MCTagger-eval

Multi-Tagger work savings analysis

new_pipline_v3.py

Retrieves Multi-Tagger predictions for specified study designs.

Inputs:
-List of PMIDs that were abstract screened.
-List of PMIDs that were full-text screened.
-List of PMIDs that were included in the review.

Outputs:
-List of the Multi-Tagger prediction scores for each PMID.

PubMed data collection.py

Retreives MeSH terms and publication types from PubMed API.

Inputs:
-List of PMIDs for each review.

Outputs:
-List of PMIDs, their MeSH terms, and their publication types.
-List of PMIDs that were not found in PubMed.

DERP-reports-pubmed_data_web_rct_data.py

Performs calculations and compiles statistics for work savings analysis.

Inputs:
-PMIDs and tagger predictions output from new_pipeline_v3.py.
-List of MeSH terms and publication types for each PMID from 'PubMed data collection.py'.
-RCT scores pulled from the web RCT Tagger API.
-List of tagger prediction thresholds for each study design.
-List of included/relevant study designs for each review.

Outputs:
-Descriptive statistics tables (one based on filtering with Multi-Tagger alone; and one based on Multi-Tagger, web RCT Tagger, and MeSH terms and publication types).
-Graphs (number filtered out during full-text screening, percentage filtered out during full-text screening, number filtered out during abstract screening, percentage filtered out during abstract screening).
-List of included articles that were filtered out.
-Results files including all relevant predictions, MeSH categories, and publication types for each PMID.