Teachable-Machine-IA/evaluation.py at main · TRIBAK-Mohamed/Teachable-Machine-IA · 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
"""
evaluation.py
Métriques et plots pour classification et régression.
"""
import streamlit as st
from sklearn.metrics import (accuracy_score, precision_score, recall_score, f1_score, roc_auc_score,
                             confusion_matrix, mean_absolute_error, mean_squared_error, r2_score)
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
import pandas as pd

def evaluate_classification(model, X_test, y_test, show_plots=True):
    preds = model.predict(X_test)
    probs = None
    try:
        probs = model.predict_proba(X_test)[:,1]
    except Exception:
        try:
            probs = model.decision_function(X_test)
        except Exception:
            probs = None

    acc = accuracy_score(y_test, preds)
    prec = precision_score(y_test, preds, average='weighted', zero_division=0)
    rec = recall_score(y_test, preds, average='weighted', zero_division=0)
    f1 = f1_score(y_test, preds, average='weighted', zero_division=0)
    auc = None
    if probs is not None and len(np.unique(y_test))==2:
        try:
            auc = roc_auc_score(y_test, probs)
        except Exception:
            auc = None

    metrics = {"accuracy": acc, "precision": prec, "recall": rec, "f1": f1, "auc": auc}
    st.write(metrics)

    if show_plots:
        # confusion matrix
        cm = confusion_matrix(y_test, preds)
        fig, ax = plt.subplots()
        sns.heatmap(cm, annot=True, fmt='d', ax=ax, cmap='Blues')
        ax.set_title("Confusion Matrix")
        st.pyplot(fig)

    return metrics

def evaluate_regression(model, X_test, y_test, show_plots=True):
    preds = model.predict(X_test)
    mae = mean_absolute_error(y_test, preds)
    mse = mean_squared_error(y_test, preds)
    rmse = np.sqrt(mse)
    r2 = r2_score(y_test, preds)
    metrics = {"mae":mae, "mse":mse, "rmse":rmse, "r2":r2}
    st.write(metrics)
    if show_plots:
        fig, ax = plt.subplots()
        ax.scatter(y_test, preds, alpha=0.6)
        ax.plot([min(y_test), max(y_test)], [min(y_test), max(y_test)], color='red')
        ax.set_xlabel("Vérité terrain")
        ax.set_ylabel("Prédictions")
        ax.set_title("Vérité terrain vs Prédictions")
        st.pyplot(fig)
    return metrics

def plot_learning_curve(history):
    import matplotlib.pyplot as plt
    st.subheader("Courbes d'entraînement")
    if isinstance(history, dict):
        keys = history.keys()
        fig, ax = plt.subplots()
        for k in keys:
            ax.plot(history[k], label=k)
        ax.legend()
        st.pyplot(fig)
    else:
        st.write("Historique non disponible.")