TODO.md

In tpq/exprso: Rapid Deployment of Machine Learning Algorithms

Works in progress

• regrso expansion:
• [] Add RegrsArray class
• [] Add RegrsModel class
• [] Watch out for x[,,drop=TRUE] errors
• [] Make ExprsBinary $defineCase a factor?
  • This change alone crashes downstream code
• [x] Make calcStats tidy
• [x] Add 2D plot
• [x] Rename getProbeSet function.
• [x] Rename probes as top

• [x] Rename top.N as top

• ExprsMulti expansion:

• [x] Fix splitStratify for multi-class
• [x] Fix compare for multi-class

• [x] Implement all build_ ExprsMulti methods

  • Uses doMulti for 1-vs-all classification
  • Add ExprsModule predict method
  • Add multi-class calcStats

• pl expansion:

• [x] Consider "random Plains" wrapper
• [] Consider plRFE with embedded RFE
• [x] plGridMulti
  • Performs fs before each 1-vs-all build
• [x] Disable ROC plotting during high-throughput pl

• [x] plGridPlus with "better" plCV?

  • Achieved with plNested(plNested) method

• fs expansion:

• [x] Remove doMulti fs methods?
• [] F-test

• [x] ANOVA

  • Add as a true multi-class method

• build expansion:

• [] Logistic regression
• [] Democratic SVM
• [] KNN

tpq/exprso documentation built on July 27, 2019, 8:44 a.m.