Improved methods for Predictive Process Monitoring:

This project is an extension to work done by irhete/predictive-monitoring-thesis as a master thesis at the University of Tartu, Estonia. In this thesis, we introduced three different methods to improve the currently existing techniques of outcome-oriented predictive process monitoring.

This repository and our contributions are organized as follows:

• Contribution 1: Adding CatBoost method to outcome-oriented PPM techniques.
• Contribution 2: Adding a new complex sequence encoding technique on the basis of discrete wavelet transform and neural networks.
• Contribution 3: Adding Inter-case features to the experiments.

Dataset:

You need to download below datasets and modify the path in dataset_confs.py script.

• Labelled datasets
• Inter-case features datasets

Reproduce results:

• To run this project, you need to install packages from requirements.txt file, and to do so run the below command:

              conda create --name env_name --file requirements.txt --yes
  

• Above command will create a virtual environment, so you need to activate it afterwards using below command:

              conda activate env_name
  

• To reproduce results for each contribution you need to go inside the corresponding folder, and then run the following commands:

  1. Hyperparameter optimization:

                          python experiments_optimize_params.py <data set> <bucketing_encoding> <classifier> <nr_iterations>
     

  2. Training and evaluating final models:

                          python experiments.py <data set> <bucketing_encoding> <classifier>
     

  3. Execution times of final models:

                          python experiments_performance.py <data set> <bucketing_encoding> <classifier> <nr_iterations>
     

• For Example_CatBoost:

              cd ./CatBoost/
              python experiments_optimize_params.py production single_laststate catboost 1
              python experiments.py production single_laststate catboost 
              python experiments_performance.py production single_laststate catboost 1
  

• For Example_Wavelet:

             cd ./Wavelet/
             python experiments_optimize_params.py production single_waveletLast catboost 1
             python experiments.py production single_waveletLast catboost 
             python experiments_performance.py production single_waveletLast catboost 1
  

• For Example_Inter-case features:

              cd ./Inter-case_features/
              python experiments_optimize_params.py production single_laststate catboost 1
              python experiments.py production single_laststate catboost 
              python experiments_performance.py production single_laststate catboost 1