# to_categorical & one hot encoding
###### tags: `python語法`

to_categorical就是將類別向量轉換為二進制（只有0和1）的矩陣類型表示。其表現為將原有的類別向量轉換為one hot encoding的形式

from keras.utils.np_utils import *
#類別向量定義
b = [0,1,2,3,4,5,6,7,8]
#使用to_categorical將b按照9個類别来進行轉換
b = to_categorical(b, 9)
print(b)
 
執行结果如下：
[[1. 0. 0. 0. 0. 0. 0. 0. 0.]
 [0. 1. 0. 0. 0. 0. 0. 0. 0.]
 [0. 0. 1. 0. 0. 0. 0. 0. 0.]
 [0. 0. 0. 1. 0. 0. 0. 0. 0.]
 [0. 0. 0. 0. 1. 0. 0. 0. 0.]
 [0. 0. 0. 0. 0. 1. 0. 0. 0.]
 [0. 0. 0. 0. 0. 0. 1. 0. 0.]
 [0. 0. 0. 0. 0. 0. 0. 1. 0.]
 [0. 0. 0. 0. 0. 0. 0. 0. 1.]]
 

one hot encoding

def convert_to_one_hot(labels, num_classes):
    #計算向量有多少行
    num_labels = len(labels)
    #生成值全為0的矩陣
    labels_one_hot = np.zeros((num_labels, num_classes))
    #計算向量中每個類别值在最终生成的矩陣的位置
    index_offset = np.arange(num_labels) * num_classes
    #為每個類别的位置標記上1
    labels_one_hot.flat[index_offset + labels] = 1
    return labels_one_hot
#進行測試
b = [2, 4, 6, 8, 6, 2, 3, 7]
print(convert_to_one_hot(b,9))
 
測試结果：
[[0. 0. 1. 0. 0. 0. 0. 0. 0.]
 [0. 0. 0. 0. 1. 0. 0. 0. 0.]
 [0. 0. 0. 0. 0. 0. 1. 0. 0.]
 [0. 0. 0. 0. 0. 0. 0. 0. 1.]
 [0. 0. 0. 0. 0. 0. 1. 0. 0.]
 [0. 0. 1. 0. 0. 0. 0. 0. 0.]
 [0. 0. 0. 1. 0. 0. 0. 0. 0.]
 [0. 0. 0. 0. 0. 0. 0. 1. 0.]]