Pretrained ML Models to predict Change Order Cycle Time

Manjunath Rg-Oracle

Description

FAW or OTBI should create ML models to predict Change Order Cycle time for different verticals. 

An example -

• A model created on Manufacturing Industries' data (public data if available) can be trained and provided to customer as a value-add
• The model can use PCA (principal component analysis) to figure out which input features are best predictors and which can be dropped. 
• This can leverage OAC's ML tools or ADW's ML algorithms.
• Customers can then input their data and use "transfer learning" to fine tune this shipped model as per their need. 

 

Use Case and Business Need

• OAC supports many ML models and provides a flow for creating ML models. 
• However, this has to be done by the implementor as of today
• Providing at least few pre-trained models via OTBI or FAW which the customer can fine tune for their data. 
• This effectively gives a jump start for customers to move into ML/AI space. 

Original Idea Number: f280843395