it21208 / MeLIR

🔎 𝗗𝗲𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻: The objective of a Technology Assisted Review (TAR) system is to minimise the period and cost of reviewing documents by decreasing the amount of human review overhead/effort required on documents categorised as possibly non-relevant. An IR pipeline methodology to minimise the human effort as well as the time and cost associated with certain stages of a Systematic Review such as retrieving, ranking and screening medical documents.

In the qrels folder the file "full.train.abs.2017.2018.2019_and_full.test.abs.2019.qrels" contains both training and test qrels.

For more information check the README.txt file in the scripts folder of this project.

Author: Alexandros Ioannidis

University of Strathclyde, Glasgow (UK)

©Copyright 2020 by Alexandros Ioannidis

To install the requirements execute the command below from the root folder of the project.

pip install -r requirements.txt

It is recommended that the viewer reads the sources below to acquire an improved understanding of the context of this repository.
⚫ Information Retrieval: https://en.wikipedia.org/wiki/Information_retrieval
⚫ Machine Learning: https://en.wikipedia.org/wiki/Machine_learning
⚫ Reinforcement Learning: https://en.wikipedia.org/wiki/Reinforcement_learning
⚫ Active Learning: https://en.wikipedia.org/wiki/Active_learning_(machine_learning)
⚫ Neural Network: https://en.wikipedia.org/wiki/Neural_network
⚫ Artificial Neural Network: https://en.wikipedia.org/wiki/Artificial_neural_network
⚫ Deep Reinforcement Learning: https://en.wikipedia.org/wiki/Deep_reinforcement_learning
⚫ Technology Assisted Review: https://en.wikipedia.org/wiki/Electronic_discovery#Technology-assisted_review