Data frame analytics documentation reorganization

[szabosteve]

Goal

The goal of this issue is to track the effort of creating a new, more user-friendly structure for the data frame analytics documentation. The documentation needs to present the natural “user flow” instead of exposing the machinery behind the features. After creating the user flow, we can revisit where we can reuse our existing materials and where we need to write pieces that are missing at the moment.

We want to provide a whole experience for the user and ensure that the docs support them in every step of the process.

This reorganization prepares the DFA docs for migration to the new docs system.

Reorganized TOC

• Overview [DOCS] Amends data frame analytics overview and adds resources section #1726
• Finding outliers [DOCS] Creates one pager for outlier detection #1732
• Predicting numerical values with regression [DOCS] Creates extended classification and regression sections #1736
• Predicting classes with classification [DOCS] Creates extended classification and regression sections #1736
• Determining language of text with language identification [DOCS] Creates extended classification and regression sections #1736
• Advanced concepts [DOCS] Amends data frame analytics overview and adds resources section #1726
  • How DFA works
  • DFA at scale
  • Feature encoding
  • Feature processors
  • Hyperparameter optimization
  • Feature importance
  • ML trained models
• API quick reference
• Resources [DOCS] Amends data frame analytics overview and adds resources section #1726
  • Limitations