One hot code

Fever alone /one-hot It is used to represent a special byte or vector in digital circuits and machine learning ,

Intuitively speaking, there are as many bits as there are states , And only one bit is 1, For all other 0 A code system of .

Only one bit of this byte or vector is allowed to be 1, All other bits must be 0.

It is called the sole heat because there can only be one 1, If the opposite is true , only one 0, Others are 1, Is called alone cold （one-cold）.

In statistics , Dummy variables represent similar concepts .

Usually , In the communication network protocol stack , Use 8-bit or 16 bit state unique hot code , And the system occupies one of the status codes , The rest can be used by users .

import numpy as np
wide = 10
test = np.random.randint(wide,size=5)
print(test)

# raw
def one_hot(num,wide):
    res = np.zeros(wide)
    res[num] = 1
    return res
print(np.array([one_hot(i,wide) for i in test]))

# batch
def one_hots(narr,wide):
    res = np.zeros((narr.size,wide))
    row = np.arange(narr.size)
    res[row, narr] = 1
    return res
print(one_hots(test,wide))

# multi-hot
test = np.array([[1,2,3,4,5],[7,8,9,0,1]])
print(np.array([one_hots(i,wide).sum(axis=0) for i in test]))

# tensorflow
import tensorflow as tf
print(tf.keras.utils.to_categorical(test,num_classes=(wide)))

# ...
from sklearn.preprocessing import LabelBinarizer
print(LabelBinarizer().fit(np.arange(wide)).transform(test))
# OneHotEncoder ...

# pandas
# import pandas as pd
# print(pd.get_dummies(test,wide))

Copy

...