Projects
Obesity Prediction
The project aims to understand and address the impacts of obesity on the population by designing and training machine learning models to comprehend and predict obesity trends. The data for this project was collected from Mexico, Peru, and Colombia, and includes various attributes related to eating habits, physical activity, and other lifestyle factors. The ultimate goal is to integrate the insights derived from these models into practical initiatives for policy formulation, resource allocation, intervention design, and preventive programs. This comprehensive approach aims to enable a deeper understanding of obesity prevalence and incidence, facilitating the creation of effective and targeted health strategies to promote a healthier society.
Read MoreCar Loan Price Estimation
This project centers on a case study for the e-car auto loan organization, aiming to evaluate and optimize the profitability of e-cars through auto loans. We reviewed the business model and past data, identifying that e-cars offered high interest rates to high-risk customers, resulting in low loan acceptance rates (22%). Our objective is to increase the acceptance rate to over 50% by adjusting pricing strategies and terms. Through value-based and cost-based pricing strategies, we aim to balance profit margins with transaction volume. Our data exploration indicated that lowering APR to match competitors' rates, extending loan terms, and enhancing marketing could improve acceptance rates. By employing logistic regression and causal inference models, we predicted loan outcomes and optimized pricing, achieving an acceptance rate of 66% and significantly increasing net profit. The findings highlight the importance of aligning loan offerings with market rates and optimizing loan terms to improve acceptance rates and profitability. Continuous evaluation and adjustment of pricing models based on market trends and customer behavior are essential for sustained success
Read More
Text to Video Generation
Welcome to my ongoing journey with the Open-Sora AI video generator model! Open-Sora revolutionizes video production by transforming text into high-quality educational or promotional videos, making it accessible for those without professional video editing skills. After experimenting with various models on platforms like Hugging Face and GitHub, I decided to delve deeper into the intricacies of video generation models to enhance their architecture. My roadmap includes in-depth exploration of models, diversifying training data, boosting capabilities with more GPUs, and integrating audio. Despite challenges like GPU limitations, I've made significant progress in generating videos at lower resolutions. This project is not only a technical challenge but a creative adventure, aiming to push the boundaries of AI video generation. Stay tuned as I continue to navigate this exciting field and contribute to the evolving landscape of AI technology.
Read More
EmpowerLend AI - Business Agenda
The EmpowerLend AI platform is designed to support women-owned small businesses by providing fair and accurate loan assessments through AI technology. The platform not only helps in assessing loan eligibility but also offers financial literacy tools and mentorship programs. This project includes a Streamlit prototype that demonstrates the loan assessment process and a user roadmap to guide entrepreneurs on their journey.
Read More