screenmed: A Function which is grouping abstracts in terms of similarity...

In thdiakon/screenmedR: Filtering Pubmed Articles for a meta-analysis.

A Function which is grouping abstracts in terms of similarity with a small group of relevant publications.

Description

It provides the following: 1) The number of groups the user choose to split the initial search 2) The cosine similarities between the small group of publications and each of the clustered groups of abstracts 3) A list of the clustered abstracts and the cluster they belong to. 4) The PMIDS of the publications that no abstract was found in Pubmed.

Usage

screenmed(initial_search, filtered, sparsity, group_number)

Arguments

initial_search	A character vector containing the PMID numbers of the articles of our search.
filtered	A character vector containing the pmid numbers of abstracts (4 or 5 typically) that belong to the study.
sparsity	A number typically in (0,99,1) that defines the sparsity of the abstracts for clustering.
group_number	The number of groups the user would like to divide the initial number of publications from pubmed search.

Value

group_number,cosine_similarity,clustering,missing_abstracts

thdiakon/screenmedR documentation built on Nov. 22, 2022, 6:26 p.m.