Machine learning life cycle

  1. Project requirement
    • Identity business problem as an opportunities
    • Analyse requirement and define target to achieve
    • Identify tools and technique to solve
    • Find subject matter expertise 
  2. Explore data
    • Gathering data from various source
    • Data cleaning and labelling
    • Feature selection
    • Data versioning 
    • Store metadata
  3. Model architecture
    • Decide target variable
    • Build appropriate model by substantial research
    • Model training & validating 
    • Versioning experimental model
    • Test & evaluate effective model
  4. Model Deployment & Documentation
    • Deploy model as a service endpoint 
    • Document model process 
  5. Model Monitoring
    • Monitor hardware and software performance
    • Customer satisfaction
    • Get an alert if anomalies detection 

Leave a Reply

Your email address will not be published. Required fields are marked *