Option-Implied Portfolio Skewness and Kurtosis (B. Tech Project)
Option-Implied Portfolio Skewness and Kurtosis (B. Tech Project)
Embedded Files
Project Guide: Prof. Piyush Pandey | B.Tech Project | Department of Mechanical Engineering & Shailesh J. Mehta School of Management, IIT Bombay)
Initiated a project on analyzing option prices as indicators of future market movements, focusing on deriving implied portfolio skewness and kurtosis, currently engaging in a deep dive into option theory, volatility, and pricing methodologies.
Conducting extensive literature review on existing models and papers related to option-based market predictions, laying the groundwork for developing an innovative approach to market analysis using option pricing.
Skills: Option Pricing Models · Literature Reviews · Statistics · Python
Glacier Monitoring Based on Sentinel-I RS Data (GNR 618: Remote Sensing and GIS Applications to Cryosphere)
Glacier Monitoring Based on Sentinel-I RS Data (GNR 618: Remote Sensing and GIS Applications to Cryosphere)
Project Guide: Prof. Gulab Singh | GNR 618: Remote Sensing and GIS Applications to Cryosphere | Centre of Studies in Resources Engineering
Developed a comprehensive classification framework for monitoring glaciers, focusing on remote areas such as Tyndall Glacier, Antarctica, and Greenland, using Sentinel-1 RS data for detailed snow and ice analysis.
Implemented dual classification approaches—unsupervised K-means clustering and supervised Random Forest classifier—to accurately distinguish between five distinct snow and ice types on glacier surfaces.
Leveraged image segmentation techniques in analyzing iceberg formations, enabling precise calculation of vital physical parameters like volume and surface area, crucial for understanding iceberg dynamics.
Conducted a systematic study of glacier and iceberg movements in the Patagonian Ice Sheet over 6 to 12 months, revealing significant seasonal variations in accumulation and ablation rates.
Executed a three-step supervised classification process—including preprocessing for enhanced color differentiation, detailed annotations for model training, and final classification using Random Forest—enhancing the accuracy of ice-water demarcation in remote-sensed imagery.
The analysis of the formation and drifting of icebergs using machine learning algorithms has shown promising results in predicting the trajectory of these floating hazards. This has practical implications for shipping and offshore operations, as well as for understanding and mitigating the impact of climate change on polar regions. In the future, this technology could be further developed to improve predictions and forecasting of iceberg movements, potentially reducing risks and improving safety in these regions.
Skills: Data Classification · Image Segmentation · Machine Learning Algorithms
Poker Push-Fold Strategy Solver (IE 616: Game Theory and Decision Analysis)
Poker Push-Fold Strategy Solver (IE 616: Game Theory and Decision Analysis)
Guide: Prof. Urban Larsson| IE 616: Game Theory and Decision Analysis | Dept of IE & OR, IIT Bombay
My classmate and I collaborated on a project spurred by our interest in an article drawing parallels between trading and poker. This led us to explore the intricate mathematics of poker. Motivated by my second-place finish in our institute-wide poker tournament, I enrolled in IE616 - Decision Analysis and Game Theory. In the course, I developed a poker push-fold strategy solver for heads-up poker. Using the Nash equilibrium, we calculated Expected Value (EV) for pushing all-in with various hands against opponent calling ranges. Through comparisons of EV, factoring in pot odds and hand equity via Monte Carlo simulations, we established guidelines for optimal plays. This project showcased the nuanced interplay of game theory, probabilistic modeling, and decision-making in poker.
Skills: Game Theory · Monte Carlo Simulation · Mathematical Modeling · Statistical Data Analysis · Python
Traffic Lights Signal Optimization (ME 308: Operations Research Project)
Traffic Lights Signal Optimization (ME 308: Operations Research Project)
Guide: Prof. Avinash Bhardwaj | ME 308: Operations Research Project | Dept of Mechanical Engineering, IIT Bombay
During my Industrial Engineering and Operations Research (ME308) course, I undertook a project centered on Traffic Lights Signal Optimization, implementing algorithms such as Max Pressure, Self-Organizing Traffic Lights, and the Genetic Algorithm using Python. Validated through SUMO, a traffic simulation platform, the effectiveness of these algorithms in minimizing congestion and optimizing waiting times emerged unmistakably. When considering real-world deployment, while Max Pressure and SOTL would require advanced sensors for real-time feedback, the Genetic Algorithm offers flexibility, being less sensor-dependent but demanding considerable computational power. The hands-on experience reinforced my appreciation for optimization techniques.
Skills: Algorithm Development · Optimization · Modeling and Simulation · Python
Colour Predictive Model for Laser Colour Marking (ME338:Manufacturing Process)
Colour Predictive Model for Laser Colour Marking (ME338:Manufacturing Process)
Guide: Prof. Deepak Malra| ME338:Manufacturing Process II Department of Mechanical Engineering, IIT Bombay
I undertook a significant project centered around the predictive analysis of color acquisition through laser processing. Employing real-world data, I engineered a sophisticated Random Forest model to discern the probability of color manifestation based on varied laser parameters. The project included a meticulous exploration of the dataset using advanced visualization techniques, uncovering precise parameter ranges where color outcomes were reliably predictable. Additionally, I played a pivotal role in the development of an Artificial Neural Network (ANN), enhancing predictive capabilities. This ANN not only forecasted RGB values corresponding to specific laser parameters but also demonstrated versatility by inversely predicting laser parameters based on desired RGB outcomes. This project represented a fusion of data science expertise and domain knowledge, contributing to advancements in predictive modeling within the context of laser technology.
Skills: Artificial Neural Networks · Data Analytics · Predictive Modeling · Data Visualization · Python
Prediction of Crude Oil Prices (Predictioneer, AZeotropy)
Prediction of Crude Oil Prices (Predictioneer, AZeotropy)
Self Project | Predictioneer, AZeotropy IIT Bombay
In my third year at IIT Bombay, I applied theory to practice, venturing into the challenge of forecasting crude oil futures. I implemented both Long Short-Term Memory (LSTM), using TensorFlow's Keras, and the ARIMA model. Through a univariate approach based solely on price data and leveraging techniques like early stopping, the LSTM model was particularly effective. The real excitement came when testing the model against real-time data; seeing my theoretical knowledge come alive and interact with live market data was profoundly rewarding and thrilling, marking a significant milestone in my initial journey into this field.
Skills: Time Series Analysis · Long Short-term Memory (LSTM) · Backtesting · Financial Markets · Python
Prediction of Results of Premier League Matches
Prediction of Results of Premier League Matches
Guide: Prof. Biplab Banerjee| DS 303: Introduction to Machine Learning | C-MInDS, IIT Bombay
Created a custom database consisting of relevant attacking and defensive statistics of all Premier League Teams over seven years, normalizing all data for every 90 minutes to improve accuracy
Predicted the result by using two simple Linear Regressors to predict goals scored by each team
Skills: Linear Regression · Sports Analytics · Python (Programming Language)
Page updated
Google Sites
Report abuse