The Regression Model Analysis can be used to predict the service life of facilities based off individual variable values.  Regression Models require much more data than Cohort Analysis.  In general, users should always run Cohort Analyses first, then if those results are determined to be calibrated and useful, then Regression Models can be attempted.

The Regression Model tool may be found in the Operation Center, within the Analysis tab, under the Deterioration Model drop-down menu.  Users can right-click the Regression Model option to create a new model.  Like many saved items in InfoAsset Planner, the ID must have no spaces or special characters, while the description has no character restrictions and is much easier to change later.

When you create a new Regression Model, you should select one of two windows.  In the first window, a target year for detailed tabular reports may be created as well as a specific selection of facilities can be input.

After clicking Next, the second window requires you to specify the type of model equation to apply as well as the Sensitivity Analysis.  In general, the Cox model is the easiest to run and converge.  For more information on these models/equations see the Model Equations page. The selected Sensitivity Analysis should already contain all the input data required to run the Regression Model.

The regression reliability summary chart shows the results of the regression analysis in either graphical or tabular form. These results are dynamic in the sense that you can change the variables (located above the graph/report) being plotted so that it is easier to visualize how these variables effect the model output.

Regression Model results are always shown with age on the x-axis reaching up to 100 years.  The Result Type dropdown holds the same graphing methods from the Cohort Models.  The only graph removed is the Residual Life Expectancy bar graph.  The main power within the Regression Models lies in the variables which the user can edit to compare different type of hypothetical pipes.  Category Variables are stored on the left while Stepwise Variables are on the right.  You can enter in any set of variable values and see how the resulting graphs compare to facilities with different attributes.

With large amounts of failure data, users can create very powerful statistical analysis.  With less data; however, you Regression Model graphs may look more like the example above than the example below.