-
Notifications
You must be signed in to change notification settings - Fork 1
/
experiments.py
109 lines (90 loc) · 3.46 KB
/
experiments.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
from tree.base import *
from tree.utils import *
from time import perf_counter as time
np.random.seed(42)
def create_binary_dataset(N, M):
# Generate a random binary matrix with N samples and M features
data = np.random.randint(0, 2, size=(N, M))
# Create column names for the features
feature_names = [f'Feature_{i+1}' for i in range(M)]
# Create a DataFrame
df = pd.DataFrame(data, columns=feature_names)
# Change datatype to float
for i in range(M):
df[f'Feature_{i+1}'] = df[f'Feature_{i+1}'].astype(float)
df.rename(columns={f'Feature_{M}': 'Output'}, inplace=True)
return df
N = 20
M = 20
iterations = 1
timeFitting = np.array([[0]*(M + 1) for i in range(N + 1)], dtype=float)
timePredicting = np.array([[0]*(M + 1) for i in range(N + 1)], dtype=float)
for n in range(1, N + 1):
for m in range(1, M + 1):
pred = 0
fit = 0
for iter in range(iterations):
DF = create_binary_dataset(2*n, m)
x = DF.drop(columns=['Output'])
y = DF['Output']
xTrain = x.head(n)
yTrain = y.head(n)
xTest = x.tail(n)
yTest = y.tail(n)
tree = DecisionTree(criterion='information_gain', maxDepth=3)
start = time()
tree.fit(xTrain, yTrain)
end = time()
fit += end - start
start = time()
ypred = tree.predict(xTest)
end = time()
pred += end - start
timeFitting[n][m] = fit / iterations
timePredicting[n][m] = pred / iterations
x_dim, y_dim = N + 1, M + 1
x = np.arange(x_dim)
y = np.arange(y_dim)
x, y = np.meshgrid(x, y)
# Create the plot with four subplots
fig = plt.figure(figsize=(12,8))
# Plot for timeFitting
ax1 = fig.add_subplot(221, projection='3d')
ax1.plot_surface(x, y, timeFitting, cmap='viridis')
ax1.set_title('Time Fitting', fontsize=14)
ax1.set_xlabel('Number of Samples (N)', fontsize=12)
ax1.set_ylabel('Number of Features (M)', fontsize=12)
ax1.set_zlabel('Time (s)', fontsize=12)
ax1.tick_params(axis='both', labelsize=10)
# Plot for timePredicting
ax2 = fig.add_subplot(222, projection='3d')
ax2.plot_surface(x, y, timePredicting, cmap='viridis')
ax2.set_title('Time Predicting', fontsize=14)
ax2.set_xlabel('Number of Samples (N)', fontsize=12)
ax2.set_ylabel('Number of Features (M)', fontsize=12)
ax2.set_zlabel('Time (s)', fontsize=12)
ax2.tick_params(axis='both', labelsize=10)
# Plot for O(n * d * log n)
O_nd_log_n = np.array([[n * m * np.log2(n + 1) for m in range(y_dim)] for n in range(x_dim)])
ax3 = fig.add_subplot(223, projection='3d')
ax3.plot_surface(x, y, O_nd_log_n, cmap='plasma')
ax3.set_title('O(n * d * log n)', fontsize=14)
ax3.set_xlabel('Number of Samples (N)', fontsize=12)
ax3.set_ylabel('Number of Features (M)', fontsize=12)
ax3.set_zlabel('Time Complexity', fontsize=12)
ax3.tick_params(axis='both', labelsize=10)
# Plot for log2(n)
log2_n = np.array([[np.log2(n + 1) for _ in range(y_dim)] for n in range(x_dim)])
ax4 = fig.add_subplot(224, projection='3d')
ax4.plot_surface(x, y, log2_n, cmap='coolwarm')
ax4.set_title('log2(n)', fontsize=14)
ax4.set_xlabel('Number of Samples (N)', fontsize=12)
ax4.set_ylabel('Number of Features (M)', fontsize=12)
ax4.set_zlabel('log2(n)', fontsize=12)
ax4.tick_params(axis='both', labelsize=10)
plt.tight_layout()
plt.show()