Skip to content

Latest commit

 

History

History
65 lines (61 loc) · 5.54 KB

Project.md

File metadata and controls

65 lines (61 loc) · 5.54 KB
Raw

Course Project (2018-19 M1)

Team List (Presentation Order)

Data Proposal (10.23 Tues)

  • Form a group (up to 4 students) and select data set
    • No restriction on data set. However, business(fin/ma/econ) related data is welcome (extra credit for creative data selection and pre-processing)
  • Create a designate repository GITHUB_ID/PHBS_MLF_2018 for your team
    • Tick Initialize this repository with a README and select python under .gitignore
    • Put team members (student # and github ID), brief description of data, and goal of the project in README.md (refer to markdown cheetsheet)
    • Put the data under /data folder (if too big, put some samples)
    • In lieu of submission, invite jaehyukchoi as a collaborator (under Setting > Collaborators & teams)
  • In the class, use README.md for brief presentation (3 min per team)

Project Guidline

  • Report should be consist of the summary in README.md and the execusion in python notebooks .ipynb. ( .pdf, .ppt, .doc NOT accepted.)
  • In the README.md summary,
    • You may update your proposal file.
    • briefly describe your motivation, goal, data source, result and conclusion.
    • A few figure or table for summary is recommended.
    • Use links to data or .ipynb files (see past year examples below)
  • In the .ipynb execusion,
    • Put command cell and edit cell (comments) in a balanced way. (Do not only put code!)
    • Put a brief table of contents with links (example: PML)
    • You may breakdown code into several .ipynb files by function (e.g., data cleaning, learning, result analysis). In that case, make sure to save intermediate result into file so that I can run the later steps (result analysis) without running previous steps (data cleaning, learning).
    • The use of .py file should be strictly restricted to function or class only. (Do not put any learning procedure in .py)
    • I should be able to reproduce the result from your code. Your code should run with no error. Code with error will be severely deduct your score. Make sure to run your code in a new session.
  • Other considerations:
    • Make sure the workload within team is balanced. (Add your team members to collaborators, each team members committ codes, etc)
    • There should be no secret component (e.g., stock trading strategy)
    • Creativive (out-of-textbook) ideas are recommended for better result or result analysis
  • Deadline for updating report is 11.13 Tues Midnight (11:59 PM)

Past year's project (2017-18 Module 3 Under Topics in Quantitative Finance: Machine Learning for Finance)

  • JiayuCai: Cross-currency-basis prediction
  • AtomMe: The Prediction of Credit User's Overdue event Based on Machine Learning Method (Competition)
  • ZheshengZhang: LSTM-based method for stock returns prediction
  • diyawang: Factors affecting bank competitive power
  • JOY199603: Forecasting the Price Change on IPO day
  • zhang-yunhe: Predicting default of credit cards clients (UCI)
  • Louie-Lin: Credit card default (UCI)
  • callmebyd: Lending Club Loan Data (Kaggle)
  • zsq96512: FX rate prediction
  • stuartphbs: Wine Quality (UCI)
  • yipanglin: NBA MVP prediction
  • DengQingqin: Sentiment Measures on Stock Market
  • LeiZHANG1995: Bitcoin price prediction
  • evanleungc: Predict low-risk profitable trading opportunity with high frequency trading data
  • labro: An investment model based on stock price info and stock revivews on the intenet
  • YedaDu: Idiosyncratic Market Value Factor: explaining market value by machine learning methods
  • UvoChow: Sentiment Analysis of Guba Forum