A Recommendation Model for Predicting Alzheimer's Drugs' Mechanism of Action

A simple binary prediction model that gets the Alzheimer's drugs' description texts as input. It classifies the drugs into two Small Molecules (SM) and Disease modifying therapies (DMT) categories.

Overview of encoding a document into a single CLS vector using BioBERT

A version of trained DT with term frequency features that is visualized. Blue

leaves indicate the DMT Biologics class and orange leaves stands for Small Molecules dt-eps-converted-to.pdf

Dataset

Download the dataset from [https://www.alzforum.org/therapeutics]

citation

@inproceedings{pouyan,
 title={A Recommendation Model for Predicting Alzheimer's Drugs' Mechanism of Action},
 author={Nahed, Pouyan and Kambar, Mina Esmail Zadeh Nojoo and Cacho, Jorge Ram{\'o}n Fonseca and Lee, Garam and Cummings, Jeffrey and Taghva, Kazem},
 booktitle={Intelligent Sustainable Systems - Selected Papers of WorldS4},
 pages={},
 year={2022},
 organization={Springer}
}