Flek ML pipeline consists of a sequence of integrated activities. When used together, they form an end-to-end machine learning workflow that starts with data preparation and ends with prediction and recommendation:

​

1. Encode & Load. Prepare the raw data and upload it into the ML engine

2. Build. Learn the probabilistic model and store the Nuggets using core ML algorithms

3. Query. Fetch the stored Nuggets from the engine

4. Mine. Iteratively scan the model store for Nuggets with given signature

5. Explore. Search for probabilistic patterns, visualize, tabulate, trace and validate

6. Discover. Auto-discover interesting associations, rules, anomalies or causal relationships

7. Predict. Run both forward and backward prediction and classification activities

8. Recommend. Perform both profile and item based recommendation tasks