Monitoring and debugging models

Debugging a model during development is a crucial development step. With Advanced Analytics Integration in on-premises environments, we have several options to debug our model and figure out how it is performing.

The first and most logical place to start debugging in an interactive scenario is to look at the chart output. If there is something wrong with the code, you will get an error message here. In the following example, we can see that a library called forecast is missing from the environment:

Figure 7.8: Error message in chart

Sometimes, you may need more comprehensive information or debug prints from the actual R code. Since R is running as a service, there is no easy way to get debug prints during execution. You can, however, use file writing. Returning to our previous Rserve example, adding the following code will produce a file called debug.txt in our Rserve home folder (the added code is shown in bold):

...