Integrating Factors of 1st order Ordinary Differential Equations

This was my Bachelor's dissertation. Here, we find a correlation between the integrating factors of a first order ordinary differential equation and infinitesimal Lie group transformations.We show that corresponding to each integrating factor of an Ordinary Differential Equation (ODE), we can find a symmetry group under which the ODE is form-invariant.

Organization: St. Xavier's College, Kolkata

Diabetic Retinopathy Detection

Created an automated analysis system capable of classifying each retinal image to one of the classes (normal, mild, moderate, severe and proliferative), corresponding to the severity of the disease. Compared the performance of various approaches and models performing this classification task.

Organization: Chennai Mathematical Institute

Facial Recognition using Eigenfaces

We compared characteristics of a face image to that of face images of known individuals and non-face images. The approach treats the images as two-dimensional, rather than requiring three-dimensional geometry, by taking advantage of the fact that faces are generally upright, and may thus be described in terms of two-dimensional characteristics. The method used is Eigenfaces, where we project the images into a low-dimensional feature space using Principal Component Analysis.

Organization: Chennai Mathematical Institute

Day Ahead Solar Photovoltaic Forecasting of Power Output : A Comparative Study

The project presents a comparative study of the prediction models, generated using the techniques of ARIMA, SVM, ANN and XGBoost, for forecasting day ahead power output of Photovoltaic systems. The aim is to give a general overview of the relative performance of the methods, thus evaluating how machine learning techniques perform relative to traditional time series forecasting techniques. The paper also revolves around finding those important features which are an integral part of day ahead energy productions.

Organization: Chennai Mathematical Institute, Delab Research

Unreeling Netflix

We built a movie recommendation engine on Apache PySpark. The project demonstrated how Netflix uses a recommendation system based on Big Data to recommend movies to its users. It focused on collaborative vs content-based filtering for reccomendation.

Organization: Chennai Mathematical Institute