Mia Lai's Projects
This project is to predict credit card customer default risk using data from April to September 2005 for customers in Taiwan. It involves preprocessing the data, including cleaning and transforming, and employing algorithms like Apriori for association analysis and Random Forest for modeling to predict next month's defaulters.
This project focused on leveraging data analysis techniques to optimize purchasing patterns and enhance consumer engagement within an e-commerce platform. By querying a dataset of over 100,000 users, the aim was to identify key insights and trends that would drive strategic decision-making and improve marketing strategies.
This project leverages LLaMA-2, a powerful and versatile language model, as the foundation to create a fine-tuned Large Language Model (LLM) for a medical AI chatbot using a dataset of 5,000 medical questions and answers involves several steps, each crucial for ensuring the model's accuracy, relevance, and safety.
This analytical project delves into the diverse ecosystem of the Google Play Store, exploring various dimensions of mobile Apps ranging from app categories and user ratings to installation counts and reviews. Utilizing Python to perform a structured data analysis and sentimental analysis that draws insights into user behavior and app trends.
The Food System Optimization Analysis Project utilizes advanced modeling and technology to enhance global food systems focusing on sustainability and fairness. Employing methods like BWM, EWM, and genetic algorithms, it develops the 4SD Model, providing actionable insights for improved food security.
This study analyzed 440 SME on Shenzhen's exchange using a Cash-Cash Flow sensitivity model, with regional and enterprise heterogeneity analysis showing fintech's significant impact on financial constraints.
A side project created by Sadhanha Anand, Cindy Jeon, and Mia Lai developed a solution for Lux(conceptual electronics company) to minimize delivery disruptions caused by weather. The project integrates real-time weather data from OpenWeatherMap with Lux's delivery system, facilitating strategic adjustments in the face of extreme weather conditions.