HOPER/case_study.yaml at main · serbulent/HOPER · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
parameters:
    choice_of_module: [case_study]
    #********************case_study Module********************************
    module_name: case_study
    choice_of_task_name:  [prepare_datasets,model_training_test,prediction] #prepare_datasets,fuse_representations,model_training_test,prediction
    fuse_representations:
        representation_files: [./data/hoper_case_study_example_data/representation_files/node2vec_d_50_p_0.5_q_0.25_multi_col.csv,./data/hoper_case_study_example_data/representation_files/multi_modal_rep_ae_multi_col_256.csv]
        min_fold_number:  2
        representation_names:  [modal_rep_ae,node2vec]
    prepare_datasets:
        positive_sample_data:  ["/media/DATA2/kemal/HOPER/hoper_case_study_example_data/prepare_datasets/positive.csv"]
        negative_sample_data:  ["/media/DATA2/kemal/HOPER/hoper_case_study_example_data/prepare_datasets/neg_data.csv"]
        prepared_representation_file:  ["/media/DATA2/sinem/isik_makale_1003/low_elenmis_yeni_datasetler/all_transfer_ae/test_fused_output_transfer_ae_2col_eklenmis.csv"]
        representation_names:  [Transfer_ae_unfiltered]
    model_training_test:
        representation_names:  [Transfer_ae_unfiltered]
        scoring_function:  ["f_max"]  #"f1_micro",f1_macro","f1_weighted"
        prepared_path:  []
        classifier_name:  ["Fully_Connected_Neural_Network"]
    prediction:
        representation_names:  [Transfer_ae_unfiltered]
        prepared_path:  ["/media/DATA2/sinem/HOPER/prediction_data.csv"]
        classifier_name:  ['Fully_Connected_Neural_Network']
        model_directory:  ["/media/DATA2/kemal/HOPER/case_study/case_study_results/training/Transfer_ae_unfiltered_Fully_Connected_Neural_Network_binary_classifier.pt"]