Jingxuan Lyu
Systems, research, visual thinking.
UC Berkeley. Computer Science and Data Science.
Selected Works
A short project selection.
Machine Learning Research
Predicting Fraudulent Transactions: A Comparative Study
A fraud-detection study comparing classical ML models, ensemble methods, and a CNN on large-scale transaction data.
Compared logistic regression, decision tree, random forest, XGBoost, stacking, and CNN approaches on the same fraud-detection task.
Built a preprocessing workflow with duplicate removal, one-hot encoding, class-ratio adjustment, and recursive feature elimination.
Exhibit 01
Model Atlas
2025
Project Signal
A fraud-detection study comparing classical ML models, ensemble methods, and a CNN on large-scale transaction data.
Stack
Python
pandas
scikit-learn
01
Assistive Technology
Braille Calculator Initiative
An accessibility-focused calculator project designed to help visually impaired students perform two-digit multiplication more efficiently.
Designed a calculator workflow to simplify two-digit multiplication for visually impaired students.
Improved software-side data transmission and iterated through user testing in real classroom settings.
Exhibit 02
Access Field
2026
Project Signal
An accessibility-focused calculator project designed to help visually impaired students perform two-digit multiplication more efficiently.
Stack
Accessibility Design
Prototyping
Software Testing
02
Sections
Contact