TABLE OF CONTENTS

  • Tutorial to learn basics of statistics by David Mease
  • Tutorial to learn advanced data mining skills by MIT professor
  • My experience in healthcare at Program for experimentation and Theoretical Modelling ,Loyola Medical  Center University
  • Text Mining Project 
    • github links
    • documentation
    • power point presentation

Watch this series from Google by David Mease to understand the basics of data mining . 

Once you have gained sufficient knowledge follow this course from MIT

In Healthcare sector , modelling of drug interactions with viruses is usually done through ODE's.I built models for ODE's on Berkeley Madonna and R.Beginners would usually found Berkeley Madonna fairly straightforward and simple but learning it on R has its own advantages as more people know R rather than Berkeley Madonna.Below I am attaching a screenshot of output from Berkeley Madonna.

This is an output of an ODE which is used to model the interaction between the virions,infected cells and T-cells(uninfected cells).

d/dt(I)=(1-h)*b*T0*V-k*delta*I  

d/dt(V)=(1-e)*p*I-m*c*V

For more information on the symbols and constants please see Development of Mathematical Models for the Analysis of
Hepatitis Delta Virus Viral Dynamics
Bruno C. de Sousa1*, Celso Cunha2

Text Mining 

The objective of this project is to classify tweets and at the same time extract what themes are talked about.Latent Dirichlet Allocation is one such technique used to classify documents and extract themes from them.Information such as which hospital organizations are talking about floods, mosquitoes or tech brands like Apple, Samsung are talking about which technology. OLED is a technology that both the brands incorporated in their latest phones.

  • Web crawler using R programming language is built. RSelenium is a package which can be used to automate browsers like chrome, firefox
  • The code for web crawler can be is published on github/aadishchopra/textMining
  • Extracted tweets are cleaned using text mining techniques which can be found easily online. "tm" package can be used
  • Topic Modelling using LDA to generate themes from handles 

To get an overview of what I have done, you can click this link to open up the slides.