Step 1
Model Training
The model learned from past, capturing market shifts, and unique lane behaviors.
Step 2
Prediction Testing
Once the model was trained, we tested it by making predictions.
Step 3
Validation & Measuring
Comparing predictions to actual data and measuring rates accuracy
To evaluate the model’s performance, we conducted a rigorous backtesting process. The goal was to assess how well the model could predict transportation costs based on historical data.
Our model’s performance was evaluated on two key fronts: its ability to capture long-term trends over multiple years and its accuracy in predicting rates for a specific year (2023).
LT Model uses advanced ML to predict transportation trends up to 12 months ahead, ensuring accurate, lane-specific forecasts.
Forecasted Prices for Historical Shipments
History Forecast provides historical forecasts for requested lanes, offering complete information about past prices on both new and existing lanes. This feature makes it easier to analyze LT Forecasts based on a comprehensive understanding of historical data.
GS Network Predictions
This parameter allows including GS Network forecasts and historical data in the Batch results.
Rate Mode
This parameter allows you to select the mode and your rates will display the output file.
Historical Data
This parameter allows you to choose between Real History and Predicted History for the lane history results.
Geography Level
This parameter allows you to select the level of history aggregation for a Lane. Predicted History does not support KMA or 3dZIP levels.