Multi Variate Time Series Forecasting for Gold Prices in lieu of COVID-19

- Python, Darts, TFT, TCN, XGBoost, LSTM

- Demonstrated a statistically significant correlation of 0.53 between Gold Prices and COVID-19 case numbers.
- Executed Time Series forecasting and dynamic model optimization, incorporating new case and vaccination data to predict Gold Prices.
- Achieved a Mean Absolute Percentage Error (MAPE) of less than 3 over a three-month period using Temporal Fusion Transformers, Temporal Convolutional Neural Networks, XGBoost and LSTMs.

Exploratory Data and Regression analysis for US-CH portfolios based on macroeconomic conditions

- Python, statsmodel, matplotlib

- Gathered and visualized data from various domestic and international sources, including the World Bank and International Monetary Fund (IMF).
- Conducted exploratory data analysis based on government bond yields, unemployment rates, inflation, and other factors, establishing a statistically significant relationship between US and Swiss macroeconomic conditions and the performance of Swiss Equities (p-values near 0).


- Flutter, Dart, Python, Django, GraphQL, Cassandra

- India faced a massive vaccine shortage for the COVID-19 vaccine from April 2021 through July 2021, which caused many people seeking the vaccine to be turned away at vaccine centres leading to crowding and confusion.
- This tool was created to aid the people of my country to have more detailed information about the availability of vaccines at vaccine centres.
- At its peak, the website and progressive web app helped as many as thousand people to have accurate information regarding the vaccines thus helping slow the spread of the virus.

Stock Market Prediction

- Python, Pandas, NumPy, Scikit-learn, Matplotlib, TensorFlow

- Reduced Test Loss: Achieved a low test loss of 0.0594, indicating a high level of accuracy in the stock market prediction model.
- Minimized Test MAE: Obtained a Mean Absolute Error (MAE) of 0.192 on the test data, demonstrating the model’s effectiveness in minimizing prediction errors.
- Managed Large Dataset: Successfully analyzed a large dataset to predict market fluctuations, with the model’s performance validated on test data.

College Social Media Platform

- Flutter, Django, Django REST Framework

- A platform for students to exchange information via instant messaging, images, and videos, providing an efficient communication tool for the college.
- This app offers a modern interface based on cutting-edge UI/UX research.

Sun Spot Prediction

- Python, Pandas, NumPy, Scikit-learn, Matplotlib, SKLearn

- Implemented ARIMA Model: Achieved superior forecasting accuracy over a baseline model in predicting sunspot activity.
- Evaluated Model Performance: Reduced Mean Squared Error (MSE) and Mean Absolute Error (MAE) in the test data predictions.
- Visualized Predictions: Plotted actual vs predicted values, providing a clear visual representation of the model’s performance.

Quiz Exit Survey

- Django, React, JavaScript, Python, Django REST, PostgreSQL

- There was no central portal for conducting quizzes and surveys in our college, which led to many complications for students and teachers.
- This platform solved this issue using Django as a backend and PostgreSQL serving as the database.
- It provided students with the required auto-signed proofs of completion, as well as teachers with the necessary analysis.

Exploration-Exploitation problem in Policy-Based Deep Reinforcement Learning for episodic and continuous environments

- International Journal of Engineering and Advanced Technology

GANs and VAEs as methods of synthetic data generation and augmentation to enhance heart disease prediction

- International Journal of Engineering and Advanced Technology

An Overview of Causal Inference and its Applications in Health-care and Finance using methods such as Bayesian Networks and Granger’s Causality

- International Journal of Emerging Technologies and Innovative Research