FederatedSpamClassifier/plot_utils.py at main · FedeOmb/FederatedSpamClassifier · 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
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

def visualize_data_distribution(stats):
    if not stats: return
    clients = [s["client"] for s in stats]
    cleans = [s["clean"] for s in stats]
    spams = [s["spam"] for s in stats]
    totals = [s["total"] for s in stats]

    x = np.arange(len(clients))
    width = 0.35

    fig, ax = plt.subplots(figsize=(10, 5))
    rects1 = ax.bar(x - width/2, cleans, width, label='Clean', color='skyblue')
    rects2 = ax.bar(x + width/2, spams, width, label='Spam', color='orange')

    for i, v in enumerate(totals):
        ax.text(i, v + 5, str(v), ha='center', fontweight='bold')

    ax.set_ylabel('Quantità Dati')
    ax.set_title('Distribuzione per classe tra i client')
    ax.set_xticks(x)
    ax.set_xticklabels(clients)
    ax.legend()

    plt.tight_layout()
    plt.show()

def plot_metrics_single_run(history):
    plt.figure(figsize=(10, 5))
    sns.lineplot(data=history, x='round', y='f1', linewidth=2, alpha=0.8)
    plt.title('F1')
    plt.xlabel('Round')
    plt.ylabel('F1')
    plt.grid(True, alpha=0.3)
    plt.show(block=False)

    plt.figure(figsize=(10, 5))
    sns.lineplot(data=history, x='round', y='recall', linewidth=2, alpha=0.8)
    plt.title('Recall')
    plt.xlabel('Round')
    plt.ylabel('Recall')
    plt.grid(True, alpha=0.3)
    plt.show(block=False)

    plt.figure(figsize=(10, 5))
    sns.lineplot(data=history, x='round', y='precision', linewidth=2, alpha=0.8)
    plt.title('Precision')
    plt.xlabel('Round')
    plt.ylabel('Precision')
    plt.grid(True, alpha=0.3)
    plt.show(block=False)

    plt.figure(figsize=(10, 5))
    sns.lineplot(data=history, x='round', y='loss', linewidth=2, alpha=0.8)
    plt.title('Loss')
    plt.xlabel('Round')
    plt.ylabel('Loss')
    plt.grid(True, alpha=0.3)
    plt.show(block=False)

    plt.show()

def plot_metrics_grid_search(results_df):
    palette = sns.color_palette("tab20", n_colors=results_df['params_str'].nunique())

    plt.figure(figsize=(10, 5))
    sns.lineplot(data=results_df, x='round', y='f1', hue='params_str', linewidth=2, alpha=0.8, palette=palette)
    plt.title('F1')
    plt.xlabel('Round')
    plt.ylabel('F1')
    plt.grid(True, alpha=0.3)
    plt.legend(title="parametri", bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0.)
    plt.tight_layout(rect=[0, 0, 0.82, 1])
    plt.show(block=False)

    plt.figure(figsize=(10, 5))
    sns.lineplot(data=results_df, x='round', y='recall', hue='params_str', linewidth=2, alpha=0.8, palette=palette)
    plt.title('Recall')
    plt.xlabel('Round')
    plt.ylabel('Recall')
    plt.grid(True, alpha=0.3)
    plt.legend(title="parametri", bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0.)
    plt.tight_layout(rect=[0, 0, 0.82, 1])

    plt.show(block=False)

    plt.figure(figsize=(108, 5))
    sns.lineplot(data=results_df, x='round', y='precision', hue='params_str', linewidth=2, alpha=0.8, palette=palette)
    plt.title('Precision')
    plt.xlabel('Round')
    plt.ylabel('Precision')
    plt.grid(True, alpha=0.3)
    plt.legend(title="parametri", bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0.)
    plt.tight_layout(rect=[0, 0, 0.82, 1])
    plt.show(block=False)

    plt.figure(figsize=(10, 5))
    sns.lineplot(data=results_df, x='round', y='loss', hue='params_str', linewidth=2, alpha=0.8, palette=palette)
    plt.title('Validation Loss')
    plt.xlabel('Round')
    plt.ylabel('Loss')
    plt.grid(True, alpha=0.3)
    plt.legend(title="parametri", bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0.)
    plt.tight_layout(rect=[0, 0, 0.82, 1])
    plt.show(block=False)

    plt.show()

def plot_alpha_comparison(df_results):
    sns.set_style("whitegrid")

    alphas = sorted(df_results['alpha'].unique(), reverse=True)

    for a in alphas:
        plt.figure(figsize=(10, 6))
        data = df_results[df_results['alpha'] == a]

        sns.lineplot(data=data, x='round', y='f1', hue='algorithm', linewidth=2.5)

        plt.title(f"Robustness Analysis alpha = {a}", fontsize=14, fontweight='bold')
        plt.xlabel("Rounds")
        plt.ylabel("Test F1-Score")
        plt.grid(True, alpha=0.3)
        plt.legend(loc='lower right')
        plt.tight_layout()

        plt.figure(figsize=(10, 6))
        sns.lineplot(data=data, x='round', y='loss', hue='algorithm', linewidth=2.5)

        plt.title(f"Robustness Analysis alpha = {a}", fontsize=14, fontweight='bold')
        plt.xlabel("Rounds")
        plt.ylabel("Test Loss")
        plt.grid(True, alpha=0.3)
        plt.legend(loc='upper right')
        plt.tight_layout()

        plt.show()