ECON 5060 Economics 2023/24 Fall
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
ECON 5060
HKUST Department of Economics
2023/24 Fall
Group Project (Instructions and Guidelines)
Choose one from the given topics below:
|
Topics |
Target Variables |
File Names |
|
P2P Lending Default Prediction |
Status: Late, Repaid, Current Define numeric variable as follows: Late = 1 (default) Repaid = 0 (not default) |
P2P Lending Dataset
|
|
Predicting US Corporate Bankruptcy |
status_label: alive or failed Define numeric variable as follows: failed = 1 (bankrupted) alive = 0 (not bankrupted) |
US Bankruptcy Dataset |
|
Predicting Taiwan Corporate Bankruptcy |
Bankrupt: 1: bankrupt 0: not bankrupt |
Taiwan Bankruptcy Dataset |
|
Forecasting US Inflation Rate
|
CPIAUCSL Define inflation as the log first difference of CPIAUCSL: Inflation = log(CPIAUCSL(t)) – log(CPIAUCSL(t-1)) |
US FRED-MD Macro Dataset US FRED-MD_Appendix |
|
Forecasting US Interest Rate
|
FEDFUNDS: Federal Fund Rate
|
US FRED-MD Macro Dataset US FRED-MD_Appendix |
|
Financial Crisis Prediction |
crisisJST: 1: financial crisis 0: no crisis |
Financial Crisis Dataset Financial Crisis Chronology |
|
Forecasting Direction of Stock Market with Technical Indicators |
Define the weekly directional movement of the “Close” price of S&P 500 as follows: 1 if the weekly return of the index is positive, i.e., Close(t) > Close(t-5) 0 otherwise |
S&P 500 Index Dataset |
|
Predicting Stock Return with Fundamental Indicators |
return_adj_12m: 12-month adjusted return of share price after financial reporting period (i.e., difference between stock return and S&P 500 Index return) |
US Stock Fundamentals Dataset |
Content Requirements:
· Choose one from the given topics above
· Formulate the ML procedures or methodologies in addressing the topic
· Collect, compile, preprocess, and analyze the data
· Apply at least five different ML methods that you learn in this course to solve your ML task
· Summarize the findings, make conclusion and recommendations
Format Requirements:
· Word or PDF
· A cover page with title and group information (group number, student names and numbers).
· The structure includes an introduction (or executive summary), main body, conclusion, and a list of references.
· A maximum of 16 pages including the cover page, tables, charts, and references
· Font size 11 or 12, double spacing
Submission of Paper:
· Please email your term paper together with the code file to me by December 13.
Guidelines on Data Preprocessing
· Data preprocessing such as removing unnecessary features and transforming features into comparable ones across observations is highly recommended.
· You may need to transform some categorical features into numeric ones. There are three ways to do that, namely, Ordinal Encoding, One-Hot Encoding, Dummy Variable Encoding. Check https://machinelearningmastery.com/one-hot-encoding-for-categorical-data/ for details.
· If there is severe class imbalance in the data (e.g., less than 5% of instances belong to one class and the remaining instances belong to another class), you may undersample the majority class or oversample the minority class.
· If you undersample the instances in the majority class, you can randomly select the instances in the majority class so that the number of instances in the majority class matches the number of instances in the minority class.
· If you oversample the instances in the minority class, you can apply the Synthetic Minority Oversampling Technique (SMOTE).
2023-11-27