Roman Dong

Detail-focused Data Analyst with 1.5 years experience. Dedicated and hard-working with passion for Big Data. Team-oriented person who values communication with stakeholders and colleagues to achieve common goal. Currently holding a visa with full time work right.

Project Name: An Enhanced Autoencoder for Anomaly Detection in Imbalanced Financial Fraud Data 

Project Description: Developed an advanced anomaly detection model using autoencoders to improve financial fraud detection in highly imbalanced datasets. 

• Utilized ggplot2 in R and matplotlib in Python for exploratory data analysis and visualization, aiding in the interpretation and presentation of results.
• Designed and implemented an improved autoencoder model incorporating class-specific reconstruction losses and gradient clipping to enhance detection accuracy for both majority and minority classes.

Project Name: Accuracy Comparison among Five Machine Learning Algorithms for Financial Risk Evaluation 

Project Description: Conducted a comparative analysis of several machine learning algorithms such as Decision Tree and Random Forest to assess their efficacy in financial risk prediction. 

• Facilitated discussions on the practical implementation of these algorithms in the finance sector, demonstrating their potential through empirical analysis and model performance.
• Authored in-depth analyses on the mathematical foundations of Decision Trees, Logistic Regression, and Random Forests, elucidating their application in financial risk management.
• Developed and tested these models using financial datasets, applying metrics such as Accuracy, Precision, Recall, F1 Score, and AUC for comprehensive performance evaluation.

Tennis, Cycling and Console Games