User Guide

• The getML suite
  • Engine
    • Starting the engine
    • Shutting down the engine
    • Logging
  • Monitor
  • Python API
    • Connecting to the getML engine
    • Session management
    • Lifecycles and synchronization between engine and API
      • Lifecycle of a DataFrame
      • Lifecycle of a Pipeline
• Managing projects
  • The relationship between projects and engine processes
  • Managing data using projects
  • Using the project module to manage your project
• Importing data
  • Unified import interface
    • CSV interface
      • Import from CSV
      • Export to CSV
    • Pandas interface
      • Import from Pandas
      • Export to Pandas
    • JSON interface
      • Import from JSON
      • Export to JSON
    • SQLite3 interface
      • Import from SQLite3
      • Export to SQLite3
    • MySQL interface
      • Import from MySQL
      • Export to MySQL
    • MariaDB interface
      • Import from MariaDB
      • Export to MariaDB
    • PostgreSQL interface
      • Import from PostgreSQL
      • Export to PostgreSQL
    • Greenplum interface
      • Import from Greenplum
      • Export to Greenplum
    • ODBC interface
      • An example: Microsoft SQL Server
      • Important: Always pass escape_chars
      • Import data using ODBC
      • Export data using ODBC
  • Data Frames
  • Handling of NULL values
• Annotating data
  • In short
  • Roles
    • A note on reproducibility and efficiency
    • Join key
    • Time stamp
    • Target
    • Numerical
    • Categorical
    • Text
    • Unused_float
    • Unused_string
  • Units
• Data model
  • Tables
    • The population table
    • Peripheral tables
  • Placeholders
  • Joins
  • Data schemata
    • The star schema
    • The snowflake schema
  • Time series
    • Self-joining a single table
    • Features based on time stamps
• Preprocessing
  • Mappings
  • Handling of free form text
• Feature engineering
  • Definition
  • Feature learning vs. propositionalization
  • Design principles
  • Algorithms
    • FastProp
    • Multirel
    • Relboost
    • RelMT
• Predicting
  • Using getML
  • Using external software
• Hyperparameter optimization
  • Tuning routines
    • Tuning recipes and hyperparameter subspaces
  • Random search
  • Latin hypercube search
  • Gaussian hyperparameter search
• Deployment
  • Batch prediction
  • HTTP Endpoints
    • Request Format
      • Time stamp formats in requests
      • Using an existing DataFrame
      • Using data from a database
    • Transform Endpoint
    • Predict Endpoint