Siqi Wang

 Avazu Ad Click Prediction – Kaggle Challenge
Date: July 2023

• Engaged in a supervised machine learning project to predict ad click-through rates.
• Developed XGBoost model through analysis with large-scale datasets and comparisons between classifiers.
• Learned and adapted strategies from champion's solution.
• Reduced log loss to 0.405.

Cartpole Problem Solution using Unsupervised Machine Learning
Date: July 2023

• Successfully solved the cartpole balancing problem with neural network.
• Self-implemented policy gradient algorithm using policy ascent as optimization.
• Achieved convergence on 300 iterations.

Householder QR Factorization - Python Implementation
Date: April 2023

• Self-studied and implemented the Householder QR factorization method.
• Employed both graphical and mathematical approaches to evaluate the time complexity.
• Conducted comprehensive comparisons with other QR algorithms, focusing on stability and execution time.