Master's Degree Workspace

Tasks

• Set up the operational system (Manjaro)
• Configure my personal Git
• Download the project
• Install and configure Pycharm
• Configure a virtual environment
• Install libraries and dependencies
  • Install Python
  • Install Sktime
  • Install other dependencies
  • Create a setup file to config the environment
• Download all timeseries datasets
• Centralize the receiving of datasets
• Test the experiments
• Create a main with all experiments
• Organize the folder's project
• Clean up the code of each variant
• Create a Abstract Class called Variant to keep duplicated codes
• Rewrite the ensemble variant and the function _fit_discretizers()
• ...

Search Technique - functions

fit( data, labels ){

    resolutions = resolutions_definition( data )
  
    number_of_samples_per_class = 2
  
    while( resolutions.size > 1 ){

        samples = get_samples( data, labels, number_of_samples_per_class )

        word_sequences = discretization_extraction( samples, resolutions.windows )

        ngram_sequences = ngrams_definition( word_sequences, resolutions )

        bag_of_bags = frequency_counter( ngram_sequences )

        resolutions_rank = calcule_separability( bag_of_bags, resolutions )

        resolutions = get_first_half( resolutions_rank )

        number_of_samples_per_class = 2*number_of_samples_per_class
  
    }
  
    word_sequences = discretization_extraction( data, resolutions.windows )

    ngram_sequences = ngrams_definition( word_sequences, resolutions )

    bag_of_bags = frequency_counter( ngram_sequences )
    
    clf = new LogisticRegression()
    
    clf = clf.fit( bag_of_bags, labels )

}

Search Technique - Sequence Diagram