MixtComp Docker

Build

sudo docker build -t mixtcomp/docker .

Learn with default files (data.csv, model.csv, algo.json, param.json)

sudo docker run --rm  -v <data_directory>:/home/docker/data mixtcomp/docker

Predict with the learnt model

sudo docker run --rm -v <data_directory>:/home/docker/data mixtcomp/docker data/data.csv data/model.csv data/algo.json data/param.json data/resLearn.RData

Learn mode

4 files:

• data csv file with semicolon (;) as separator and the first line as column names
• model csv file with semicolon (;) as separator and the first line as column names
• algo json file containing the SEM algorithm's parameters (see associated documentation in RMixtComp or data[-functional]-example/algo.json file)
• param json file containing the mixtCompLearn/mixtCompredict function parameters (see associated documentation in RMixtComp or data[-functional]-example/param.json file)

The output is saved in <data_directory>/resLearn.RData.

Predict mode

5 files:

• same 4 first files as learn mode
• res containing the learnt model (".RData" file with a saved MixtComp object named res)

The output is saved in <data_directory>/resPredict.RData.