ML Pipeline Design

Flek's ML pipeline consists of a sequence of multiple activities. When used together, they form an integrated pipeline starting at data preparation and ending with making predictions. Here are the 6 steps:

​

1. Encode & Load the raw data into the ML engine

2. Build the model and run the ML engine on a local machine or a remote server

3. Store the generated Nuggets into the Universal Model

4. Query the model to get statistics, association rules and association maps

5. Rule Mining and analyze results to gain insight

6. Predict, get scores and test for accuracy

​