讲解 BUS265 Machine Learning and Digital Technology 2023/2024讲解 Python编程

BUS265

Machine Learning and Digital Technology

2023/2024

No   Module Outcome Learning Description

1   A1   At the end of this module, students will understand the role of artificial intelligence, machine learning, and text analytics in business research.

2   A2   At the end of this module, students will understand the difference between supervised versus unsupervised learning algorithms.

3   A3   At the end of this module, students will be able to employ, evaluate, and interpret results from major machine learning algorithms used in business research.

4   A4   At the end of this module, students will be able to build basic supervised prediction models and understand the bias-variance trade-off and its implications for applications in international business.

5   A5   At the end of this module, students will understand the process of model building and evaluation, be able to find the best model for a particular business problem and address potential issues of overfitting.

6   B1   Demonstrate a clear understanding of the basics of artificial intelligence, machine learning, and text analytics relevant to business and international business.

7   B2   Develop quantitative skills including machine learning algorithms, interpretation and extrapolation.

8   B3   Develop coding skills in Python.

9   B4   Organise and complete an individual or group project.

10   B5   Analyse complex data.

11   C1   Explain, interpret, present and critique machine learning analysis in the area of international business.

12   C2   Use quantitative and text data confidently and competently (e. g. descriptive statistics, machine learning algorithms).

13   C3   Apply analytical and problem-solving skills to produce evidence-based and creative thinking.

Assurance of Learning (selected modules only): contribution to Programme Level Learning Outcomes (see programme rubric map)

No Programme Learning Outcome (LO) Description

1   LO 1.1   Develops an informed argument on the selected topic/issue based on appropriate sources

2   LO 1.2   Communicate original ideas and/or findings in a concise and coherent manner in an essay, report or examination

3   LO 2.1   Collects and analyses information and data from a variety of sources

4   LO 2.2   Applies relevant theories to a specific case or scenario

5   LO 2.3   Proposes an effective solution based on the supporting evidence

6   LO 3.1   Develops an understanding of different research methods and approaches

7   LO 3.2   Applies appropriate statistical techniques to a specific case or problem

8   LO 3.3   Develops an understanding of different research methods and approaches

9   LO 4.1   Demonstrates awareness of global, environmental, social and ethical implications

10   LO 4.2   Links academic knowledge to professional practice

11   LO 5.1   Develops awareness of the international business environment, socio-political debates and globalization

12   LO 5.3   Manages the time effectively and meets the deadlines

13   LO 5.4   Promotes a culture of integrity, ethics, diversity and inclusion

Assessment instructions for students (as per QMPlus ‘Assessment Information’ tab)

1. The module learning outcomes being assessed [See above table]

2. Instructions and guidance

Your individual research report is an opportunity for you to apply machine learning concepts and techniques to an interesting business analytics problem of your choice. Your report must be about machine learning concepts and techniques learned in this module, including supervised and unsupervised machine learning, text analysis, and network analysis.

The report should follow the business analytics process: (1) Business Understanding, (2) Data Understanding, (3) Data Preparation, (4) Modelling, (5) Evaluation. Importantly, before you focus on data and machine learning techniques, you need to first convert a business problem into an analytics solution answering the following key questions: (i) What is the business problem? (ii) What are the goals that the business wants to achieve? (iii) In what ways could an analytics model help to address the business problem (e.g., prediction)? Then, another essential part of your report is to connect your business problem to a suitable dataset and appropriate machine learning techniques in Python and Colab.

Topics (Please select ONE topic)

1. Customer churn prediction

2. Insurance fraud prediction

3. Sentiment analysis of review data

Datasets

Below are the datasets (corresponding to the topics above) you can use (Please select ONE dataset):

1. Mobile Phone Operator (Acme Telephonica) Customer Churn. The data can be accessed here. For details about the dataset, see slides from Kelleher et al. 2020 here.

o A cleaned data file is available via the following link and can be loaded directly in Colab via the link using Pandas: https://gitlab.com/valdanchev/data-storage-for-teaching-ml/-/raw/main/ACMETelephoneABT_cleaned.csv

2. Motor Insurance Fraud Dataset. For motor insurance fraud detection, please use the following dataset described in Phua et al. (2004) and available here.

o The data file is also available via the following link and can be loaded directly in Colab via the link using Pandas: https://gitlab.com/valdanchev/data-storage-for-teaching-ml/-/raw/main/automobile_insurance_claims.csv

3. Yelp reviews data for sentiment analysis available here: https://gitlab.com/valdanchev/data-storage-for teaching-ml/-/raw/main/yelp_reviews_data_500.csv

The data consist of 500 reviews from the website Yelp, each review is labelled 1 (positive review) or 0 (negative review).

Write-up

Write a clear and informative report. Imagine a very data savvy client who would read your report. The word limit of the report is 2000 words (+/- 10%), including text, computer code, and comments (The word count is approximate, and only provided for guidance; Colab notebook, as an interactive computational environment, can count the number of code/text cells and the number of lines but has no word count capabilities). The word count of the research report excludes references and data outputs (e.g., DataFrames, tables). Your report should summarize what was the goal of the project, what data you used, what machine learning techniques you have applied to your data and why. Summarise also what you learned from the project. Did you manage to solve a particular business problem using machine learning tools? If your method did not do as well as you expected, could you elaborate as to why? What conclusions can you draw from the project?

You should not worry if your results do not look 'great' (e.g., low accuracy score) or if your coefficients are not 'statistically significant'. Please report any results you found, including those that turned out to be 'useless'. This is a good research practice. The assessment is about you demonstrating your skills in performing machine learning analysis to address a business analytics problem; 'unexpected' results in the process are perfectly fine, and better be acknowledged.

You are encouraged to write and submit your report entirely in Colab (see below for details). Alternatively, you can write your report in Word or Pages or a similar application and provide your Colab notebook as a supplementary file. In either case, you will need to submit your computational notebook as it is a core component of your report.

Reproducible research report and good research practices

To create a reproducible research report, use throughout your computational (Colab) notebook:

• Python code in Code cells

• Hashtag symbol `#` in Code cells to introduce a comment line describing your Python code. Code commenting is an important part of the assignment.

• Markdown language in Text cells to write up your methods, results, and interpretation.

Before submission, please ‘Restart and run all’ (under ‘Runtime’ in the Colab menu bar) to make sure that your data analysis is computationally reproducible.

Feel free to reuse and reference code from the module notebooks as well as from the module textbooks and tutorials, providing that your project is sufficiently different from the sources you are drawing upon (for example, you may use the same data but define a slightly different business problem and apply a different model; or you may use the methodology and Python code from a source but on a new problem and data). You are also encouraged to consult and refer to online tutorials and sources, including DataCamp courses (e.g., Machine Learning with scikit-learn) and Stack Overflow. You can refer to such sources by providing links pointing to them or by in-text citation and end-of-report references.

Submission

You need to submit your notebook by 25th April 2024, 4 pm. You can download your Colab notebook from ‘File’ and then ‘Download .ipynb’, and submit your downloaded notebook. Alternatively, you can download the notebook as a PDF: from ‘File’ select ‘Print’ and then ‘Save as PDF’ (notebooks submitted as PDF should include in the front page a sharable link to the actual Colab notebook).

Schedule

• Decision about research problem (and variables you plan to use) due 25 March 2024 (to be indicated in this spreadsheet; you will be allowed to change your research problem and data afterwards if needed but it is important to think about and indicate your working ideas about your research problem by 25 March 2024).

• Research Report due 25th April 2024, 4 pm.

3. Assessment rubric with weighted criteria

4. Assurance of Learning measures: performance thresholds for assessment criteria

For UG: “significantly exceeds expectations” [outstanding/excellent] at equivalent of 70+, “exceeds expectations” [good] at equivalent of 60-69, “meets expectations” [average] at equivalent of 40-59; “does not meet expectations” [poor/outright fail] at equivalent of 39 or less.

Submission Information

Please observe the following style. guide. Unless otherwise specified,

• All work must be typed and submitted in MS Word or Adobe PDF format

• Font size should be 12 point (unless otherwise specified)

• Font style. should be Arial

• Lines should be double-spaced

• Leave margins for comment Insert page numbers

• Use a header containing your student ID number, the module code to which your work applies, and the date. And Please:

• Always spell-check and proof read your work before handing it in (once you have submitted your work you will not be permitted to retrieve it)

• Keep your own electronic back-up copy of your work and if possible save on two devices Avoid plagiarism

• Submit your work on time

Guidelines and Late-Work Policy

1. Coursework submitted late (and there are no extenuating circumstances) will incur a late penalty. Five per

2. cent of the total marks available shall be deducted for every period of 24 hours, or part thereof, that an assignment is overdue there shall be a deduction of five per cent of the total marks available (i.e. five marks for an assessment marked out of 100). After seven calendar days (168 hours or more late) the mark shall be reduced to zero, and recorded as 0FL (zero, fail, late).

3. Each module has word limits for coursework assignments; however, the decision about whether to impose a penalty mark for exceeding the word limit is made by each module organiser. You must check the module handbook and the assignment briefing documents to see whether the particular module organiser has adopted a penalty system. It is your responsibility to read the handbook and assignment briefing carefully. If no penalty is specified then the module organiser will take into account the word length under the standard marking conventions. For example, if you have exceeded the word limit then it might be that you have not been sufficiently succinct or focused in your assignment and therefore might be penalised for these weaknesses. Please note that word limits do NOT include references or appendices. However if excessive material is included on appendices then this too will be judged accordingly and you may be awarded a lower mark.

4. You should ensure that your submission is in either Microsoft Word or PDF format.

5. Failure to submit in either one of these formats will result in a mark of 0 being awarded for the particular assessment. It is therefore your responsibility to ensure that the file format is correct and it can be opened by the receiving party.

6. You should ensure that the correct piece of assignment is uploaded as the document downloaded on the due date by the module organiser will be marked regardless of content. You will not have another opportunity to submit the work again if you mistakenly uploaded the wrong document.

7. ALLOW YOURSELF PLENTY OF TIME TO SUBMIT YOUR COURSEWORK. DO NOT LEAVE IT UNTIL THE LAST MINUTE

8. Computer problems, such as computer viruses, failure to make a back-up copy or temporary internet access problems, will NOT be viewed as a valid reason for late submission.

9. Check that your assignment submission has been successful, and print a copy of the confirmation screen.

10. If you submit your assignment after the deadline, you will still be able to submit your coursework via QMplus however you will be penalised for late submission, the only exception to this is if you have an approved extension due to extenuating circumstances.

Submission Deadline

Date 25/04/2024

Time

QMplus Link/Page

Help and Support

• Module Organiser

• Disability and Dyslexia Service (DDS)

• Quantitative Skills Tutor

• Academic Writing Skills Tutor

Marking Rubric (A clear marking criteria should be provide breakdown of marks)

See table “Assessment rubric with weighted criteria” above.