Kelly Du Email & Phone Number

New York, Ny, Us

New York, New York, Us

• Led a team of 3 to develop a cancer prediction model using advanced machine learning algorithms such as Random Forest, K-nearest Neighbors, and Logistic Regression. Addressed class imbalance with Synthetic Minority Oversampling Technique (SMOTE), enhancing model accuracy by 21% and improving prediction reliability• Executed Principal Component Analysis (PCA) to transform a complex dataset with over 1000 features into a concise 30-component representation, simplifying data analysis and enhancing model performance by reducing computational complexity• Implemented fuzzy string matching technique to merge multiple datasets effectively, correcting inconsistencies to produce a dataset of 27 optimized features for more accurate analysis and modeling.• Delivered weekly progress reports and presentations on model developments and project milestones to managers and data scientists

New York, Ny, Us

• Directed the ETL (Extract, Transform, Load) process for over 2,000 lines of detailed financial time series data, including prices, dates, and ticker information, laying the groundwork for effective analysis• Achieved (high accuracy) up to 15.27% average Mean Absolute Percentage Error (MAPE) in forecasting daily stock prices for 2 months using SARIMAX (Seasonal AutoRegressive Integrated Moving Average with eXogenous factors) machine learning model in Python, significantly enhancing the model's accuracy and reliability• Worked collaboratively with a diverse team of data scientists and financial analysts, fostering an environment of mutual learning and innovation

New York, Ny, Us

• Conducted market research to identify and evaluate the top 5 M&A(Merger& Acquisition) targets with $2 million+ annual revenue and high potential (ROI, growth rate, etc.), analyzing market dynamics and trends to align with strategic investment goals• Visualized 30k+ customers’ segmentation, industry trends, and competitor analysis using SQL and Excel to inform strategic decisions• Prepared strategic expansion plans for 5 investment targets and presented actionable insights and recommendations to senior managers

Ithaca, Ny, Us

• Enhanced a Logistic Regression model's accuracy by 12%, achieving an 86% success rate, through optimization of feature selection and hyperparameters using Python• Leveraged BERT-based natural language processing (NLP) techniques to conduct sentiment analysis and text classification of Twitter posts, identifying and categorizing the latest news on natural disasters. • Prepared a comprehensive report analyzing the applications of Convolutional Neural Networks (CNN) in image recognition and object detection within computer vision

Orlando, Fl, Us

• Achieved 90% industry standard alignment in nationwide hydraulic conductivity of compacted soil predictions by implementing a Python-based Gradient Boosting model• Attained a statistically significant 0.878 R-squared value by optimizing Random Forest classifier through grid search and cross-validation techniques• Contributed as a co-author to the paper “Predicting Hydraulic Conductivity of Compacted Soil Liners/Covers Using Boosting Algorithm,” which was presented at The 9th International Conference on Water Resources and Environment Research

Master Of Science - Ms, Data Science

Bachelor Of Science - Bs, Major In Physics, Minor In Statistics