Hyderabadi girl uses Machine Learning to crack Covid-19 pattern 

Hyderabadi girl uses Machine Learning to crack Covid-19 pattern 

She extracted data from John Hopkins University dashboard.

Published: 21st April 2020 10:09 AM  |   Last Updated: 21st April 2020 10:10 AM   |  A+A-

Express News Service

HYDERABAD: “I have done an analysis on Coronavirus (Covid-19) by using Machine Learning methods. Firstly, I checked the survivability rate among three age groups: 0-40 years, 41- 64 years and 65+ years by using Classification Algorithms in Machine Learning. Next, I checked the most common symptoms among the confirmed patients and dead patients.

I realised that few of the symptoms stayed until the patient died,” says Bhagyashree Kottoori, a Hyderabadi girl currently pursuing her Master’s Programme in Computer Science from Wayne State University, Michigan state in the United States. “There were 40 unique symptoms registered. I found the most common symptom sets among two categories i.e. confirmed patients and dead patients by using Association Analysis in Machine Learning,” she says.

She extracted data from John Hopkins University dashboard. What made her take up this analysis? She says, “I realised that not many people know about such analysis. So working on something that people aren’t aware of was my sole intention. I presented my results and model to my Professor. Dr Suzan Arslanturk in my Master’s program and she guided me through it,” informs this alumnus of CVSR College of Engineering in Hyderabad. She spent around three weeks to pre-process the data and implement my idea to get the final results.

“Next, I am keen to research and analyse how long India will take to reach a saturation point with the positive cases like it happened in China. If only the preventive measure like social distancing are followed well,” says this RK Puramgirl who moved to the US recently. She is now looking for the right platforms to publish her work.

Results of the analysis:
Prediction of common symptoms among confirmed patients category were fever, cough, sorethroat and fatigue.

Among dead patients, the common symptoms were fever, difficulty in breathing and
pneumonia.

Coming to the survivability predictions from my model: Age group 0-40 years has 98 percent of survivability rate
Age group 41-64 years has 83 percent of survivability rate.
Age group of 65+ years has 78 percent of survivability rate.

— Manju Latha Kalanidhi kalanidhi@ newindianexpress.com @mkalanidhi

flipboard facebook twitter whatsapp