Research Projects

This project examines various machine learning models for classifying text (restaurant and movie reviews) and images (CIFAR-10 dataset). It compares the accuracy of traditional models like Naïve Bayes and SVM with deep learning models like CNNs across different setups. The analysis highlights the best approaches for each type of data and discusses the implications of model complexity and dataset characteristics on performance.

Developed predictive model with 71.01% accuracy to forecast flight delays, clustered airports to enhance efficiency, and used Dijkstra’s algorithm for shortest flight paths, leading to fuel savings. Optimized intra-state connectivity with Kruskal's algorithm, reducing total delay by 93%, and created a linear programming model that minimized delay penalties by 20.15%.

We explored the effectiveness of data in predicting consumer preferences compared to traditional creative methods. Using advanced web data analytics, we analyzed customer feedback, sentiment from platforms like Reddit, and sales metrics from Google Shopping. Our goal was to understand the digital presence of brands and evaluate how data could influence their strategic direction.

Developed a predictive real estate model leveraging XG Boost Regressor, integrating web-scraped market data with existing datasets to forecast daily store visits, achieving a MAPE of 13.3%, enabling strategic retail location decisions.

Developed a predictive model for bankruptcy using 64 financial indicators, an ensemble of XGBoost and neural networks, with 91% accuracy, to significantly improve financial risk assessment.

Optimized Craigslist's classification system by creating an algorithm combining LSTM and Random Forest for Text and Image Classification respectively, reducing misclassifications by 31%.

Developed an LSTM model to generate text, mimicking the style of Nietzsche's writings.

Enhanced a local caterer’s pricing strategy by conducting detailed conjoint and Van Westendorp analyses, leading to optimized subscription models and menu offerings for increased profitability and customer satisfaction.

Developed an advanced scoring and network model using Bayesian analysis to identify Key Opinion Leaders (KOLs) for a neurology product launch, utilizing clinical trials data, research publications, leadership positions, and open payment data. Secured a top-three position among 42 teams.

Predicted consumer activity type for the NCAA Women’s Basketball Tournament using an ensemble of XG Boost, Random Forest and Logistic Regression, achieving an accuracy of 98.5%. Team secured 3rd place at the university level.

Developed a Predictive Fraud Detection framework combining Time Series Analysis, Rule Based Heuristics, and K means Clustering, enabling a major telecom client to identify and eliminate 50% of fraudulent clone accounts in their network, thus preventing an annual revenue loss of approximately $7 million.

Developed a Bidirectional LSTM model for predicting next day closing prices with a MAPE of 19% and built a dynamic portfolio optimization algorithm incorporating profit-taking and stop-loss strategies.

Leveraged BERT to distinguish AI generated from human written texts in a Kaggle competition, using datasets from Mistral AI (to train the model) and achieving 72% training and 58% testing accuracy.

Forecasted Airbnb 'Super host' status in Chicago with an 84% accuracy using Logistic Regression and assessed potential returns on investment employing the Herfindahl Index for strategic investment insights.