The “Options” tab allows you to set options for the regression calculations made in the Prediction Modeling Module.
The critical condition value is a user-defined trigger value used in M&R calculations to determine what type of maintenance a pavement should receive. The “Suggested Critical Condition” is the FAA PAVEAIR suggestion for the critical condition and is determined during the regression analysis as the steepest part of the regression curve between a condition of 50 and 80. Note that at this time, M&R calculations use a default critical condition of 60.
The “Intercept” specifies the y-intercept, that is, where the family curve crosses they y-axis. The default value is 100 to reflect that most pavements have a PCI of 100 at an age of 0. Other indices, such as FOD Index, typically have a value of 0 at an age of zero. Select an appropriate intercept based on the type of family curve you are developing.
Using the radio buttons next to "Slope," you may influence the graph's curve direction by selecting either "Downward" or "Unconstrained". The “Unconstrained” option places no restrictions on the curve, which sometimes results in a model with positive slope (condition increasing with time).
FAA PAVEAIR has the ability to constrain the slope downward using an iterative calculation. If a regression curve has a positive slope anywhere, it identifies the point near the positive slope with the most error, marks it bad, and re-runs the regression. The process is repeated until either there are no points left, or the model has a downward slope. If the iterative process deletes all the datapoints, an error is returned and the user should examine the data for issues such as bad inspection data and mixed pavement types or uses.
The “Automatically Calculate Number of Regression Coefficients” checkbox allows the user to specify the order of the regression equation or allow FAA PAVEAIR to determine the order that provides the best fit. “Best Fit” is defined as the equation has having the highest r² value. Available options are 2 (quadratic regression equation or x²), 3 (cubic regression equation or x³), and 4 (quartic regression equation or x4).