The Significance of Data Accuracy

Data accuracy is a critical measure in evaluating the performance of our Predicted Estimated Time of Arrival (PETA) model. Accurate predictions enable clients to coordinate their supply chain more effectively, helping them identify containers at risk of delays and take proactive measures to mitigate disruptions. This capability is essential for optimizing logistics operations and maintaining customer satisfaction.

<aside> 💡

Please contact [email protected] or your dedicated Client Success Manager to obtain customized accuracy reports.

</aside>

Methodology for Measuring Accuracy

Accuracy in our system is determined by comparing the final PETA (Predictive Estimated Time of Arrival) to the ATA (Actual Time of Arrival).

The formula for defining accuracy is:

X% accurate on Y days ahead, where:

• X% indicates the percentage of accurate predictions.
• Y days ahead specifies the time frame for when the prediction was made relative to the ATA.

Critical Determinants of Accuracy

Several factors influence how accuracy is defined:

1. Accuracy Timeframe

Accuracy is defined by the user, though by default, we use a 1-day timeframe in our accuracy dashboard. This means a prediction is considered accurate if it is within one day of the ATA.

• Custom Timeframes: Some customers may choose a narrower definition of accuracy, such as 12 hours. In such cases, any prediction that deviates by more than 12 hours is considered inaccurate.

2. Forecast Horizon (Y)

The value of Y days ahead represents how far in the future the prediction is made. Customers can configure this based on their specific needs:

• A larger Y (e.g., 5 days ahead) indicates the focus is on predictions further into the future.
• Customers prioritizing short-term predictions may set Y to a smaller value (e.g., 1 day ahead).

Practical Example