Package: AIscreenR 0.1.1.9000

AIscreenR: AI Screening Tools in R for Systematic Reviewing

Provides functions to conduct title and abstract screening in systematic reviews using large language models, such as the Generative Pre-trained Transformer (GPT) models from 'OpenAI' <https://platform.openai.com/>. These functions can enhance the quality of title and abstract screenings while reducing the total screening time significantly. In addition, the package includes tools for quality assessment of title and abstract screenings, as described in Vembye, Christensen, Mølgaard, and Schytt (2024) <doi:10.31219/osf.io/yrhzm>.

Authors:Mikkel H. Vembye [aut, cre, cph], Thomas Olsen [ctb]

AIscreenR_0.1.1.9000.tar.gz
AIscreenR_0.1.1.9000.zip(r-4.5)AIscreenR_0.1.1.9000.zip(r-4.4)AIscreenR_0.1.1.9000.zip(r-4.3)
AIscreenR_0.1.1.9000.tgz(r-4.5-any)AIscreenR_0.1.1.9000.tgz(r-4.4-any)AIscreenR_0.1.1.9000.tgz(r-4.3-any)
AIscreenR_0.1.1.9000.tar.gz(r-4.5-noble)AIscreenR_0.1.1.9000.tar.gz(r-4.4-noble)
AIscreenR_0.1.1.9000.tgz(r-4.4-emscripten)AIscreenR_0.1.1.9000.tgz(r-4.3-emscripten)
AIscreenR.pdf |AIscreenR.html✨
AIscreenR/json (API)
NEWS

# Install 'AIscreenR' in R:
install.packages('AIscreenR', repos = c('https://mikkelvembye.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/mikkelvembye/aiscreenr/issues

Pkgdown site:https://mikkelvembye.github.io

Datasets:

On CRAN:

Conda:

gptopenaiscreeningsystematic-review

6.11 score 10 stars 7 scripts 237 downloads 16 exports 36 dependencies

Last updated 2 months agofrom:472378964d. Checks:4 OK, 4 NOTE. Indexed: yes.

TargetResultLatest binary
Doc / VignettesOKFeb 15 2025
R-4.5-winOKFeb 15 2025
R-4.5-macOKFeb 15 2025
R-4.5-linuxOKFeb 15 2025
R-4.4-winNOTEFeb 15 2025
R-4.4-macNOTEFeb 15 2025
R-4.3-winNOTEFeb 15 2025
R-4.3-macNOTEFeb 15 2025

Exports:approximate_price_gptgenerate_ft_dataget_api_keyis_chatgptis_chatgpt_tblis_gptis_gpt_agg_tblis_gpt_tblrate_limits_per_minutesample_referencesscreen_analyzerscreen_errorsset_api_keytabscreen_gpttabscreen_gpt.originalwrite_ft_data

Dependencies:askpassclicodetoolscpp11curldigestdplyrfansifurrrfuturegenericsglobalsgluehttr2jsonlitelifecyclelistenvmagrittropensslparallellypillarpkgconfigpurrrR6rappdirsrlangstringistringrsystibbletictoctidyrtidyselectutf8vctrswithr

Using OpenAI's GPT API models for Title and Abstract Screening in Systematic Reviews

Mikkel H. Vembye

Rendered fromUsing-GPT-API-Models-For-Screening.Rmdusingknitr::rmarkdownon Feb 15 2025.

Last update: 2025-01-16
Started: 2024-10-28

Readme and manuals

Help Manual

Help pageTopics
Approximate price estimation for title and abstract screening using OpenAI's GPT API modelsapproximate_price_gpt
RIS file data from Functional Family Therapy (FFT) systematic reviewfilges2015_dat
Function to generate dataset to be used for fine-tuning modelsgenerate_ft_data
Get API key from R environment variable.get_api_key
Test if the object is a ''chatgpt'' objectis_chatgpt
Test if the object is a ''chatgpt_tbl'' objectis_chatgpt_tbl
Test if the object is a ''gpt'' objectis_gpt
Test if the object is a ''gpt_agg_tbl'' objectis_gpt_agg_tbl
Test if the object is a ''gpt_tbl'' objectis_gpt_tbl
Model prize data (last updated November 5, 2024)model_prizes
Print methods for ''chatgpt'' objectsprint.chatgpt
Print methods for ''gpt'' objectsprint.gpt
Print methods for ''gpt_price'' objectsprint.gpt_price
Find updated rate limits for API modelsrate_limits_per_minute
Random sample referencessample_references
Analyze performance between the human and AI screening.screen_analyzer
Generic function to re-screen failed title and abstract requests.screen_errors
Re-screen failed requests.screen_errors.chatgpt
Re-screen failed requests.screen_errors.gpt
Creating a temporary R environment API key variableset_api_key
Title and abstract screening with GPT API models using function calls via the original function call argumentstabscreen_gpt.original
Title and abstract screening with GPT API models using function calls via the tools argumenttabscreen_gpt tabscreen_gpt.tools
Function to write/save fine tune dataset in required jsonl formatwrite_ft_data