0
# Performance Evaluation
1

2
Comprehensive model performance analysis tools with interactive visualizations for classification and regression tasks.
3

4
## Capabilities
5

6
### ROC Analysis
7

8
Receiver Operating Characteristic curves for binary and multi-class classification performance evaluation.
9

10
```python { .api }
11
class ROC:
12
    def __init__(self, **kwargs):
13
        """ROC curve analyzer."""
14
    
15
    def explain_perf(self, y_true, y_prob, name=None):
16
        """
17
        Generate ROC curve analysis.
18
        
19
        Parameters:
20
            y_true (array-like): True binary labels
21
            y_prob (array-like): Predicted probabilities
22
            name (str, optional): Name for explanation
23
            
24
        Returns:
25
            Explanation object with ROC curves and AUC scores
26
        """
27
```
28

29
### Precision-Recall Analysis
30

31
Precision-Recall curves particularly useful for imbalanced datasets.
32

33
```python { .api }
34
class PR:
35
    def __init__(self, **kwargs):
36
        """Precision-Recall curve analyzer."""
37
    
38
    def explain_perf(self, y_true, y_prob, name=None):
39
        """
40
        Generate Precision-Recall curve analysis.
41
        
42
        Parameters:
43
            y_true (array-like): True binary labels
44
            y_prob (array-like): Predicted probabilities
45
            name (str, optional): Name for explanation
46
            
47
        Returns:
48
            Explanation object with PR curves and average precision scores
49
        """
50
```
51

52
### Regression Performance
53

54
Comprehensive regression metrics including residual analysis and error distributions.
55

56
```python { .api }
57
class RegressionPerf:
58
    def __init__(self, **kwargs):
59
        """Regression performance analyzer."""
60
    
61
    def explain_perf(self, y_true, y_pred, name=None):
62
        """
63
        Generate regression performance analysis.
64
        
65
        Parameters:
66
            y_true (array-like): True target values
67
            y_pred (array-like): Predicted values
68
            name (str, optional): Name for explanation
69
            
70
        Returns:
71
            Explanation object with regression metrics and residual plots
72
        """
73
```
74

75
## Usage Examples
76

77
```python
78
from interpret.perf import ROC, PR, RegressionPerf
79
from interpret import show
80
from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor
81
from sklearn.datasets import make_classification, make_regression
82
from sklearn.model_selection import train_test_split
83

84
# Classification performance
85
X, y = make_classification(n_samples=1000, n_classes=2, random_state=42)
86
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
87

88
clf = RandomForestClassifier(random_state=42)
89
clf.fit(X_train, y_train)
90
y_prob = clf.predict_proba(X_test)[:, 1]
91

92
# ROC analysis
93
roc = ROC()
94
roc_exp = roc.explain_perf(y_test, y_prob, name="ROC Analysis")
95
show(roc_exp)
96

97
# PR analysis
98
pr = PR()
99
pr_exp = pr.explain_perf(y_test, y_prob, name="PR Analysis")
100
show(pr_exp)
101

102
# Regression performance
103
X_reg, y_reg = make_regression(n_samples=1000, noise=0.1, random_state=42)
104
X_train_reg, X_test_reg, y_train_reg, y_test_reg = train_test_split(X_reg, y_reg, test_size=0.2, random_state=42)
105

106
reg = RandomForestRegressor(random_state=42)
107
reg.fit(X_train_reg, y_train_reg)
108
y_pred_reg = reg.predict(X_test_reg)
109

110
reg_perf = RegressionPerf()
111
reg_exp = reg_perf.explain_perf(y_test_reg, y_pred_reg, name="Regression Performance")
112
show(reg_exp)
113
```