This group project represents 15% of the final mark for this module. The submission deadline is April 25, 2025 (Week 10, Friday) by 5pm. A submission link will be provided on the LM Core one week before the deadline. University policies on late submission penalty and academic integrity will be followed.

Task Description:

Each group will apply two machine learning (ML) models to predict AU9999 gold prices on the Shanghai Gold Exchange. The objective is to critically evaluate how ML techniques can capture complex relationships between gold prices and multiple factors, while addressing real-world challenges in financial forecasting.

Report Structure:

1.   Introduction

2.   Data Collection & Preprocessing

•   Use reliable data sources https://libguides.lib.xjtlu.edu.cn/financial_data (e.g., CSMAR, Wind, FRED, Bloomberg, Yahoo Finance) with explicit citations.

•   Sample period: January 2000 – December 2024

•   Predictor selection: Include at least 5 predictors and justify your choices using economic/financial theory. For example:

o Macroeconomic: CPI, Treasury yields, GDP growth rates, etc.

o Commodity: Brent crude oil, LBMA silver prices, etc.

o Currency: USD/CNY exchange rate, USD Index (DXY), etc.

o Sentiment: S&P 500 volatility (VIX), gold ETF holdings, etc.

o Technical: 50/200-day moving averages, RSI, etc.

•   Use suitable data preprocessing methods to handle missing data, ensuring that all affected observations are explicitly documented. Select a data frequency (daily, weekly, or monthly) that aligns with the predictive requirements of your models and the availability of your chosen predictors.

3.   Model Development

•   Model selection: Choose two ML approaches and briefly describe them.

4.   Results & Discussion

•   Conduct a thorough statistical and graphical analysis of the dataset.

•   Identify correlations between gold prices and selected predictors.

•   Report and interpret model evaluation metrics, e.g., RMSE, MAE, and R² .

•   Visualize findings using appropriate charts and graphs to communicate insights effectively. Discuss the practical implications of your findings.

5.   Conclusion

•   Identify limitations and propose extensions.

6.   References

Other Requirements:

•   Reference:

o Use of APA or Harvard Referencing, both citation and reference

o Minimum of five (5) credible references, e.g., peer-reviewed journals, industry reports, textbooks, newspapers, magazines, etc.

o No references from Wikipedia, UKEssays, students’ papers, or other unreliable information sources

•   Generative AI:

o The use of Generative AI for code generation is permitted.

•   Submission package:

o Group report: 1000 words+/-20% excluding tables, figures, references, and appendix. Name the group report by using Group ID and the team  members’ family names, e.g., Group12_Wang_Li_Liu_Shen. Save the  report as a Word document.

o Data file: Save it as a CSV or Excel file containing all the data used for analysis.

o Python script: Ensure all codes are well-documented and reproducible.

o GenAI  statement:  Save  it  as  a  Word  document  containing  a  brief description of how Generative AI is used and a list of the key prompts.