Machine Learning & AI

Propose and develop a Machine Learning based solution for your workplace or for a problem domain agreed in advance with your Tutor.
The problem domain is finance but I haven’t chosen a detailed topic yet and my tutor doesn’t mind as long as there’s evidence of understanding of Machine Learning and business context.

My initial thoughts are for a solution:
”
Financial Services product recommendation. (Pensions firms typically don’t. Financial Advisers step in at this point. Commercial banking loans/mortgage etc. Would be more appropriate)

Lots of financial companies currently have online messaging support baked into their website offering answering basic questions to their users before connecting them to a person.

Machine Learning could be used to identify and recommend products to consumers based on information present about them, and asking them additional questions through chat.

Same recommendations could be used to suggest additional products to customers and send them an email.
”

Your submission at a minimum should:

• Include an assessment of the technological and business need driving the solution.
• Critically assess the relevant ML technologies involved compared to those not selected.
• Apply and evaluate your chosen ML technique to the problem.
• Discuss the challenges of applying ML practices for full benefits realization in your organisation.

Some potential sources of data sets that could be used for your solution include:

• https://archive.ics.uci.edu/ml/datasets.php
• https://www.kaggle.com/datasets
• https://ai.google/tools/datasets/

Assessment Criteria

The following high-level assessment criteria should provide a useful guide to the general characteristics you should seek to develop in your assessment.

• Identification of a business problem and an available dataset that has the potential to be amenable to the application of ML within a relevant context (10 marks)
• Application of one or more relevant machine learning algorithm(s) to the dataset, and discussion of the learned models and how the performance of these algorithms can be adjusted (30 marks)
• Evaluation of the performance of the machine learning model(s) that were produced, and critical discussion of how their results could be improved upon (30 marks)
• Discussion about whether (and if so, how) or not (and if not, why) your solution could (not) be applied within the relevant context previously identified, and what benefits and risks may have been incurred by doing so (20 marks)
• Use of relevant academic and commercial literature, and overall professionalism of the report (10 marks)