Study

Course Resources

Here I’ve compiled some useful course resources that I discovered during my undergraduate studies at USTC, hoping they can assist students who are passionate about learning.

Mathematics and Statistics

Mathematical analysis
Linear Algebra
Discrete Mathematics
Probability Theory
Mathematical Statistics
Stochastic Process
- University of Science and Technology of China Course Evaluation Community
- MIT: Stochastic Process
Real Analysis
Complex Analysis
- Online Courses
- “Concise Complex Analysis” Textbook Exercises
Regression Analysis
- Course Home Page
- MIT: Regression Analysis
Convex optimization
- Understanding strong duality and weak duality, KKT
- Textbook Exercises (Extraction Code: z51k)
Time Series Analysis
Multivariate Statistical Analysis
- Multivariate Statistical Analysis from the USTC
- Applied Multivariate Statistical Analysis
Non-parametric Statistics

Computer Science

CS Self-study guide
- CS Self-study Guide
Data structure
- Data structure Junhui Deng
Computer Algorithm
Database Systems
- “Close Contact with MySQL” Notes Table of Contents
- Introduction to Database Systems, UC Berkeley
Git
- Runoob Tutorial
- Xuefeng Liao’s Git Tutorial
R Language
- Dongfeng Li’s R Language Tutorial
- R Markdown Tutorial
Python and Data Analysis
HTML CSS Javascript Vue
Machine Learning
Deep Learning

Sociology, History and Economics

Course Projects

Here I have listed some of the course projects I completed either individually or in collaboration with a group during my undergraduate studies. Most of these projects achieved the highest scores in the entire statistics department. If you are interested in these projects, please feel free to communicate with me via email or WeChat.

Multi-layer Community Detection with Network Embedding and Dependent Connectivity

Course Name: Fundamentals of Statistical Algorithm (Achieved the Highest Score in the Statistics Department)

Extended network embedding model for detecting inner and outer communities in directed multi-layer networks and incorporating correlation information between connections.
Designed an optimization iterative algorithm for the proposed model to maximize the joint likelihood function.
Conducted simulation experiments to evaluate the model’s performance and applied it to real-world datasets for research.

Time-series Forecasting Using ARIMA-GARCH and Neural Network

Course Name: Time Series Analysis (Selected for Exhibition as an Outstanding Course Work)

Researched contemporary popular time series ensemble models and learned signal processing algorithms like Empirical Mode Decomposition.
Expanded the VMD-GARCH/LSTM-LSTM model into a more diverse range of neural networks, exploring the performance of different neural networks.
Utilized MAE and RMSE metrics to compare our experimental results, finding that replacing the LSTM model with a classic RNN model significantly improved prediction accuracy on some datasets.

Vehicle Data Analysis and Driver Behavior Prediction

Course Name: Applied Statistical Software (Achieved the Highest Score in the Statistics Department)

Rectified measurement errors within experimental data and executed fundamental data processing.
Carried out a detailed analysis of vehicle networking data using statistical methods, including: box plot analysis, fitting error analysis, quantile regression analysis, time series analysis and cluster analysis.
Investigated the correlation between car speed change rates and wheel speed, and visualized the car gear’s time series diagram based on the findings.
Analyzed driver’s behavior habits and inferred road conditions, including road openess and the presence of traffic lights.

Zhengtao Gui