Spaces:
Running
Running
File size: 5,163 Bytes
71188bb |
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 |
import streamlit as st
st.set_page_config(page_title="Model Evaluation Metrics", page_icon="📊", layout="wide")
st.sidebar.title("📊 Model Evaluation Metrics")
st.sidebar.markdown("Select a metric category from below.")
st.markdown("<h1 style='text-align: center;'>📏 Model Evaluation Metrics</h1>", unsafe_allow_html=True)
metric_type = st.radio(
"Select the type of model evaluation:",
["🎯 Classification Metrics", "📈 Regression Metrics"]
)
if metric_type == "🎯 Classification Metrics":
st.markdown("## 🎯 Classification Metrics")
st.write("Used when the target variable is **categorical**.")
st.markdown("### 1. Accuracy")
st.write("""
- **Definition**: Correct predictions out of total predictions
- **Formula**:
Accuracy = (TP + TN) / (TP + FP + FN + TN)
- ⚠️ Avoid using when classes are imbalanced.
""")
st.markdown("### 2. Confusion Matrix")
st.write("""
A matrix that compares actual and predicted labels.
Useful for understanding **true positives**, **false positives**, **true negatives**, and **false negatives**.
| | Predicted Positive | Predicted Negative |
|---------------|--------------------|--------------------|
| Actual Positive | True Positive (TP) | False Negative (FN) |
| Actual Negative | False Positive (FP) | True Negative (TN) |
- Use for binary and multiclass classification.
""")
st.markdown("### 3. Precision")
st.latex(r"Precision = \frac{TP}{TP + FP}")
st.write("Of all predicted positives, how many were correct.")
st.markdown("### 4. Recall (Sensitivity)")
st.latex(r"Recall = \frac{TP}{TP + FN}")
st.write("Of all actual positives, how many were correctly identified.")
st.markdown("### 5. F1 Score")
st.latex(r"F1 = 2 \cdot \frac{Precision \cdot Recall}{Precision + Recall}")
st.write("Harmonic mean of precision and recall. Good for imbalanced classes.")
st.markdown("### 6. Specificity (True Negative Rate)")
st.latex(r"Specificity = \frac{TN}{TN + FP}")
st.write("Measures how well the model identifies negatives.")
st.markdown("### 7. ROC Curve and AUC")
st.write("""
- **ROC Curve**: Plot of True Positive Rate (Recall) vs False Positive Rate
- **AUC** (Area Under the Curve): Measures model's ability to distinguish classes.
- AUC = 1: Perfect
- AUC = 0.5: Random
""")
st.markdown("### 8. Log Loss (Logarithmic Loss)")
st.latex(r"LogLoss = -\frac{1}{n} \sum \left[ y \log(\hat{y}) + (1 - y) \log(1 - \hat{y}) \right]")
st.write("""
- Evaluates predicted probabilities instead of just labels
- Lower log loss indicates better performance
- Especially useful for probabilistic models
""")
elif metric_type == "📈 Regression Metrics":
st.markdown("## 📈 Regression Metrics")
st.write("Used when the target variable is **continuous**.")
st.markdown("### 1. Mean Absolute Error (MAE)")
st.latex(r"MAE = \frac{1}{n} \sum_{i=1}^{n} |y_i - \hat{y}_i|")
st.write("Measures average absolute difference between actual and predicted values. More robust to outliers.")
st.markdown("### 2. Mean Squared Error (MSE)")
st.latex(r"MSE = \frac{1}{n} \sum_{i=1}^{n} (y_i - \hat{y}_i)^2")
st.write("Penalizes large errors more than MAE. Sensitive to outliers.")
st.markdown("### 3. Root Mean Squared Error (RMSE)")
st.latex(r"RMSE = \sqrt{MSE}")
st.write("Square root of MSE. Easy to interpret since it has same units as output.")
st.markdown("### 4. R² Score (Coefficient of Determination)")
st.latex(r"R^2 = 1 - \frac{SS_{res}}{SS_{tot}}")
st.write("""
Indicates how well model explains variation in data:
- **1.0** → perfect
- **0.0** → same as predicting mean
- **< 0** → worse than mean
""")
st.markdown("### 5. Adjusted R² Score")
st.latex(r"\text{Adjusted } R^2 = 1 - \left( \frac{(1 - R^2)(n - 1)}{n - k - 1} \right)")
st.write("""
- Adjusts R² for number of predictors (k)
- Prevents overestimating performance from adding irrelevant features
""")
st.markdown("### 6. Mean Absolute Percentage Error (MAPE)")
st.latex(r"MAPE = \frac{100\%}{n} \sum_{i=1}^{n} \left| \frac{y_i - \hat{y}_i}{y_i} \right|")
st.write("Expresses error as a percentage. Avoid if actual values can be 0.")
st.markdown("### 7. Median Absolute Error")
st.write("Robust metric not influenced by outliers. Takes the median of all absolute differences.")
st.markdown("---")
st.markdown("### ✅ Choosing the Right Metric")
st.write("""
- **Classification**:
- Use **F1-score** for imbalanced data.
- Use **AUC-ROC** for probabilistic classifiers.
- Use **Log Loss** if working with predicted probabilities.
- **Regression**:
- Use **RMSE** when large errors are more serious.
- Use **MAE** when all errors matter equally.
- Use **R²** to evaluate explained variance.
- Always compare with a **baseline model**.
""")
st.success("Choosing the right metric helps you evaluate and improve your model with confidence!")
|