Skip to content

Latest commit

 

History

History
65 lines (48 loc) · 3.13 KB

BioGITOM_Usage.md

File metadata and controls

65 lines (48 loc) · 3.13 KB

Reproducibility

  1. Usage

    To execute a specific task using BioGITOM, run the following command in your terminal:

    python run_biogitom.py --task <task>

    Replace <task> with the appropriate task name.

    Example: To run the omim2ordo task:

    python run_biogitom.py --task omim2ordo

    List of Predefined Tasks

    BioGITOM includes the following predefined ontology matching tasks:

    1. omim2ordo: Aligns the omim ontology with the ordo ontology.
    2. body: Matches the snomed.body ontology with the fma.body ontology.
    3. ncit2doid: Maps the ncit ontology to the doid ontology.
    4. neoplas: Aligns the snomed.neoplas ontology with the ncit.neoplas ontology.
    5. pharm: Matches the snomed.pharm ontology with the ncit.pharm ontology.

    Each task is organized under the Tasks/ directory, containing all necessary resources such as implementation scripts, and results.

  2. Outputs and Results Visualization

    Executing a task generates the following output files in TSV format:

    • {task}_all-predictions: Includes all candidate mappings with their respective mapping scores.
    • {task}_matching_results: Contains the filtered mappings based on a predefined threshold.
    • {task}_all_predictions_ranked: Lists all candidate mappings considered for rank-based metrics calculation, along with their mapping scores.
    • {task}_formatted_predictions: Provides predictions reformatted specifically for calculating rank-based metrics.

    Instructions for Evaluating Ontology Matching Results

    • Evaluating Global Metrics (Precision, Recall, F1-Score) To evaluate the generated mappings included in the {task}_matching_results.tsv file in terms of global metrics such as Precision, Recall, and F1-Score, use the following command:

      python Scripts/evaluate_global_metrics.py --task <task> --src_ent <src_ent> --tgt_ent <tgt_ent>

      Replace <task>, <src_ent>, and <tgt_ent> with the corresponding task name, source ontology name, and target ontology name.

      Example:

       python Scripts/evaluate_global_metrics.py --task omim2ordo --src_ent omim --tgt_ent ordo

    • Evaluating Ranked-Based Metrics (MRR and Hits@k) To assess the {task}_matching_results.tsv file in terms of ranked-based metrics such as Mean Reciprocal Rank (MRR) and Hits@k, use the following command:

       python Scripts/evaluate_ranked_based_metrics.py --task <task> --src_ent <src_ent> --tgt_ent <tgt_ent>

      Replace <task>, <src_ent>, and <tgt_ent> with the corresponding task name, source ontology name, and target ontology name. Example:

       python Scripts/evaluate_ranked_based_metrics.py --task omim2ordo --src_ent omim --tgt_ent ordo

    • Viewing Logs for a Specific Task Task execution logs can be found in the Tasks/ directory. Each task folder includes a file named Results_log_{task}.md, which contains detailed logs of the task's execution.