Please enable JavaScript.

Coggle requires JavaScript to display documents.

Machine Learning (Process (ML2.0 (Steps (Prediction engineering: Assemble…

- - - - Steps
        
        Prepare data
        
        "cutoff_time"
        
        Extract relevant data subset
        
        metadata.json
        
        Engineer features
        
        Deep Feature Synthesis
        
        Learn a model
        
        Auto-Tuned Models (ATM)
        
        Evaluate a model and report
        
        model_provenance.json
        
        Formulate problem, assemble training examples
        
        Label Segment Featurize
  - - - Prediction engineering: Assemble training examples
        
        custom labeling function
      - Feature engineering: Generate features
        
        Deep Feature Synthesis
      - Modelling and operationalization: Generate a model M
        
        custom cost function
      - Integration testing in production
        
        API driven
      - Validate in production
        
        evaluate model in production environment
      - Deploy
        
        same steps as in 5
      - Data organization: Form an Entityset
        
        Featuretools, metadata.json
    - - be end-to-end
      - enable confidence
      - provide deployable code
      - provide provisions for update
      - subsume discovery