January 2018 - April 2018, Web based interface for Autism School, SMILES.
- Developed a website for an Autism School to store, calculate, and maintain the medical history of autistic children at the SMILES Foundation, Hyderabad, India.
- The website was built using NodeJS and for the database MongoDB was used.
November 2017, April 2018, Binary Sentiment analysis using Naive Bayes Classifier, Multi-class Logistic Regression with Parameter Optimization (using JAVA), BITS Pilani Hyderabad, India.
- Code for Logistic Regression and 2-Layer Neural Networks can be found here.
- A binary sentiment analyzer was developed in JAVA using modifications of the Naive Bayes classifier and a multi-class logistic regression algorithm was implemented with grid search to deal with any generic dataset.
September 2017, Implemented Random Forest, ID3 with pruning in JAVA, Project done as part of the machine learning course at BITS Pilani Hyderabad, India.
- The model works for any categorical data-set. The execution time of the model is comparable to the CARET Package of R.
August 2017-December 2017, Fractal Theory, Design Project, BITS Pilani Hyderabad, India, Guide: Prof.Sharan Gopal.
- Completed a study/design project on Fractal Theory in order to understand the properties of fractals and visualize them using programming tools.
- I am also exploring the applications of fractal dimensions for fast feature selection.
March 2017, Student Healthcare Database using MongoDB, Project done as part of the Database Systems course, BITS Pilani Hyderabad, India.
- A student health care database was created for the students on campus using MongoDB.
- The front end was a website, created using NodeJS. Students could update their health records, and query relevant information such as, 'Contact Details of students with a certain Blood Group' through the website.
November 2015, Program to Find the Shortest Path to various places within college, Project done as part of the Graph Theory course, BITS Pilani Hyderabad, India.
- The C++ program was based on 'Dijkstra's Shortest Path Algorithm', with weighted edges.
- The weights for the edges were generated keeping in mind the 'Sun-Index', a measure of the intensity of sun light on the path.