SigMA/plots.py at main · hammouc/SigMA · GitHub

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
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import time
import pickle
from matplotlib.lines import lineStyles
from tabulate import tabulate
import torch
import torch.nn.functional as F
import torch.optim as optim
import utils
import NNmodels

__name__ = "__example2__" # "__example2__", "__example4__", "__example5__"

if __name__ == '__example2__':
    '''输入误差数据'''
    fBm_100 = np.array([[9.52e-2, 1.72e-2, 1.34e-2],
                        [1.18e-1, 1.15e-2, 1.06e-2],
                        [2.85e-1, 1.32e-1, 1.09e-1]])
    fOU_100 = np.array([[1.34e-1, 3.34e-2, 3.41e-2],
                        [1.34e-1, 2.96e-2, 2.44e-2],
                        [1.36e-1, 1.35e-1, 1.35e-1]])
    rHeston_100 = np.array([[5.89e-2, 5.81e-2, 5.78e-2],
                            [5.89e-2, 5.44e-2, 5.48e-2],
                            [1.15e-1, 5.89e-2, 5.89e-2]])
    fBm_500 = np.array([[9.32e-2, 8.41e-3, 4.53e-2],
                        [1.26e-1, 8.89e-3, 8.24e-2],
                        [2.74e-1, 1.33e-1, 1.08e-1]])
    fOU_500 = np.array([[1.37e-1, 2.18e-2, 2.93e-2],
                        [1.37e-1, 1.59e-2, 1.26e-2],
                        [1.39e-1, 1.39e-1, 1.39e-1]])
    rHeston_500 = np.array([[5.92e-2, 5.03e-2, 3.44e-2],
                            [5.92e-2, 1.02e-2, 1.01e-2],
                            [7.59e-2, 5.92e-2, 5.92e-2]])
    nn_models = ['DeepSigNet', 'SigMA', 'SigSA']
    truncation_order = ['1', '3', '5']
    v_model_dict = {'fBm_100': fBm_100, 'fOU_100': fOU_100, 'rHeston_100': rHeston_100,
                    'fBm_500': fBm_500, 'fOU_500': fOU_500, 'rHeston_500': rHeston_500}
    '''选择随机过程'''
    v_model = 'fBm_100'  # 'fBm_100', 'fOU_100', 'rHeston_100', 'fBm_500', 'fOU_500', 'rHeston_500'
    '''绘制误差折线图'''
    plt.figure(figsize=(10, 10))
    colors = ['blue', 'orange', 'red']
    markers = ['o', 's', '^']
    lines = ['--', '-', ':']
    for i in range(len(nn_models)):
        plt.plot(truncation_order, v_model_dict[v_model][i], label=nn_models[i], color=colors[i],
                 marker=markers[i], linewidth=3.5, linestyle=lines[i], markersize=10)
    plt.xticks(truncation_order, fontsize=20)
    plt.yticks(fontsize=20)
    plt.xlabel('Truncation order', fontsize=20, fontweight='bold')
    plt.ylabel('Test RMSE', fontsize=20, fontweight='bold')
    legend = plt.legend(mode='expand', bbox_to_anchor=(0, 1, 1, 0), ncol=3, prop={'size': 20})
    '标注SigMA'
    for text in legend.get_texts():
        if text.get_text() == 'SigMA':
            text.set_fontweight('bold')
        else:
            text.set_fontweight('normal')
    plt.savefig(f'data/results/numerical_example2/plots/{v_model}_example2.eps')
    plt.show()

if __name__ == "__example4__":
    '''输入误差数据'''
    fBm = np.array([[1.90e-2, 4.68e-2, 7.75e-2, 8.94e-2],
                    [1.05e-1, 1.07e-1, 1.12e-1, 1.10e-1],
                    [7.70e-2, 5.35e-2, 3.83e-2, 3.89e-2],
                    [1.60e-2, 9.02e-3, 6.84e-3, 3.13e-2],
                    [1.66e-2, 8.51e-3, 8.16e-3, 1.09e-2]])
    fOU = np.array([[9.22e-2, 8.93e-2, 9.53e-2, 9.24e-2],
                    [4.30e-2, 4.05e-2, 4.01e-2, 3.85e-2],
                    [9.22e-2, 4.69e-2, 3.19e-2, 3.46e-2],
                    [2.82e-2, 1.61e-2, 2.03e-2, 2.28e-2],
                    [3.51e-2, 3.01e-2, 2.96e-2, 2.51e-2]])
    rHeston = np.array([[5.72e-2, 5.24e-2, 4.37e-2, 4.25e-2],
                        [8.05e-2, 7.65e-2, 6.60e-2, 5.79e-2],
                        [7.25e-2, 6.29e-2, 6.16e-2, 6.11e-2],
                        [4.78e-2, 1.01e-2, 7.23e-3, 7.06e-3],
                        [5.77e-2, 5.13e-2, 1.97e-2, 9.05e-3]])
    nn_models = ['Transformer', 'CNN', 'LSTM', 'SigMA', 'DeepSigNet']
    input_lengths = ['100', '500', '1000', '1500']
    v_model_dict = {'fBm': fBm, 'fOU': fOU, 'rHeston': rHeston}
    '''选择随机过程'''
    v_model = 'rHeston' # 'fBm', 'fOU', 'rHeston'
    '''绘制误差折线图'''
    plt.figure(figsize=(10, 10))
    colors = ['blue', 'red', 'green', 'orange', 'purple']
    markers = ['o', 's', '^', 'D', 'v']
    lines = ['--', ':', (0, (3, 2, 1, 2, 1, 2)), '-', '-.']
    for i in range(len(nn_models)):
        plt.plot(input_lengths, v_model_dict[v_model][i], label=nn_models[i], color=colors[i],
                 marker=markers[i], linewidth=3.5, linestyle=lines[i], markersize=10)
    plt.xticks(input_lengths, fontsize=20)
    plt.yticks(fontsize=20)
    plt.xlabel('Input Length', fontsize=20, fontweight='bold')
    plt.ylabel('Test RMSE', fontsize=20, fontweight='bold')
    legend = plt.legend(mode='expand', bbox_to_anchor=(0, 1, 1, 0), ncol=3, prop={'size': 20})
    '标注SigMA'
    for text in legend.get_texts():
        if text.get_text() == 'SigMA':
            text.set_fontweight('bold')
        else:
            text.set_fontweight('normal')
    plt.savefig(f'data/results/numerical_example4/plots/{v_model}_example4.eps')
    plt.show()

if __name__ == '__example5__':
    '''选择随机过程'''
    v_model = 'rHeston'  # 'fOU', 'rHeston'
    '''导入误差数据'''
    errors_average_rse = {'Transformer': [], 'SigMA': [], 'Deepsignet': [], 'CNN': [], 'LSTM': []}
    errors_average_rmse = {'Transformer': [], 'SigMA': [], 'Deepsignet': [], 'CNN': [], 'LSTM': []}
    for round in range(10):
        data = np.load(f'data/results/numerical_example5/errors/{v_model}/errors_se_round{round}.npz')
        for nn_model_name in errors_average_rse:
            error_average_rse = np.mean(np.sqrt(data[nn_model_name].squeeze(0)), axis=1)
            error_average_rmse = np.mean(np.sqrt(np.mean(data[nn_model_name].squeeze(0), axis=0)))
            errors_average_rse[nn_model_name].append(error_average_rse)
            errors_average_rmse[nn_model_name].append(error_average_rmse)

    '打印10轮平均Average RMSE'
    print(f'{v_model}_Average_RMSE_average:--------------')
    for nn_model_name in errors_average_rmse:
        print(f"{nn_model_name:15}: {np.mean(errors_average_rmse[nn_model_name]):7.3f}")

    '打印并绘制10轮Average RSE的分布情况'
    print(f'{v_model}_Average_RSE:--------------')
    colors = {'Transformer': 'blue', 'SigMA': 'orange', 'Deepsignet': 'purple', 'CNN': 'red', 'LSTM': 'green'}
    lines = {'Transformer': '--', 'SigMA': '-', 'Deepsignet': '-.', 'CNN': ':', 'LSTM': (0, (3, 2, 1, 2, 1, 2))}
    plt.figure(figsize=(20, 8))
    for nn_model_name in errors_average_rse:
        rse = np.concatenate(errors_average_rse[nn_model_name])
        q1 = np.percentile(rse, 25)
        q3 = np.percentile(rse, 75)
        maxi = np.max(rse)
        print(f'{nn_model_name:15} max:{maxi:7.3f} q3:{q3:7.3f} q1:{q1:7.3f}')
        sns.kdeplot(rse, label=nn_model_name, bw_adjust=5, linewidth=3.5, color=colors[nn_model_name], linestyle=lines[nn_model_name])
    plt.xlim(0, 2)
    plt.xticks(fontsize=20)
    plt.yticks(fontsize=20)
    plt.xlabel('Average RSE', fontsize=20, fontweight='bold')
    plt.ylabel('Probability density', fontsize=20, fontweight='bold')
    legend = plt.legend(loc='best', prop={'size': 20})
    '标注SigMA'
    for text in legend.get_texts():
        if text.get_text() == 'SigMA':
            text.set_fontweight('bold')
        else:
            text.set_fontweight('normal')
    plt.tight_layout()
    plt.savefig(f'data/results/numerical_example5/plots/{v_model}_example5.eps')
    plt.show()