x2007tro / ShinyMLHome

Machine learning home shiny app

ShinyMLHome

Machine learning home shiny app

Model addition

To add a new model, do the following steps:

Preparation

1. Add a parameters table with the format * Input # : xxx Parameters * to app.accdb
   • Add relavant parameters to the table
2. Add ui file reference to main.R (ui)
   • The files don't exist yet, they will be created in step 6 and 7
3. Add server file reference to main.R (server)
   • The files don't exist yet, they will be created in step 6 and 7
4. Add code in ./global.R to read the table created above
5. Add code in ./global.R to load library and ./Model/mxxx.R
   • mxxx.R doesn't exist yet, it will be created in step 6 and 7
6. Copy and paste the .R files below
   • ./iUI/model_xxx_par.R
   • ./iUI (server)/model_xxx_grid.R
   • ./iUI (server)/model_xxx_bayesian.R (if applicable)
   • ./iUI/model_xxx_plot.R (if applicable)
   • ./Model/mxxx.R
7. Rename the above files to reflect the appropriate model name

Update ui files

7. Update ./iUI/model_xxx_par.R
   • update the tab object name - tp_model_xxx_par
   • update the parameter variable name - xxx_pars
   • update the ui component name - mxxxp
8. Update ./iUI/model_xxx_grid.R
   • update the tab object name - tp_model_xxx_grid
   • update the ui component name - mxxxg
   • add/remove ui output depending on the model
9. Update ./iUI/model_xxx_bayesian.R (if applicable)
   • update the tab object name - tp_model_xxx_bayesian
   • update the ui component name - mxxxb
10. Update ./iUI/model_xxx_plot.R (if applicable)
    • update the ui component name - mxxxb
    • add/remove ui output depending on the model

Update model file

10. Update ./Model/mxxx.R

    • replace the model part of the function names
    • update score_board related codes (two places) to include model specific parameters
    • update function CoreTrainxxx2 for specific model
    • update bayesian search for specic model
      • update input parameters for function BsOpUtil
      • update field names for data.frame mdl_pars_bayesian

11. Update ./Helper/ml_helper.R (It might be very easy to do, refer to package pdf)

    • update function PredictMe
    • update function FormatData4Model

Update server file

12. Update ./iServer/model_xxx_grid.R
    • update model name parameter - mdl_nm
    • update the parameter variable name - xxx_pars
    • update the ui component name - mxxxg/p
    • replace GridSearchxxx2 with new model name
    • add/remove ui output depending on the model
13. Update ./iServer/model_xxx_bayesian.R (if applicable)
    • update model name parameter - mdl_nm
    • update the parameter variable name - xxx_pars
    • update the ui component name - mxxxb/p
    • replace BayesianSearchxxx2 with new model name