Get a numpy arrays indices in order of occurrence

Question

I am using train_test_split. My training set, X[], is an array of file paths. Then I have another array y[] that is composed on one hot encoded labels. They are related by the array row index. So if I pass X it looks like this:

Index path

4, data\djip2\DJIP2.5844MHz.10MSPS.fc32.2016-07-01_000000000001.npy

20, data\taigentank\USRP-2_420GHz-1MSps-1MHztaigentank1_000000000000.npy

2, data\866_300_1\USRP-866_300MHz-1MSps-1MHz_lte_1_000000000002.npy

And y[] looks like this:

Index label

4, 00000001

20, 00000010

2, 01000000

These arrays are then passed to a batch generator after being randomized. At the batch generator I need to make sure each X array value can be mapped back to a y[] array label.

So, I want to be able to get the X array indexes which are now in a random order like:

2, path

4, path

20, path

And pass them to another function in this order. I need the indices because I need to pass the path as well as its associated label. Is there a simple way to do this with numpy?

FranklynJey · Accepted Answer · 2019-11-11 07:37:15Z

1

One solution might be:

n = range(numberOfInstances)

which creates a list of integers like [0,1,2,3,4...numberOfInstances-1]. Then shuffle the list

random.shuffle(n)

Save this list as a numpy array

n_np = np.array(n)

and finally reorder your data and groundtruth accordingly like

y = y[n_np]
x = x[n_np]

which should reorder both arrays in regarding the same way.

I hoipe this helps :)! I am a little bit confused why you have an unordered sequence of IDs which somehow are shuffled again....

answered Nov 11, 2019 at 7:37

FranklynJey

6765 silver badges9 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Robi Sen Over a year ago

The reason is I am doing this labelbinarizer = LabelBinarizer() X = train.path y = labelbinarizer.fit_transform(train.word) X, Xt, y, yt = train_test_split(X, y, test_size=0.3, stratify=y) -- where X is a array of paths and Y is array of labels. The split funcation associates the right labels to each path and makes a random dataset so on each epoc the samples come in at diffrent orders to

Collectives™ on Stack Overflow

Get a numpy arrays indices in order of occurrence

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related