Blogs

---

BioMedical Data - A Sea to Explore

This blog is just an overview of what I think of how Bio Medical Data can be explored and how can use Machine Learning to get some useful insights out of it.

Y = MX + C

The above equation is a linear equation, what are your thoughts when you see this equation. In every Machine learning class or “A.I” class, maybe the main focus is on M and C but the crucial part is X - DATA. Most of the biomedical data contains the direct information for eg. Have a look at the ECG signal.

Fig 1

Reading the Signals

All details are present in the few data points on the left and right of R or the peak in the above figure 1. It’s not necessary to be a cardiologist; we just need the data points.

As these raw data points make up the information, I made use of it for preparing a feed-forward network using MATLAB (ANN). As a beginner, I made use of this raw data for detecting the disease patterns.

I knew nothing about the preprocessing & data science at all. But however I knew only one thing, differentiating factors, different patterns possess distinct data-points.

Lets check some of the examples:

Visually the differences can be plotted; such data points can be directly fed to the Neural Networks for classification. This helped me get started into the importance of data and Data Science.

Here's a paper I have written detect various patterns using neural network

Detecting the patterns of coronary artery disease and hypertension using artificial neural network

Hope you enjoyed this blog. Please follow us on facebook , LinkedIn. Have a great day!