Paper Description

This contains the codebase and the annotated relevant trial set for each of the 25 queries, for the paper titled Towards an Aspect-based Ranking Model for Clinical Trial Search accepted in the 8th International Conference on Computational Data and Social Networks (CSoNet 2019).

Paper link:

UMLS concept based retrieval and aspect based ranking model of clinical trials

Requirements

1. Download the dump of the clinical trials from the link
2. Setup QuickUMLS tool
3. Download the adversity events reported from the site
4. Base Line Setup
   1. Download and setup Elastic search for the baseline.
   2. For the baseline code refer
5. All environment packages are listd in [requirements.txt] file.

Folder strucutre & description

allAnnotatedData

1. It contains the 25 annotated queries.
2. 5 queries for each disease class.

baselineSetup refer

1. Scripts for indexing trials for each class of disease and script for retriving and ranking trials on the basis of query.

bigTableAndSmallTable

1. Calculates the precision, speraman's rank order correlation and overlap across 25 final queries retrieved trials and ranked on the basis of relevancy(5-methods)

commonPatientSearchedTerms

1. QuickUMLS tool applied over 1440 lexicons of medDRA common patient terms
2. Finds the problematic queries.

datasetPreparation

1. Get PubMedIds for the clinical trials.
2. Map clinical trials linked with PubMed-Ids to different classes(26) of disease with all extra fields(Adversity, popularity) appended.

goldStandardDataTREC_2018

Robustness study

1. Precision@10
2. Recall

rankedTrialsOnDifferentRelevancyBasedSchemes

1. Contains trials in ranked order on the basis of different relevancy (pageRank, Exact Match, SynSet) based approach.

rankingOnTheBasisOfRelevancyAndAddingColumnsForInclusionExclusionCounts

1. Rank retrieved trials for 25 quries on the basis of different relevancy(25) methods.

All_Importan_Data_And_Dump_Data

1. 5 class Csv Files with all fields appended.
2. Pickle File of UMLS concepts for each trial across 5 disease classes.
3. Ranked trials on the basis of 5 relevancy based methodologies.
4. Dump files.

clusteringSourceFiles

1. Contains application of different clustering algorithm like DBScan, Affinity Algorithms on different variations of the data.

Citation information

If you use the codes or the dataset, please cite the paper.

Bibtex

@InProceedings{10.1007/978-3-030-34980-6_25,
author="Roy, Soumyadeep
and Rudra, Koustav
and Agrawal, Nikhil
and Sural, Shamik
and Ganguly, Niloy",
editor="Tagarelli, Andrea
and Tong, Hanghang",
title="Towards an Aspect-Based Ranking Model for Clinical Trial Search",
booktitle="Computational Data and Social Networks",
year="2019",
publisher="Springer International Publishing",
address="Cham",
pages="209--222",
doi="10.1007/978-3-030-34980-6_25",
isbn="978-3-030-34980-6"
}