# AutoGrab Confidence Score

AutoGrab’s confidence score is derived from a statistical process that reflects how confident the valuation model is in its predicted value. The score ranges between 0 and 1, where higher values indicate stronger confidence in the accuracy of the prediction, and lower values suggest greater uncertainty.

 

### **Interpreting the Confidence Score**

* **High Confidence (e.g., >0.8):** The valuation model is highly certain about the estimated retail value, supported by a large dataset and strong model accuracy.
* **Medium Confidence (e.g., 0.5–0.8):** The prediction is reasonable but may require additional validation due to a smaller sample size or price volatility.
* **Low Confidence (e.g., <0.5):** The prediction is uncertain, often due to limited data availability or high market volatility.

&#x20;

### **Factors Affecting Confidence Scores**

* **Data Quality and Availability:**
  * High-quality, complete data for a vehicle variant increases the confidence score.
  * Listings with poor data quality (e.g., missing mileage, unknown price) are excluded from AutoGrab’s valuation model.
* **Model Fit:**
  * Vehicles closely aligned with the training data generally have higher confidence scores.
  * Outliers and rare vehicle types may result in lower confidence.
* **Recency of Data:**
  * AutoGrab incorporates recency weighting in its valuation model.
  * Vehicles with more recent and high number of listings achieve higher confidence scores.

&#x20;

### **Applications of Confidence Scores**

* **Informed Decision-Making:** AutoGrab’s confidence score helps customers determine whether additional inspection or pricing validation is needed before acquiring a vehicle.
* **Risk Mitigation:** A low confidence score highlights higher risks, enabling AutoGrab’s customers to approach vehicle transactions with caution and avoid potential losses.
* **Enhanced Transparency:** Including confidence scores alongside valuations enhances trust and confidence with AutoGrab’s clients, allowing them to assess the reliability of predictions.

&#x20;

This framework ensures AutoGrab provides reliable valuations while highlighting areas that may require further analysis.