CODEWITHSUNDEEP
pythonnumpy
user10163680
user10163680

Reputation: 263

np.insert command changing the format of zero and one

I want to insert zero and one at the end of last column of my whole data.

These are actually my labels which I have to convert in one hot vector later. But when I insert them at the end of column and when I print that column it prints as [0., 0., 1.,...] rather it should be [0,0,1,....].

What should I do to overcome this error? Because with 0. and 1. I can't get one hot representation and with these numbers I get error of unsupported iterator index. I need the last column as 0 and 1s.

Can someone please guide me what should I do to fix this error? 

def set_label(data_all, Labels, hc):
    loc_1=len(data_all[0])
    labels= Labels
    insert_h = np.insert(data_all,loc_1,labels,axis=1)

    label_store =[]
    for i in range(len(labels)):
        hmax = labels[i]
        if hmax < hc:
            label_store.append(0)
        elif hmax > hc:
            label_store.append(1)
    loc_2= len(insert_h[0])

    insert_label =np.insert(insert_h, loc_2, label_store,axis=1)
    return insert_label

When I print insert_label of above function I get following result. zero and one is appearing as 0.0000e+00 and 1.0000e+00. I want it to be 0 and 1.

[[1.11886399e-01 2.83619339e+00 3.14394233e+001.00000000e+00 0.00000000e+00]
[2.29317271e-01 1.87929947e+00 3.14713896e+00 1.00000000e+00 0.00000000e+00]
[3.69530173e-02 3.52560100e+00 5.05051823e+00 1.50000000e+00 0.00000000e+00]
[1.56559513e-01 2.10847705e+00 3.88715380e+00 1.50000000e+00 0.00000000e+00] 
[2.44876157e-03 6.02824866e+00 1.02350953e+01 5.00000000e+00 1.00000000e+00]
[9.38766706e-03 4.89698950e+00 6.28692713e+00 5.00000000e+00 1.00000000e+00]]

Here in the input data for calling above function.

Labels=[1.  1.  1.5 1.5 5.  5. ]
data_all=[[1.11886399e-01, 2.83619339e+00, 3.14394233e+00]
[2.29317271e-01, 1.87929947e+00, 3.14713896e+00]
[3.69530173e-02, 3.52560100e+00, 5.05051823e+00]
[1.56559513e-01, 2.10847705e+00, 3.88715380e+00]
[2.44876157e-03, 6.02824866e+00, 1.02350953e+01]
[9.38766706e-03, 4.89698950e+00, 6.28692713e+00]]`

Function call "set_label(data_all,Labels,2)

Upvotes: 1

Views: 164

Answers (1)

Javad Sameri
Javad Sameri

Reputation: 1319

you can convert your data type to object like this :

def set_label(data_all, Labels, hc):
    loc_1=len(data_all[0])
    labels= Labels
    insert_h = np.insert(data_all,loc_1,labels,axis=1)

    label_store =[]
    for i in range(len(labels)):
        hmax = labels[i]
        if hmax < hc:
            label_store.append(0)
        elif hmax > hc:
            label_store.append(1)
    loc_2= len(insert_h[0])
    #this is what I added to your code 
    insert_h = np.insert(np.array(insert_h,dtype='O'),loc_2,np.array(labels,dtype = np.int),axis=1)
    return insert_h

and your output would be like this :

array([[0.111886399, 2.83619339, 3.14394233, 1.0, 1],
       [0.229317271, 1.87929947, 3.14713896, 1.0, 1],
       [0.0369530173, 3.525601, 5.05051823, 1.5, 1],
       [0.156559513, 2.10847705, 3.8871538, 1.5, 1],
       [0.00244876157, 6.02824866, 10.2350953, 5.0, 5],
       [0.00938766706, 4.8969895, 6.28692713, 5.0, 5]], dtype=object)

or when you want to select your label and convert it to one hot convert labels type to integer :

np.array('your selected array',dtype = np.int)

Upvotes: 1

Related Questions