Skip to content

This package applies the different aspects of SEARCH (store, explore, assess, reduce, confirm, and holistic) to initial machine learning results to an unknown data set.

License

Notifications You must be signed in to change notification settings

IdahoLabResearch/Digital-Twin-Analytics-SEARCH

Repository files navigation

Digital-Twin-Analytics-SEARCH

Data scientists use machine learning to analyze and predict information for digital twins, but the selection of an appropriate model is time consuming, repetitive, requires specialized knowledge, and can lead to an unintended bias or preference for the chosen model. The SEARCH (Store, Explore, Assess, Reduce, Confirm, and Holistic) digital analytics platform will provide initial data exploration of an unknown data set. The initial framework will perform analysis on a given dataset to pre-process the data, determine which algorithms are candidates for analysis, perform dimension reduction, validate results over multiple imputations of the data, provide documentation for ease of use, and display the analytics on a graphical user interface (GUI).

About

The SEARCH package currently supports regression analysis.

High Dimensional Models

  • Principal Componet Regression (PCR)
  • Elastic Net Regression (ENET)
  • Ridge Regresssion (RIDGE)
  • Least Absolute Shrinkage and Selection Operator Regression (LASSO)
  • Random Forest (RF)
  • Support Vector Machine (SVM)
  • MultiLayer Perceptron (MLP)

Low Dimensional Models

  • Includes all high dimensional models plus ...
  • Linear Model (LM)
  • Generalized Additive Model (GAM)

Getting Started

  1. Install R

  2. Install RStudio

  3. Install Miniconda

  4. Install package dependencies

Run install_dependencies_rbased.R
  1. Build and check a package in devtools
devtools::check()
  1. Install the SEARCH package
Build -> Install 
  1. Navigate to Shiny App
inst/shiny/app.R
  1. Start the Shiny app
Run App

About

This package applies the different aspects of SEARCH (store, explore, assess, reduce, confirm, and holistic) to initial machine learning results to an unknown data set.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages