

[array([[-2.44141527e-01, 8.39016956e-01],
        [ 1.37468561e+00, 4.97114860e-01],
        [ 3.08071887e-02, -2.03260255e-01],...

While y looks like:

y is array([[0.],
            [0.],
            [0.],...

(it also contains 1s)

So, I'm trying to split [X, y] into training and testing sets. The training set is suppose to be a random selection of 120 of the randomly generated data points. Here is how I'm generating the training set:

nTrain = 120

maxIndex = len(X)
randomTrainingSamples = np.random.choice(maxIndex, nTrain, replace=False)
trainX = X[randomTrainingSamples, :]  # training samples
trainY = y[randomTrainingSamples, :]  # labels of training samples    nTrain X 1

Now, what I can't seem to figure out is, how to get the testing set, which is the 130 other randomly generated data points that are not included in the training set:

testX =  # testing samples
testY =  # labels of testing samples nTest x 1

Suggestions are much appreciated. Thank you!

Upvotes: 0

Views: 1364

Answers (3)

feed liu
feed liu

Reputation: 36

You can try this.

randomTestingSamples = [i for i in range(maxIndex) if i not in randomTrainingSamples]
testX =  X[randomTestingSamples, :]  # testing samples
testY =  y[randomTestingSamples, :]  # labels of testing samples nTest x 1

Upvotes: 1

Ernest S Kirubakaran
Ernest S Kirubakaran

Reputation: 1564

You can shuffle the index and pick the first 120 as train and the next 130 as test

random_index = np.random.shuffle(np.arange(len(X)))
randomTrainingSamples = random_index[:120]
randomTestSamples = random_index[120:250]

trainX = X[randomTrainingSamples, :] 
trainY = y[randomTrainingSamples, :] 

testX = X[randomTestSamples, :]
testY = y[randomTestSamples, :]

Upvotes: 0

Chris
Chris

Reputation: 29742

You can use sklearn.model_selection.train_test_split:

import numpy as np
from sklearn.model_selection import train_test_split

X, y = np.ndarray((250, 2)), np.ndarray((250, 1))

trainX, testX, trainY, testY = train_test_split(X, y, test_size= 130)

trainX.shape
# (120, 2)
testX.shape
# (130, 2)
trainY.shape
# (120, 1)
testY.shape
# (130, 1)

Upvotes: 3

Related Questions