Predicting the pre-occurrence of acute graft-versus-host disease before transplantation using machine learning algorithms can lead to improved care plans for patients receiving allogeneic stem cell transplantation.



About CDSS


This CDSS is the final product of a Ph.D. thesis conducted at Shahid Beheshti University of Medical Sciences, Tehran, Iran that approved by Iran National Committee for Ethics in Biomedical Research with Approval ID: IR.SBMU.REC.1397.128

Published articles from this project are:

Vast and Reliable Database

The participants included 182 Allogeneic Hematopoietic Stem Cell Transplantation recipients with 51 attributes transplanted at Taleghani hospital, Tehran, Iran between 2009 and 2017.



Algorithms used

In this project we have use various algorithms like Ensembles, Naive Bayes, SVM, Random Forest, and etc. in training our various machine learning models.



Analysis Using Python and Spyder

The code used for analysis of data and getting prediction rates is pretty simple. The code is written by Python in Spyder, Pycharm, Jupyter, Atom and Thony IDEs.