Prescription Suggesting System For a Health Condition

Owned by Amrit Singh Bains (Unlicensed)
Last updated: May 19, 2014 by piyush dane (Unlicensed)

Overview: 

The project Idea is to use machine learning algorithms to learn the prescription patterns for a health Condition and then use the results to help the doctors by giving the suggestion for the prescription.

Main Features:

  • Given the database for  patient treatment use the data mining technique to learn the prescription pattern for diagnosis.
  • Make a service that can use the results from above feature and suggest the prescription for a diagnosis.

  • A general architecture to use any database by the service.

  • A general architecture of the project to add further new machine learning module to expand the domain of search and results.

Learning Algorithm:

Association Rule Mining will be used as the machine learning Algorithm for the 1st chunk. This is a Research Paper which use Association Rule mining to find relationships between procedures performed on patients and reported diagnosis. In the 1st chunk of this project, we are planing to find the relationships between the health condition of patient and prescriptions they have been provided.

Implementation details:

  • Requirements-

    • Machine learning library - Weka (Language : java).

    • Server - Tomcat(java).
    • JDBC driver to access the database.

  • Modules-

    • Pre-Processing for database.

    • Learning Module for Association Rule Mining on Prescriptions.
    • Prediction module for Association Rule Mining on Prescriptions.
    • Rest API for querying.

Architecture:

This is the Architecture of the project that I'm about to implement.

To Edit or Leave your comments please go to this link and to login use userid - piyush.dane.29@gmail.com, password - geek001.

 

Time Line :-

19/05/14 - 23/05/14 - Familiarize with weka and OpenMrs database.

24/05/14 - 30/05/14 - Complete the preprocessing module to load the database and get the useful data to be converted into format required by Weka.

31/05/14 - 6/06/14 - Complete the learning module. Save the results of Weka in a HashTable.

7/06/14 - 8/06/14 - Won’t be able to to work since I’ll be travelling to europe.

9/06/14 - 13/06/14 - Make the prediction module for the Association rule mining.

14/06/14 - 20/06/14 - Make a REST API to query a diagnosis and get the results from association rule mining module.

--------------------------------------------MID TERM EVALUATION-------------------------------------------

Till this part will have a service up in running for raxa to use.


23/06/14 - 4/07/14 - make a generalised Sencha application to use the service.

5/07/14 - 18/07/14 - Make changes in the architecture of the project to make it as much as generalised as possible so that other machine learning modules can be incorporated later with this project.

19/07/14 - 25/07/14 - Travelling back to india and going home for a week. Wont be able to work much in this week.

26/07/14 - 1/08/14 - Work on the documentation of this project.

2/08/14 - 22/08/14 - Make the machine learning module for idea 2.


Mentors: 

  • Surajit Nundy
  • Piyush Madan.

 

Â