乐趣区

关于程序员:QBUS6840

2019S1 QBUS6840 Group Project Page 1 of 5
QBUS6840 Assignment 2 – Group Assignment:
Due dates: Friday 24 May 2019
Value: 30%
Rationale
This assignment has been designed to help students develop valuable communication and
collaboration skills and to allow students to apply their predictive analytics skills on a real
world datasets.
Description
The assignment will be done in groups of 5 (or 4 or 6 depending on the total number of
students in the class) without exception. The group can be formed freely or assigned by the
Coordinator. Please get close contact with your members in earlier stage. A group leader will
be automatically assigned by the Canvas system.
Notes

  1. The assignment will be done in groups of 5 (or 4 or 6 depending on the total of students
    in the class) without exception. The group can be formed freely or assigned by the
    Coordinator. Please get close contact with your members in earlier stage. A group
    leader for each group shall be automatically assigned on Canvas.
  2. The assignment is due at Friday 16:00pm 24 May 2019. The late penalty for the
    assignment is 5% of the assigned mark per day, starting after 16:00 pm on the due date.
    The closing date, 31 May 2019, 16:00pm is the last date on which an assessment will
    be accepted for marking.
  3. Your answers shall be provided as a word-processed. Prepare one single report. Do not
    have separate report for each question/task. Add your Python code as appendix to the
    report. At the same time, we will ask you to upload your python code to your Canvas
    folder.
  4. Your report should include the Group ID and SIDs of all group members. No names!
    You may stay with the report cover sheet provided.
  5. You need to provide full explanation and interpretation of any results you obtain.
    Output without explanation will receive zero marks.
  6. Be warned that plagiarism between individuals is always obvious to the markers of the
    assignment and can be easily detected by Turnitin.
  7. The data sets for this assignment can be downloaded from Canvas.
  8. Presentation of the assignment is part of the assignment. Markers will assign up to 10%
    of the mark for clarity of writing and presentation. It is recommended that you should
    include your Python code as appendix to your report, however you may insert small
    section of your code into the report for better interpretation when necessary. Think
    about the best and most structured way to present your work, summarise the procedures
    implemented, support your results/findings and prove the originality of your work.
  9. Numbers with decimals should be reported to the Second decimal point.
    2019S1 QBUS6840 Group Project Page 2 of 5
  10. The report should be NOT more than 25 pages, with font size no smaller than 11pt,
    including everything like text, figure, tables, small sections of inserted codes etc but
    excluding the cover pages, appendix containing Python code and the meeting minutes.
    A violation of this rule will attract a penalty of 5% marks.
    Meeting Minutes
  11. Your group is required to submit meeting minutes, which are to be attached to the
    report as the second appendix. Your group may use the attached templates for preparing
    agendas and meetings minutes.
  12. You should document at least 3 meeting minutes for this group assignment, using the
    template provided/or a template you choose. Each minutes should at least record the
    following information:
    a. Meeting dates/time/venue/duration;
    b. Key points of the process of discussion such as who said what (key points);
    c. Action list and responsible members, task due time etc
    d. Review/group judgement on the quality of individually completed/responsible
    tasks; The purpose of this is to infer whether a member is doing their share of
    jobs
    Note: Any unsatisfactory meeting minutes may attract a penalty of up to 15% of
    the total marks.
  13. In case of a problem within a group we will request minutes of the previous meetings.
    We can make an individual adjustment to the group mark if there is sufficient evidence
    that a student has done very little. If the student has truly done little, we will award a
    mark of 0.
    Peer Assessment, Marks and Feedback
  14. We may ask for peer assessment from each student. The instruction how to do this will
    be released later on.
  15. Each group will be awarded a group mark per the marking criteria. In some cases,
    individual marks may be applied if there is dispute in a group and the quality or
    quantity of contributions made by individuals are significantly different, in which cases
    the unit coordinator will seek peer assessments reports from individuals in a group and
    meeting minutes.
  16. We will allocate 15% marks for competition among the groups. The group with the
    highest test score will secure full 15% marks while other groups will secure a mark
    according to their test score against the best test score.
  17. Feedback will be provided on the marked submission.
    Background and Dataset
    The S&P/ASX 200 index is used as the benchmark for Australian equity performance. It is a
    market-capitalisation weighted and adjusted stock market index, see Wikipedia. ASX 200
    index is calculated based on the 200 largest ASX listed stocks. It starts from 31 March 2000
    with a value of 3133.3
    The ASX200 historical and current data can be downloaded from
    https://au.finance.yahoo.com/…^AXJO. The historical data can be
    downloaded at three different frequencies of Daily, Weekly and Monthly between any
    2019S1 QBUS6840 Group Project Page 3 of 5
    specified duration since 31 March 2000. In this project, your group will be asked to analyze
    the data at both frequency of daily and monthly. To align with the due date of the project, the
    following strategy is recommended:
  18. Use 31 March 2000 as the starting date for downloading
  19. The end date can be set as a date when your group starts work on the project. You
    may explore this dataset and do all the possible tasks for the project.
  20. The Project due date is 24 May 2019. You can download the most recent full dataset
    on 23 May 2019. You may use this dataset to finalise your best model and make all
    the predictions. Even the data on 24 May 2019 may become available before you
    submit your report, you shall not use it at all.
  21. For the group(s) who may be granted extension due to unforeseen reasons, the
    competition forecasts will be the next FIVE days from the granted due date.
    The dataset shall contain information of dates (daily or months), open price, highest price,
    lowest price, close price and adjusted close price. Your work in this project is to analyze the
    time series of close price. A set of Daily and Monthly datasets have been downloaded for
    your convenience if you don’t wish to download it at early stage. You can get them directly
    from Canvas.
    Tasks
    Please note most tasks are deliberately designed open. This gives more freedom for you to
    explore a better solution.
    Data Pre-processing: Conduct initial analysis over the entire data. Write python program to
    clean up the data, e.g., checking/deleting incomplete information if any, to make sure data is
    complete, or normalising the data, etc. It is up to you how to normalise or transform the data
    so that the resulting dataset can be well incorporated in training your chosen model(s). You
    MUST retain your python program (or code section) used for all the pre-processing work.
    Exploratory Time Series: Analyze the entire time series for both daily and monthly data.
    You may plot them or do what you can to reveal any patterns. Summarise what you have
    revealed or observed. In your report, carefully present your analysis and findings. Any
    different patterns between Daily and Monthly time series?
    Benchmark Model: Based on what you have found from exploring the time series, consider
    using a classical model to build your benchmark model for forecasting. This could be the
    moving average, or decomposition method, exponential moving average etc. It is always a
    good idea to split the given time series data into a training subset and a validation subset.
    Document your findings and justification. This should be done for both Daily and Monthly
    time series.
    Build Advanced Models: You are requested to build at least TWO advanced models such as
    ARIMA, State Space Model, Deep Neural Networks and Recurrent Neural Networks etc.
    This is your choice. In building your chosen models, you need to at least optimise models in
    terms of e.g. the orders in ARIMA or State Space Model, and/or other parameters as well.
    Simply building a model without any consideration of validation and tuning hyperparameter
    2019S1 QBUS6840 Group Project Page 4 of 5
    does not meet the minimal requirement for this task. Document your findings and
    justification. This should be done for both Daily and Monthly time series.
    Competent Model and Final Result(s): Finally, according to your work, decide your best
    model for both Daily and Monthly time series. For all of them, please make five forecasts
    ahead. For the monthly time series, you may report your five forecast in a table in your
    report, but we ask you to forecast five daily results (i.e., the forecasts for Monday 27 May to
    Friday 31 May 2019) and save your results into a csv file containing two columns, one for the
    date (named Dates) and the other column for the predicted values with the second decimal
    point. Name your file as GroupXXX_Results_ASX200.csv. where XXX is your group
    number in form of e.g., 008 (for Group 8), 085 (for Group 85) and 123 (for Group 123). The
    results will be assessed against the actual close prices on the dates in order to decide your
    group performance among the entire class (competition!). It is important for you to name
    your csv file in the above format, otherwise our program may fail your results.
    Note:
  22. The score will be based on the mean squared error.
  23. The close price on 24 May 2019 may become available, but you shall not use that
    information to train you model. In the forecasting stage, for some models, you may
    need predict the close price on 24 May 2019, then predict close prices for the next
    five days based on the prediction for 24 May 2019. However, the forecast for 24 May
  24. will not be used for assessment in competition.
    Presentation
    Please submit your project through the electronic system on Canvas
    The assignment material to be handed in will consist of a final report that:
    i) Takes a research article form in which you shall have a number of sections
    such as introduction, methodology, experiment results, findings/interpretation,
    and conclusion. All references should be properly cited and take a full
    bibliographical format. Here are couple of examples
    http://cs229.stanford.edu/pro…
    http://cs229.stanford.edu/pro…
    http://cs229.stanford.edu/pro…
    ii) Details ALL steps and decisions taken by the group regarding requirements
    above.
    iii) Demonstrates an understanding of the relevant principles of predictive
    analytics approaches used.
    iv) Clearly and appropriately presents any relevant graphs and tables.
    The MAXIMUM page limit is 25 pages, including any computer output, graphs,
    tables, etc.
    2019S1 QBUS6840 Group Project Page 5 of 5
    Your group is required to submit meetings minutes. Your group may use the attached
    templates for preparing agendas and meetings minutes. You should document at least
  25. meetings during the semester. Documentation should be in terms of attendance,
    discussion points, actions decided, review etc. You may use your own form or find
    something online.
    You, as a member of a group, may be also required to submit your peer assessment.
    Please use the attached criteria sheet and assessment form for this purpose. You will
    be informed of how to use online form when it becomes available.

http://www.daixie0.com/conten…

退出移动版