Assign cost pools with machine learning
Applies to: TBM Studio 12.7 and later. TBM Studio 12.7 and later can use machine learning to speed up Costing Standard configuration. Previously, mapping accounts to cost pools required manually creating mapping files with one-by-one assignment to reach the taxonomy-standard cost pools. Now, machine learning maps the majority of accounts for you in milliseconds.
Watch this demo video from Apptio Education Services: Machine Learning in 12.7. Or, browse all Apptio videos.
Apptio collects information and develops prediction models. As more data is classified, these models learn over time.
Here is an overview of how machine learning in Apptio works.
- The existing information Apptio has is classified to the desired outcome, such as cost pool, and is put into the training service (hosted in AWS).
- Apptio identifies appropriate algorithms, determines the best parameters, and establishes a prediction model. The model is published to the catalog.
- You can then access that model, apply it to your data, and review the results directly within Apptio. When the predictions have been reviewed as part of your configuration process, the now-classified data will be sent to our training service.
- The new data triggers the training service to anonymize and generate a new model. Apptio tracks and adjusts the models over time to improve the quality and quantity of our predictions.
- When a model meets our quality review, it is published in the catalog.
- You can update to the latest model as desired.
Here is a diagram of the process:
