Skip to content

collab-uniba/sensor-based-classifierTuning

Repository files navigation

sensor-based-classifierTuning

How to use the scripts

  • Tuning_10k : Script to build and train the classifier with 10 fold cross validation

  • Run_HoldOut_10k : is the Script to Run 10 times the script Tuning_10k; after that the result will be 10 model trained with ten fold cross validation

nohup ./run_HoldOut_10k.sh <input.csv> models/models.txt <output_folder> <label> &> log.txt &

where, <input> is the input dataset, <output_folder> is the name of the folder where saving the output, <label> is the class to classify (valence for example).

For example:

nohup bash run_HoldOut_10k.sh Train_Empatica_10sec_valence_noicse.csv models/models.txt results_10sec_valence_noicse valence &>10sec_noicse.txt &
  • Results parsing.ipynb The script essentially parses and summarizes the performance metrics from multiple text files, calculates mean values, and exports the summarized data to a CSV file. This is useful to uderstand which is the best model!

  • TestModel : is the script to Run the trained model on the test Dataset

Rscript testModel.R <input_test.csv> <best_model.rds> <prediction.csv> <result.csv>

where, <input_test.csv> is the input testing dataset, <best_model.rds> is the best model trained with the command before, <result.csv> is the name of the file where saving the output metrics.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •