templates/model_info.html

{% extends "base.html" %}
{% block title %}Model Info - PriceMyCar{% endblock %}
{% block content %}
<section class="model-page">
  <h1>How It Works</h1>
  <p class="section-sub">Learn how our predictive AI analyzes vehicle data to estimate the best market price.</p>

  <h2 style="margin-top:40px;font-size:1.1rem;font-weight:600">Prediction Pipeline</h2>
  <div class="pipeline">
    <div class="pipe-step"><div class="pipe-icon">📝</div><h4>1. Input Data</h4><p>You provide details like brand, model, year, and transmission.</p></div>
    <div class="pipe-step"><div class="pipe-icon">🔍</div><h4>2. Condition Analysis</h4><p>The system calculates deductions based on 10 physical factors.</p></div>
    <div class="pipe-step active"><div class="pipe-icon" style="color:#fff">⚙️</div><h4>3. AI Valuation</h4><p>Our predictive AI matches your inputs with real-world market patterns.</p></div>
    <div class="pipe-step"><div class="pipe-icon">💵</div><h4>4. Price Estimation</h4><p>The app generates a price estimate and a detailed deviation analysis.</p></div>
  </div>

  <div class="algo-metrics">
    <div class="algo-section">
      <h2>AI Prediction System</h2>
      <p>PriceMyCar uses <strong>Predictive AI technology</strong> powered by the <strong>HistGradientBoostingRegressor (HGBR)</strong> algorithm. This model was trained on thousands of real-world used car sales transactions to capture non-linear depreciation trends—for example, how the value of a luxury vehicle drops much faster than an economy family car as mileage increases.</p>
      
      <h3 style="margin-top:20px;font-size:.95rem;font-weight:600">Why HGBR was selected:</h3>
      <p>HGBR groups continuous numerical variables (like mileage or age) into 256 integer bins (histogram-based binning). This significantly reduces computation time, improves noise resistance, and acts as a strong regularizer. We applied L2 regularization (2.0) and restricted tree complexity (max leaf nodes = 25) to prevent overfitting.</p>

      <h3 style="margin-top:20px;font-size:.95rem;font-weight:600">Key Parameters Evaluated by AI:</h3>
      <ul class="feature-list">
        <li><strong>Brand &amp; Model:</strong> Popularity and demand representation via Frequency Encoding.</li>
        <li><strong>Car Age:</strong> Derived feature (2025 - year) capturing the primary driver of value loss.</li>
        <li><strong>Mileage (Odometer):</strong> Cumulative physical wear and tear.</li>
        <li><strong>Fuel &amp; Transmission:</strong> Encoded preferences (manual vs automatic and fuel types).</li>
        <li><strong>Ownership History:</strong> Encoded ordinal tiers (First Owner to Test Drive Car).</li>
        <li><strong>Interaction Terms:</strong> Engineered feature (age × km) capturing cumulative depreciation.</li>
      </ul>
    </div>
    
    <div class="metrics-section">
      <h2>AI Valuation Metrics</h2>
      <div class="metric-row">
        <div><div class="metric-name">AI Model Accuracy (R²)</div><div class="metric-desc">The proportion of used car price variance successfully explained by our machine learning features on the held-out test split.</div></div>
        <div class="metric-val">80.6%</div>
      </div>
      <div class="metric-row">
        <div><div class="metric-name">Average Price Deviation (MAE)</div><div class="metric-desc">The average absolute nominal error of predictions relative to actual market transactions (equivalent to 107,252 INR).</div></div>
        <div class="metric-val" style="color:var(--green)">Rp 20.1M</div>
      </div>
      <div class="metric-row">
        <div><div class="metric-name">Generalization Gap</div><div class="metric-desc">The difference between Train R² and Test R² (86.8% vs 80.6%). A low gap indicates the model is highly stable and does not overfit.</div></div>
        <div class="metric-val" style="color:var(--green)">6.2%</div>
      </div>
    </div>
  </div>
</section>
{% endblock %}