NumPy 是 Python 科學計算的基礎軟件包,
提供多維數組對象,多種派生對象(掩碼數組、矩陣等)以及用於快速操作數組的函數及 API,它包括數學、邏輯、數組形狀變換、排序、選擇、I/O 、離散傅立葉變換、基本線性代數、基本統計運算、隨機模擬等等。
70道NumPy測試題~
1. 將 NumPy 導入為 np,並查看版本
問題:將 NumPy 導入為 np,並輸出版本號。

1 import numpy as np 2 print(np.__version__) 3 #> 1.13.3
2. 如何創建 1 維數組?
問題:創建數字從 0 到 9 的 1 維數組。
期望輸出:#> array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

1 import numpy as np 2 3 arr = np.arange(10) 4 print(arr)
3. 如何創建 boolean 數組?
問題:創建所有 True 的 3×3 NumPy 數組。

1 import numpy as np 2 3 ns1 = np.full((3, 3), True, dtype=bool) 4 #> array([[ True, True, True], 5 #> [ True, True, True], 6 #> [ True, True, True]], dtype=bool) 7 print(ns1) 8 9 # Alternate method: 10 ns = np.ones((3,3), dtype=bool) 11 print(ns) 12
4. 如何從 1 維數組中提取滿足給定條件的項?
問題:從 arr 中提取所有奇數。
輸入: >>arr = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])`
期望輸出:>>#> array([1, 3, 5, 7, 9])

1 import numpy as np 2 3 arr = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) 4 ns = arr[arr % 2 == 1] 5 print(ns)
5. 如何將 NumPy 數組中滿足給定條件的項替換成另一個數值?
問題:將 arr 中的所有奇數替換成 -1。
輸入:arr = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
期望輸出:#> array([ 0, -1, 2, -1, 4, -1, 6, -1, 8, -1])

1 import numpy as np 2 3 arr = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) 4 arr[arr % 2 == 1] = -1 5 print(arr) 6
6. 如何在不影響原始數組的前提下替換滿足給定條件的項?
問題:將 arr 中所有奇數替換成 -1,且不改變 arr。
輸入:arr = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
期望輸出:out
#> array([ 0, -1, 2, -1, 4, -1, 6, -1, 8, -1])
arr
#> array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

1 import numpy as np 2 3 arr = np.arange(10) 4 out = np.where(arr % 2 == 1, -1, arr) 5 print(arr) 6 print(out)
7. 如何重塑(reshape)數組?
問題:將 1 維數組轉換成 2 維數組(兩行)。
輸入: np.arange(10)
array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
期望輸出:#> array([[0, 1, 2, 3, 4],
#> [5, 6, 7, 8, 9]])

1 import numpy as np 2 3 arr = np.arange(10) 4 arr.reshape(2, -1) 5 print(arr)
8. 如何垂直堆疊兩個數組?
問題:垂直堆疊數組 a 和 b。
輸入:a = np.arange(10).reshape(2,-1)
b = np.repeat(1, 10).reshape(2,-1)
期望輸出:#> array([[0, 1, 2, 3, 4],
#> [5, 6, 7, 8, 9],
#> [1, 1, 1, 1, 1],
#> [1, 1, 1, 1, 1]])

1 import numpy as np 2 3 a = np.arange(10).reshape(2,-1) 4 b = np.repeat(1, 10).reshape(2,-1) 5 6 r1 = np.concatenate([a, b], axis=0) # 方法一 7 r2 = np.vstack([a, b]) # 方法二 8 r3 = np.r_[a, b] # 方法三 9 10 print(r1) 11 print(r2) 12 print(r3) 13
9. 如何水平堆疊兩個數組?
問題:水平堆疊數組 a 和 b。
輸入:a = np.arange(10).reshape(2,-1)
b = np.repeat(1, 10).reshape(2,-1)
期望輸出:#> array([[0, 1, 2, 3, 4, 1, 1, 1, 1, 1],
#> [5, 6, 7, 8, 9, 1, 1, 1, 1, 1]])

1 import numpy as np 2 3 a = np.arange(10).reshape(2,-1) 4 b = np.repeat(1, 10).reshape(2,-1) 5 6 r1 = np.concatenate([a, b], axis=1) # 方法一 7 r2 = np.hstack([a, b]) # 方法二 8 r3 = np.c_[a, b] # 方法三 9 print(r1) 10 print(r2) 11 print(r3)
10. 在不使用硬編碼的前提下,如何在 NumPy 中生成自定義序列?
問題:在不使用硬編碼的前提下創建以下模式。僅使用 NumPy 函數和以下輸入數組 a。
輸入:a = np.array([1,2,3])`
期望輸出:#> array([1, 1, 1, 2, 2, 2, 3, 3, 3, 1, 2, 3, 1, 2, 3, 1, 2, 3])

1 import numpy as np 2 3 a = np.array([1,2,3]) 4 r1 = np.r_[np.repeat(a, 3), np.tile(a, 3)] 5 print(r1)
11. 如何獲得兩個 Python NumPy 數組中共同的項?
問題:獲取數組 a 和 b 中的共同項。
輸入:a = np.array([1,2,3,2,3,4,3,4,5,6])
b = np.array([7,2,10,2,7,4,9,4,9,8])
期望輸出:array([2, 4])

1 import numpy as np 2 3 a = np.array([1,2,3,2,3,4,3,4,5,6]) 4 b = np.array([7,2,10,2,7,4,9,4,9,8]) 5 r1 = np.intersect1d(a,b) 6 print(r1)
12. 如何從一個數組中移除與另一個數組重復的項?
問題:從數組 a 中移除出現在數組 b 中的所有項。
輸入:a = np.array([1,2,3,4,5])
b = np.array([5,6,7,8,9])
期望輸出:array([1,2,3,4])

1 import numpy as np 2 3 a = np.array([1,2,3,4,5]) 4 b = np.array([5,6,7,8,9]) 5 r1 = np.setdiff1d(a,b) 6 print(r1)
13. 如何獲取兩個數組匹配元素的位置?
問題:獲取數組 a 和 b 中匹配元素的位置。
輸入:a = np.array([1,2,3,2,3,4,3,4,5,6])
b = np.array([7,2,10,2,7,4,9,4,9,8])
期望輸出:#> (array([1, 3, 5, 7]),)

1 import numpy as np 2 3 a = np.array([1,2,3,2,3,4,3,4,5,6]) 4 b = np.array([7,2,10,2,7,4,9,4,9,8]) 5 6 r1 = np.where(a == b) 7 print(r1)
14. 如何從 NumPy 數組中提取給定范圍內的所有數字?
問題:從數組 a 中提取 5 和 10 之間的所有項。
輸入:a = np.arange(15)
期望輸出:(array([ 5, 6, 7, 8, 9, 10]),)

1 import numpy as np 2 3 a = np.arange(15) 4 5 index = np.where((a >= 5) & (a <= 10)) # 方法一 6 r1 = a[index] 7 8 index = np.where(np.logical_and(a >= 5, a <= 10)) # 方法二 9 r2 = a[index] 10 11 r3 = a[(a >= 5) & (a <= 10)] # 方法三 12 print(r1) 13 print(r2) 14 print(r3) 15
15. 如何創建一個 Python 函數以對 NumPy 數組執行元素級的操作?
問題:轉換函數 maxx,使其從只能對比標量而變為對比兩個數組。
輸入:
期望輸出:a = np.array([5, 7, 9, 8, 6, 4, 5])
b = np.array([6, 3, 4, 8, 9, 7, 1])
pair_max(a, b)
#> array([ 6., 7., 9., 8., 9., 7., 5.])

1 import numpy as np 2 3 4 def maxx(x, y): 5 """Get the maximum of two items""" 6 if x >= y: 7 return x 8 else: 9 return y 10 11 12 pair_max = np.vectorize(maxx, otypes=[float]) 13 14 a = np.array([5, 7, 9, 8, 6, 4, 5]) 15 b = np.array([6, 3, 4, 8, 9, 7, 1]) 16 17 r1 = pair_max(a, b) 18 print(r1)
16. 如何在 2d NumPy 數組中交換兩個列?
問題:在數組 arr 中交換列 1 和列 2。
arr = np.arange(9).reshape(3,3)
arr

1 import numpy as np 2 3 arr = np.arange(9).reshape(3,3) 4 r1 = arr[:, [1,0,2]] 5 print(r1)
17. 如何在 2d NumPy 數組中交換兩個行?
問題:在數組 arr 中交換行 1 和行 2。
arr = np.arange(9).reshape(3,3)
arr

1 import numpy as np 2 3 arr = np.arange(9).reshape(3, 3) 4 r1 = arr[[1, 0, 2], :] 5 print(r1) 6
18. 如何反轉 2D 數組的所有行?
問題:反轉 2D 數組 arr 中的所有行。
# Input
arr = np.arange(9).reshape(3,3)

1 import numpy as np 2 3 arr = np.arange(9).reshape(3,3) 4 r1 = arr[::-1] 5 print(r1)
19. 如何反轉 2D 數組的所有列?
問題:反轉 2D 數組 arr 中的所有列。
# Input
arr = np.arange(9).reshape(3,3)

1 import numpy as np 2 3 arr = np.arange(9).reshape(3,3) 4 r1 = arr[:, ::-1] 5 print(r1)
20. 如何創建一個包含 5 和 10 之間隨機浮點的 2 維數組?
問題:創建一個形態為 5×3 的 2 維數組,包含 5 和 10 之間的隨機十進制小數。

1 import numpy as np 2 3 arr = np.arange(9).reshape(3,3) 4 # Solution Method 1: 5 rand_arr = np.random.randint(low=5, high=10, size=(5,3)) + np.random.random((5,3)) 6 print("1",rand_arr) 7 8 # Solution Method 2: 9 rand_arr = np.random.uniform(5,10, size=(5,3)) 10 print("2",rand_arr)
21. 如何在 Python NumPy 數組中僅輸出小數點后三位的數字?
問題:輸出或顯示 NumPy 數組 rand_arr 中小數點后三位的數字。
輸入:rand_arr = np.random.random((5,3))

1 import numpy as np 2 3 rand_arr = np.random.random((5,3)) 4 # Create the random array 5 rand_arr = np.random.random([5,3]) 6 # Limit to 3 decimal places 7 np.set_printoptions(precision=3) 8 rand_arr[:4] 9 print(rand_arr) 10
22. 如何通過禁用科學計數法(如 1e10)打印 NumPy 數組?
問題:通過禁用科學計數法(如 1e10)打印 NumPy 數組 rand_arr。
輸入# Create the random array
np.random.seed(100)
rand_arr = np.random.random([3,3])/1e3
rand_arr
期望輸出:#> array([[ 0.000543, 0.000278, 0.000425],
#> [ 0.000845, 0.000005, 0.000122],
#> [ 0.000671, 0.000826, 0.000137]])

1 import numpy as np 2 3 np.set_printoptions(suppress=False) 4 # Create the random array 5 np.random.seed(100) 6 rand_arr = np.random.random([3,3])/1e3 7 8 print(rand_arr)
23. 如何限制 NumPy 數組輸出中項的數目?
問題:將 Python NumPy 數組 a 輸出的項的數目限制在最多 6 個元素。
輸入:a = np.arange(15)
期望輸出:#> array([ 0, 1, 2, ..., 12, 13, 14])

1 mport numpy as np 2 3 np.set_printoptions(threshold=6) 4 a = np.arange(15) 5 6 print(a)
24. 如何在不截斷數組的前提下打印出完整的 NumPy 數組?
問題:在不截斷數組的前提下打印出完整的 NumPy 數組 a。
輸入:np.set_printoptions(threshold=6)
a = np.arange(15)
期望輸出:a
#> array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14])
25. 如何向 Python NumPy 導入包含數字和文本的數據集,同時保持文本不變?
問題:導入 iris 數據集,保持文本不變。

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 names = ('sepallength', 'sepalwidth', 'petallength', 'petalwidth', 'species') 6 7 r1 = iris[:3] # Print the first 3 rows 8 print(r1)
26. 如何從 1 維元組數組中提取特定的列?
問題:從前一個問題導入的 1 維 iris 中提取文本列 species。
輸入:rl = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_1d = np.genfromtxt(url, delimiter= , , dtype=None)

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris_1d = np.genfromtxt(url, delimiter=',', dtype=None) 5 print(iris_1d.shape) 6 7 species = np.array([row[4] for row in iris_1d]) 8 r1 = species[:5] 9 print(r1) 10
27. 如何將 1 維元組數組轉換成 2 維 NumPy 數組?
問題:忽略 species 文本字段,將 1 維 iris 轉換成 2 維數組 iris_2d。
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_1d = np.genfromtxt(url, delimiter= , , dtype=None)

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris_1d = np.genfromtxt(url, delimiter=',', dtype=None) 5 6 iris_2d = np.array([row.tolist()[:4] for row in iris_1d]) #方法一 7 r1 = iris_2d[:4] 8 9 iris_2d = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0,1,2,3]) #方法二 10 r2 = iris_2d[:4] 11 print(r1) 12 print(r2)
28. 如何計算 NumPy 數組的平均值、中位數和標准差?
問題:找出 iris sepallength(第一列)的平均值、中位數和標准差。
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris = np.genfromtxt(url, delimiter= , , dtype= object )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 sepallength = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0]) 6 7 mu, med, sd = np.mean(sepallength), np.median(sepallength), np.std(sepallength) 8 print(mu, med, sd)
29. 如何歸一化數組,使值的范圍在 0 和 1 之間?
問題:創建 iris sepallength 的歸一化格式,使其值在 0 到 1 之間。
輸入:url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
sepallength = np.genfromtxt(url, delimiter= , , dtype= float , usecols=[0])

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 sepallength = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0]) 5 6 Smax, Smin = sepallength.max(), sepallength.min() 7 S1 = (sepallength - Smin)/(Smax - Smin) 8 S2 = (sepallength - Smin)/sepallength.ptp() # Thanks, David Ojeda! 9 print(S1) 10 print(S2)
30. 如何計算 softmax 分數?
問題:計算 sepallength 的 softmax 分數。
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
sepallength = np.genfromtxt(url, delimiter= , , dtype= float , usecols=[0])

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 sepallength = np.array([float(row[0]) for row in iris]) 6 7 def softmax(x): 8 """Compute softmax values for each sets of scores in x. 9 https://stackoverflow.com/questions/34968722/how-to-implement-the-softmax-function-in-python""" 10 e_x = np.exp(x - np.max(x)) 11 return e_x / e_x.sum(axis=0) 12 13 r1 = softmax(sepallength) 14 print(r1) 15 16
31. 如何找到 NumPy 數組的百分數?
問題:找出 iris sepallength(第一列)的第 5 個和第 95 個百分數。
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
sepallength = np.genfromtxt(url, delimiter= , , dtype= float , usecols=[0])

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 sepallength = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0]) 5 r1 = np.percentile(sepallength, q=[5, 95]) 6 print(r1)
32. 如何在數組的隨機位置插入值?
問題:在 iris_2d 數據集中的 20 個隨機位置插入 np.nan 值。
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_2d = np.genfromtxt(url, delimiter= , , dtype= object )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris_2d = np.genfromtxt(url, delimiter=',', dtype='object') 5 6 i, j = np.where(iris_2d) 7 np.random.seed(100) 8 iris_2d[np.random.choice((i), 20), np.random.choice((j), 20)] = np.nan 9 10 # Method 2 11 np.random.seed(100) 12 iris_2d[np.random.randint(150, size=20), np.random.randint(4, size=20)] = np.nan 13 14 # Print first 10 rows 15 print(iris_2d[:10])
33. 如何在 NumPy 數組中找出缺失值的位置?
問題:在 iris_2d 的 sepallength(第一列)中找出缺失值的數目和位置。
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_2d = np.genfromtxt(url, delimiter= , , dtype= float )
iris_2d[np.random.randint(150, size=20), np.random.randint(4, size=20)] = np.nan

1 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 2 iris_2d = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0,1,2,3]) 3 iris_2d[np.random.randint(150, size=20), np.random.randint(4, size=20)] = np.nan 4 5 print("Number of missing values: \n", np.isnan(iris_2d[:, 0]).sum()) 6 print("Position of missing values: \n", np.where(np.isnan(iris_2d[:, 0])))
34. 如何基於兩個或以上條件過濾 NumPy 數組?
問題:過濾 iris_2d 中滿足 petallength(第三列)> 1.5 和 sepallength(第一列)< 5.0 的行。
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_2d = np.genfromtxt(url, delimiter= , , dtype= float , usecols=[0,1,2,3])

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris_2d = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0,1,2,3]) 5 6 condition = (iris_2d[:, 2] > 1.5) & (iris_2d[:, 0] < 5.0) 7 r1 = iris_2d[condition] 8 print(r1)
35. 如何在 NumPy 數組中刪除包含缺失值的行?
問題:選擇 iris_2d 中不包含 nan 值的行。
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_2d = np.genfromtxt(url, delimiter= , , dtype= float , usecols=[0,1,2,3])

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris_2d = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0,1,2,3]) 5 iris_2d[np.random.randint(150, size=20), np.random.randint(4, size=20)] = np.nan 6 7 # Method 1: 8 any_nan_in_row = np.array([~np.any(np.isnan(row)) for row in iris_2d]) 9 r1 = iris_2d[any_nan_in_row][:5] 10 11 # Method 2: (By Rong) 12 r2 = iris_2d[np.sum(np.isnan(iris_2d), axis = 1) == 0][:5] 13 print(r1) 14 print(r2)
36. 如何找出 NumPy 數組中兩列之間的關聯性?
問題:找出 iris_2d 中 SepalLength(第一列)和 PetalLength(第三列)之間的關聯性。
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_2d = np.genfromtxt(url, delimiter= , , dtype= float , usecols=[0,1,2,3])

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0,1,2,3]) 5 6 r1 = np.corrcoef(iris[:, 0], iris[:, 2])[0, 1] 7 print(r1) 8
37. 如何確定給定數組是否有空值?
問題:確定 iris_2d 是否有缺失值。
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_2d = np.genfromtxt(url, delimiter= , , dtype= float , usecols=[0,1,2,3])

1 import numpy as np 2 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 3 iris_2d = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0,1,2,3]) 4 5 r1 = np.isnan(iris_2d).any() 6 print(r1)
38. 如何在 NumPy 數組中將所有缺失值替換成 0?
問題:在 NumPy 數組中將所有 nan 替換成 0。
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_2d = np.genfromtxt(url, delimiter= , , dtype= float , usecols=[0,1,2,3])
iris_2d[np.random.randint(150, size=20), np.random.randint(4, size=20)] = np.nan

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris_2d = np.genfromtxt(url, delimiter=',', dtype='float', usecols=[0,1,2,3]) 5 iris_2d[np.random.randint(150, size=20), np.random.randint(4, size=20)] = np.nan 6 iris_2d[np.isnan(iris_2d)] = 0 7 r1 = iris_2d[:4] 8 print(r1)
39. 如何在 NumPy 數組中找出唯一值的數量?
問題:在 iris 的 species 列中找出唯一值及其數量。
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris = np.genfromtxt(url, delimiter= , , dtype= object )
names = ( sepallength , sepalwidth , petallength , petalwidth , species )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 names = ('sepallength', 'sepalwidth', 'petallength', 'petalwidth', 'species') 6 species = np.array([row.tolist()[4] for row in iris]) 7 r1 = np.unique(species, return_counts=True) 8 print(r1)
40. 如何將一個數值轉換為一個類別(文本)數組?
問題:將 iris_2d 的 petallength(第三列)轉換以構建一個文本數組,按如下規則進行轉換:
-
Less than 3 –> ‘small’
-
3-5 –> medium
-
>=5 –> large
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris = np.genfromtxt(url, delimiter= , , dtype= object )
names = ( sepallength , sepalwidth , petallength , petalwidth , species )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 names = ('sepallength', 'sepalwidth', 'petallength', 'petalwidth', 'species') 6 7 petal_length_bin = np.digitize(iris[:, 2].astype('float'), [0, 3, 5, 10]) 8 label_map = {1: 'small', 2: 'medium', 3: 'large', 4: np.nan} 9 petal_length_cat = [label_map[x] for x in petal_length_bin] 10 r = petal_length_cat[:4] 11 print(r)
41. 如何基於 NumPy 數組現有列創建一個新的列?
問題:為 iris_2d 中的 volume 列創建一個新的列,volume 指 (pi x petallength x sepal_length^2)/3。
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris_2d = np.genfromtxt(url, delimiter= , , dtype= object )
names = ( sepallength , sepalwidth , petallength , petalwidth , species )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris_2d = np.genfromtxt(url, delimiter=',', dtype='object') 5 6 sepallength = iris_2d[:, 0].astype('float') 7 petallength = iris_2d[:, 2].astype('float') 8 volume = (np.pi * petallength * (sepallength**2))/3 9 volume = volume[:, np.newaxis] 10 out = np.hstack([iris_2d, volume]) 11 12 r1 = out[:4] 13 print(r1)
42. 如何在 NumPy 中執行概率采樣?
問題:隨機采樣 iris 數據集中的 species 列,使得 setose 的數量是 versicolor 和 virginica 數量的兩倍。
# Import iris keeping the text column intact
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris = np.genfromtxt(url, delimiter= , , dtype= object )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 6 species = iris[:, 4] 7 np.random.seed(100) 8 a = np.array(['Iris-setosa', 'Iris-versicolor', 'Iris-virginica']) 9 species_out = np.random.choice(a, 150, p=[0.5, 0.25, 0.25]) 10 np.random.seed(100) 11 probs = np.r_[np.linspace(0, 0.500, num=50), np.linspace(0.501, .750, num=50), np.linspace(.751, 1.0, num=50)] 12 index = np.searchsorted(probs, np.random.random(150)) 13 species_out = species[index] 14 print(np.unique(species_out, return_counts=True)) 15
參考 View Code
43. 如何在多維數組中找到一維的第二最大值?
問題:在 species setosa 的 petallength 列中找到第二最大值。
# Input
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris = np.genfromtxt(url, delimiter= , , dtype= object )
names = ( sepallength , sepalwidth , petallength , petalwidth , species )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 6 petal_len_setosa = iris[iris[:, 4] == b'Iris-setosa', [2]].astype('float') 7 8 r1 = np.unique(np.sort(petal_len_setosa))[-2] 9 print(r1)
44. 如何用給定列將 2 維數組排序?
問題:基於 sepallength 列將 iris 數據集排序。
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris = np.genfromtxt(url, delimiter= , , dtype= object )
names = ( sepallength , sepalwidth , petallength , petalwidth , species )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 names = ('sepallength', 'sepalwidth', 'petallength', 'petalwidth', 'species') 6 7 print(iris[iris[:,0].argsort()][:20])
45. 如何在 NumPy 數組中找到最頻繁出現的值?
問題:在 iris 數據集中找到 petallength(第三列)中最頻繁出現的值。
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris = np.genfromtxt(url, delimiter= , , dtype= object )
names = ( sepallength , sepalwidth , petallength , petalwidth , species )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 vals, counts = np.unique(iris[:, 2], return_counts=True) 6 print(vals[np.argmax(counts)]) 7
參考 View Code
46. 如何找到第一個大於給定值的數的位置?
問題:在 iris 數據集的 petalwidth(第四列)中找到第一個值大於 1.0 的數的位置。
# Input:
url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris = np.genfromtxt(url, delimiter= , , dtype= object )

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 6 r1 = np.argwhere(iris[:, 3].astype(float) > 1.0)[0] 7 print(r1)
參考 View Code
47. 如何將數組中所有大於給定值的數替換為給定的 cutoff 值?
問題:對於數組 a,將所有大於 30 的值替換為 30,將所有小於 10 的值替換為 10。
輸入:np.random.seed(100)
np.random.uniform(1,50, 20)

1 import numpy as np 2 3 np.set_printoptions(precision=2) 4 np.random.seed(100) 5 a = np.random.uniform(1,50, 20) 6 7 np.clip(a, a_min=10, a_max=30) 8 9 print(np.where(a < 10, 10, np.where(a > 30, 30, a)))
48. 如何在 NumPy 數組中找到 top-n 數值的位置?
問題:在給定數組 a 中找到 top-5 最大值的位置。
np.random.seed(100)
a = np.random.uniform(1,50, 20)

1 import numpy as np 2 3 np.random.seed(100) 4 a = np.random.uniform(1,50, 20) 5 6 # Solution: 7 print(a.argsort()) #> [18 7 3 10 15] 8 # Solution 2: 9 print(np.argpartition(-a, 5)[:5]) #> [15 10 3 7 18] 10 11 r1 = a[a.argsort()][-5:] # Method 1: 12 print(r1) 13 14 r2 = np.sort(a)[-5:] # Method 2: 15 print(r2) 16 17 r3 = np.partition(a, kth=-5)[-5:] # Method 3: 18 print(r3) 19 20 r4 = a[np.argpartition(-a, 5)][:5] # Method 4: 21 print(r4)
49. 如何逐行計算數組中所有值的數量?
問題:逐行計算唯一值的數量。
輸入:np.random.seed(100)
arr = np.random.randint(1,11,size=(6, 10))
arr
> array([[ 9, 9, 4, 8, 8, 1, 5, 3, 6, 3],
> [ 3, 3, 2, 1, 9, 5, 1, 10, 7, 3],
> [ 5, 2, 6, 4, 5, 5, 4, 8, 2, 2],
> [ 8, 8, 1, 3, 10, 10, 4, 3, 6, 9],
> [ 2, 1, 8, 7, 3, 1, 9, 3, 6, 2],
> [ 9, 2, 6, 5, 3, 9, 4, 6, 1, 10]])
期望輸出:> [[1, 0, 2, 1, 1, 1, 0, 2, 2, 0],
> [2, 1, 3, 0, 1, 0, 1, 0, 1, 1],
> [0, 3, 0, 2, 3, 1, 0, 1, 0, 0],
> [1, 0, 2, 1, 0, 1, 0, 2, 1, 2],
> [2, 2, 2, 0, 0, 1, 1, 1, 1, 0],
> [1, 1, 1, 1, 1, 2, 0, 0, 2, 1]]
輸出包含 10 個列,表示從 1 到 10 的數字。這些數值分別代表每一行的計數數量。例如,Cell(0,2) 中有值 2,這意味着,數字 3 在第一行出現了兩次。

1 import numpy as np 2 3 np.random.seed(100) 4 arr = np.random.randint(1,11,size=(6, 10)) 5 print(arr)
50. 如何將 array_of_arrays 轉換為平面 1 維數組?
問題:將 array_of_arrays 轉換為平面線性 1 維數組。
# Input:
arr1 = np.arange(3)
arr2 = np.arange(3,7)
arr3 = np.arange(7,10)
array_of_arrays = np.array([arr1, arr2, arr3])
array_of_arrays#> array([array([0, 1, 2]), array([3, 4, 5, 6]), array([7, 8, 9])], dtype=object)
期望輸出:#> array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

1 import numpy as np 2 3 arr1 = np.arange(3) 4 arr2 = np.arange(3,7) 5 arr3 = np.arange(7,10) 6 7 array_of_arrays = np.array([arr1, arr2, arr3]) 8 print('array_of_arrays: ', array_of_arrays) 9 10 arr_2d = np.array([a for arr in array_of_arrays for a in arr]) 11 12 arr_2d = np.concatenate(array_of_arrays) 13 print(arr_2d)
51. 如何為 NumPy 數組生成>
問題:計算>
輸入:np.random.seed(101)
arr = np.random.randint(1,4, size=6)
arr
#> array([2, 3, 2, 2, 2, 1])
輸出:> array([[ 0., 1., 0.],
#> [ 0., 0., 1.],
#> [ 0., 1., 0.],
#> [ 0., 1., 0.],
#> [ 0., 1., 0.],
#> [ 1., 0., 0.]])

1 import numpy as np 2 3 np.random.seed(101) 4 arr = np.random.randint(1,4, size=6) 5 print(arr) 6 #> array([2, 3, 2, 2, 2, 1]) 7 8 def one_hot_encodings(arr): 9 uniqs = np.unique(arr) 10 out = np.zeros((arr.shape[0], uniqs.shape[0])) 11 for i, k in enumerate(arr): 12 out[i, k-1] = 1 13 return out 14 15 r1 = one_hot_encodings(arr) 16 print("r1",r1) 17 r2 = (arr[:, None] == np.unique(arr)).view(np.int8) 18 print("r2",r2)
52. 如何創建由類別變量分組確定的一維數值?
問題:創建由類別變量分組的行數。使用以下來自 iris species 的樣本作為輸入。
輸入:url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
species = np.genfromtxt(url, delimiter= , , dtype= str , usecols=4)
species_small = np.sort(np.random.choice(species, size=20))
species_small
#> array([ Iris-setosa , Iris-setosa , Iris-setosa , Iris-setosa ,
#> Iris-setosa , Iris-setosa , Iris-versicolor , Iris-versicolor ,
#> Iris-versicolor , Iris-versicolor , Iris-versicolor ,
#> Iris-versicolor , Iris-virginica , Iris-virginica ,
#> Iris-virginica , Iris-virginica , Iris-virginica ,
#> Iris-virginica , Iris-virginica , Iris-virginica ],
#> dtype= <U15 )
期望輸出:#> [0, 1, 2, 3, 4, 5, 0, 1, 2, 3, 4, 5, 0, 1, 2, 3, 4, 5, 6, 7]

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 species = np.genfromtxt(url, delimiter=',', dtype='str', usecols=4) 5 np.random.seed(100) 6 species_small = np.sort(np.random.choice(species, size=20)) 7 print(species_small) 8 print([i for val in np.unique(species_small) for i, grp in enumerate(species_small[species_small==val])])
53. 如何基於給定的類別變量創建分組 id?
問題:基於給定的類別變量創建分組 id。使用以下來自 iris species 的樣本作為輸入。
輸入:url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
species = np.genfromtxt(url, delimiter= , , dtype= str , usecols=4)
species_small = np.sort(np.random.choice(species, size=20))
species_small
#> array([ Iris-setosa , Iris-setosa , Iris-setosa , Iris-setosa ,
#> Iris-setosa , Iris-setosa , Iris-versicolor , Iris-versicolor ,
#> Iris-versicolor , Iris-versicolor , Iris-versicolor ,
#> Iris-versicolor , Iris-virginica , Iris-virginica ,
#> Iris-virginica , Iris-virginica , Iris-virginica ,
#> Iris-virginica , Iris-virginica , Iris-virginica ],
#> dtype= <U15 )
期望輸出:#> [0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2]

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 species = np.genfromtxt(url, delimiter=',', dtype='str', usecols=4) 5 np.random.seed(100) 6 species_small = np.sort(np.random.choice(species, size=20)) 7 print(species_small) 8 9 output = [np.argwhere(np.unique(species_small) == s).tolist()[0][0] for val in np.unique(species_small) for s in species_small[species_small==val]] 10 11 # Solution: For Loop version 12 output = [] 13 uniqs = np.unique(species_small) 14 15 for val in uniqs: # uniq values in group 16 for s in species_small[species_small==val]: # each element in group 17 groupid = np.argwhere(uniqs == s).tolist()[0][0] # groupid 18 output.append(groupid) 19 20 print(output)
54. 如何使用 NumPy 對數組中的項進行排序?
問題:為給定的數值數組 a 創建排序。
輸入:np.random.seed(10)
a = np.random.randint(20, size=10)print(a)#> [ 9 4 15 0 17 16 17 8 9 0]
期望輸出:[4 2 6 0 8 7 9 3 5 1]

1 import numpy as np 2 3 np.random.seed(10) 4 a = np.random.randint(20, size=10) 5 print('Array: ', a) 6 7 print(a.argsort().argsort()) 8 print('Array: ', a)
55. 如何使用 NumPy 對多維數組中的項進行排序?
問題:給出一個數值數組 a,創建一個形態相同的排序數組。
輸入:np.random.seed(10)
a = np.random.randint(20, size=[2,5])print(a)#> [[ 9 4 15 0 17]#> [16 17 8 9 0]]
期望輸出:#> [[4 2 6 0 8]
#> [7 9 3 5 1]]

1 import numpy as np 2 3 np.random.seed(10) 4 a = np.random.randint(20, size=[2,5]) 5 print(a) 6 7 print(a.ravel().argsort().argsort().reshape(a.shape))
56. 如何在 2 維 NumPy 數組中找到每一行的最大值?
問題:在給定數組中找到每一行的最大值。
np.random.seed(100)
a = np.random.randint(1,10, [5,3])
a
#> array([[9, 9, 4],
#> [8, 8, 1],
#> [5, 3, 6],
#> [3, 3, 3],
#> [2, 1, 9]])

1 import numpy as np 2 3 np.random.seed(100) 4 a = np.random.randint(1,10, [5,3]) 5 print("a=",a) 6 7 r1 = np.amax(a, axis=1) 8 print("r1=",r1) 9 r2 = np.apply_along_axis(np.max, arr=a, axis=1) 10 print("r2",r2) 11
57. 如何計算 2 維 NumPy 數組每一行的 min-by-max?
問題:給定一個 2 維 NumPy 數組,計算每一行的 min-by-max。
np.random.seed(100)
a = np.random.randint(1,10, [5,3])
a
#> array([[9, 9, 4],
#> [8, 8, 1],
#> [5, 3, 6],
#> [3, 3, 3],
#> [2, 1, 9]])

1 import numpy as np 2 3 np.random.seed(100) 4 a = np.random.randint(1,10, [5,3]) 5 print("a=",a) 6 7 8 r1 = np.apply_along_axis(lambda x: np.min(x)/np.max(x), arr=a, axis=1) 9 print("r1",r1) 10
58. 如何在 NumPy 數組中找到重復條目?
問題:在給定的 NumPy 數組中找到重復條目(從第二次出現開始),並將其標記為 True。第一次出現的條目需要標記為 False。
# Input
np.random.seed(100)
a = np.random.randint(0, 5, 10)
print( Array: , a)
#> Array: [0 0 3 0 2 4 2 2 2 2]
期望輸出:#> [False True False True False False True True True True]

1 import numpy as np 2 3 np.random.seed(100) 4 a = np.random.randint(0, 5, 10) 5 out = np.full(a.shape[0], True) 6 unique_positions = np.unique(a, return_index=True)[1] 7 out[unique_positions] = False 8 9 print(out)
59. 如何找到 NumPy 的分組平均值?
問題:在 2 維 NumPy 數組的類別列中找到數值的平均值。
輸入url = https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
iris = np.genfromtxt(url, delimiter= , , dtype= object )
names = ( sepallength , sepalwidth , petallength , petalwidth , species )
期望解:#> [[b Iris-setosa , 3.418],
#> [b Iris-versicolor , 2.770],
#> [b Iris-virginica , 2.974]]

1 import numpy as np 2 3 url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data' 4 iris = np.genfromtxt(url, delimiter=',', dtype='object') 5 names = ('sepallength', 'sepalwidth', 'petallength', 'petalwidth', 'species') 6 7 numeric_column = iris[:, 1].astype('float') # sepalwidth 8 grouping_column = iris[:, 4] # species 9 10 11 [[group_val, numeric_column[grouping_column==group_val].mean()] for group_val in np.unique(grouping_column)] 12 13 output = [] 14 for group_val in np.unique(grouping_column): 15 output.append([group_val, numeric_column[grouping_column==group_val].mean()]) 16 17 r1 = output 18 print(r1)
60. 如何將 PIL 圖像轉換成 NumPy 數組?
問題:從以下 URL 中導入圖像,並將其轉換成 NumPy 數組。
URL = https://upload.wikimedia.org/wikipedia/commons/8/8b/Denali_Mt_McKinley.jpg

1 import numpy as np 2 from io import BytesIO 3 from PIL import Image 4 import PIL, requests 5 6 # Import image from URL 7 URL = 'https://upload.wikimedia.org/wikipedia/commons/8/8b/Denali_Mt_McKinley.jpg' 8 response = requests.get(URL) 9 10 I = Image.open(BytesIO(response.content)) # Read it as Image 11 I = I.resize([150,150]) # Optionally resize 12 arr = np.asarray(I) # Convert to numpy array 13 14 # Optionaly Convert it back to an image and show 15 im = PIL.Image.fromarray(np.uint8(arr)) 16 r1 = Image.Image.show(im) 17 print(r1) 18
61. 如何刪除 NumPy 數組中所有的缺失值?
問題:從 1 維 NumPy 數組中刪除所有的 nan 值。
輸入:np.array([1,2,3,np.nan,5,6,7,np.nan])
期望輸出:array([ 1., 2., 3., 5., 6., 7.])

1 import numpy as np 2 3 a = np.array([1, 2, 3, np.nan, 5, 6, 7, np.nan]) 4 r1 = a[~np.isnan(a)] 5 print(r1) 6
62. 如何計算兩個數組之間的歐幾里得距離?
問題:計算兩個數組 a 和 b 之間的歐幾里得距離。
輸入:a = np.array([1,2,3,4,5])
b = np.array([4,5,6,7,8])

1 import numpy as np 2 3 a = np.array([1,2,3,4,5]) 4 b = np.array([4,5,6,7,8]) 5 dist = np.linalg.norm(a-b) 6 print(dist)
63. 如何在一個 1 維數組中找到所有的局部極大值(peak)?
問題:在 1 維數組 a 中找到所有的 peak,peak 指一個數字比兩側的數字都大。
輸入:a = np.array([1, 3, 7, 1, 2, 6, 0, 1])
期望輸出:#> array([2, 5])

1 import numpy as np 2 3 a = np.array([1, 3, 7, 1, 2, 6, 0, 1]) 4 doublediff = np.diff(np.sign(np.diff(a))) 5 peak_locations = np.where(doublediff == -2)[0] + 1 6 print(peak_locations)
64. 如何從 2 維數組中減去 1 維數組,從 2 維數組的每一行分別減去 1 維數組的每一項?
問題:從 2 維數組 a_2d 中減去 1 維數組 b_1d,即從 a_2d 的每一行分別減去 b_1d 的每一項。
輸入:a_2d = np.array([[3,3,3],[4,4,4],[5,5,5]])
b_1d = np.array([1,1,1]
期望輸出:#> [[2 2 2]
#> [2 2 2]
#> [2 2 2]]

1 import numpy as np 2 3 a_2d = np.array([[3,3,3],[4,4,4],[5,5,5]]) 4 b_1d = np.array([1,2,3]) 5 6 print(a_2d - b_1d[:,None])
65. 如何在數組中找出某個項的第 n 個重復索引?
問題:找到數組 x 中數字 1 的第 5 個重復索引。
x = np.array([1, 2, 1, 1, 3, 4, 3, 1, 1, 2, 1, 1, 2])

1 import numpy as np 2 3 x = np.array([1, 2, 1, 1, 3, 4, 3, 1, 1, 2, 1, 1, 2]) 4 n = 5 5 6 [i for i, v in enumerate(x) if v == 1][n-1] 7 r1 = np.where(x == 1)[0][n-1] 8 print(r1)
66. 如何將 NumPy 的 datetime64 對象(object)轉換為 datetime 的 datetime 對象?
問題:將 NumPy 的 datetime64 對象(object)轉換為 datetime 的 datetime 對象。
Input: a numpy datetime64 object
dt64 = np.datetime64( 2018-02-25 22:10:10 )

1 import numpy as np 2 3 dt64 = np.datetime64('2018-02-25 22:10:10') 4 5 from datetime import datetime 6 r1 = dt64.tolist() 7 print(r1) 8 9 r2 = dt64.astype(datetime) 10 print(r2)
67. 如何計算 NumPy 數組的移動平均數?
問題:給定 1 維數組,計算 window size 為 3 的移動平均數。
輸入:np.random.seed(100)
Z = np.random.randint(10, size=10)

1 import numpy as np 2 3 def moving_average(a, n=3): 4 ret = np.cumsum(a, dtype=float) 5 ret[n:] = ret[n:] - ret[:-n] 6 return ret[n - 1:] / n 7 8 np.random.seed(100) 9 Z = np.random.randint(10, size=10) 10 print('array: ', Z) 11 12 r1 = moving_average(Z, n=3).round(2) 13 print("r1=", r1) 14 15 r2 = np.convolve(Z, np.ones(3) / 3, mode='valid') 16 print("r2=", r2) 17
68. 給定起始數字、length 和步長,如何創建一個 NumPy 數組序列?
問題:從 5 開始,創建一個 length 為 10 的 NumPy 數組,相鄰數字的差是 3。

1 import numpy as np 2 3 length = 10 4 start = 5 5 step = 3 6 7 def seq(start, length, step): 8 end = start + (step*length) 9 return np.arange(start, end, step) 10 11 r1 = seq(start, length, step) 12 print(r1)
69. 如何在不規則 NumPy 日期序列中填充缺失日期?
問題:給定一個非連續日期序列的數組,通過填充缺失的日期,使其變成連續的日期序列。
輸入:# Input
dates = np.arange(np.datetime64( 2018-02-01 ), np.datetime64( 2018-02-25 ), 2)
print(dates)
#> [ 2018-02-01 2018-02-03 2018-02-05 2018-02-07 2018-02-09
#> 2018-02-11 2018-02-13 2018-02-15 2018-02-17 2018-02-19
#> 2018-02-21 2018-02-23 ]

1 import numpy as np 2 3 dates = np.arange(np.datetime64('2018-02-01'), np.datetime64('2018-02-25'), 2) 4 print("dates=", dates) 5 6 filled_in = np.array([np.arange(date, (date+d)) for date, d in zip(dates, np.diff(dates))]).reshape(-1) 7 8 output = np.hstack([filled_in, dates[-1]]) # add the last day 9 print("output=", output) 10 11 out = [] 12 for date, d in zip(dates, np.diff(dates)): 13 out.append(np.arange(date, (date+d))) 14 15 filled_in = np.array(out).reshape(-1) 16 output = np.hstack([filled_in, dates[-1]]) # add the last day 17 print("output", output) 18 19
70. 如何基於給定的 1 維數組創建 strides?
問題:給定 1 維數組 arr,使用 strides 生成一個 2 維矩陣,其中 window length 等於 4,strides 等於 2,例如 [[0,1,2,3], [2,3,4,5], [4,5,6,7]..]。
輸入:arr = np.arange(15)
arr
#> array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14])
期望輸出:#> [[ 0 1 2 3]
#> [ 2 3 4 5]
#> [ 4 5 6 7]
#> [ 6 7 8 9]
#> [ 8 9 10 11]
#> [10 11 12 13]]

1 import numpy as np 2 3 4 def gen_strides(a, stride_len=5, window_len=5): 5 n_strides = ((a.size - window_len) // stride_len) + 1 6 return np.array([a[s:(s + window_len)] for s in np.arange(0, n_strides * stride_len, stride_len)]) 7 8 print(gen_strides(np.arange(15), stride_len=2, window_len=4))