Data Mining Solutions to Reduce Direct Marketing Costs: A Decision Tree Analysis

1. Analyse a Data Mining technique capable of supporting practitioners to make reliable decisions which require predictive modelling, for example, in a Business scenario

2. Demonstrate results of using an efficient technique which is capable of finding a solution to a given predictive problem represented by a data set

3. Evaluate the accuracy of the technique in terms of differences between the predicted values and the given data.

What am I required to do in this assignment?

Task

Students will develop a Data Mining (DM) solution for saving the cost of a direct marketing campaign by reducing false positive (wasted call) and false negative (missed customer) decisions. Working on this assignment, students can consider the following scenario. A Bank has decided to save the cost of a direct marketing campaign based on phone calls offering a product to a client. A cost efficient solution is expected to support the campaign with predictions for a given client profile whether the client accepts or rejects the offer.

A company wants to develop an innovative DM technology which will be competitive on the market. The team will analyse the existing technologies to design a DM solution winning the competition. A team Manager will choose the best solution for the market competition in terms of cost efficiency. The evaluation of the developed solutions will be made on the test data. The costs will be defined for both the false positive and false negative predictions. Students will apply for one of the roles (i) group manager, (ii) group member or (iii) work individually. The group manager will arrange comparison and ranking of solutions designed in a group. Each student will run individual experiments to find an efficient solution and describe the obtained results.

Assignment Options

Students are welcome to explore possible assignment options within the unit scope.

Method and Technology

Examples of cost-efficient technologies for direct marketing are provided on the UCI

Machine Learning repository describing a Bank Marketing problem.

To design a solution, students will use Data Mining techniques such as Decision Tree.

Examples of solutions will be provided within an R Cloud technology CoCalc available

online. An alternative is RStudio which can be used freely by students.

Project Data and Script

Download the required Bank marketing data (.csv) and R scripts (.sagews).

Individual Report

Each solution will be evaluated in terms of the prediction accuracy (costs of false decisions). All submissions are made via BREO. A template can be used for reports.

Is there a size limit?

What do I need to do to pass? (Threshold Requirements)

1. Create a Cocalc project account (5%)

2. Upload the project data and script to the CoCalc account (5%)

3. Using CoCalc, run the project script to build a Decision Tree on the data (10%)

4. Analyse and describe the Decision Tree and the project script outcomes (22%).

5. Total to pass 42% How do I produce high quality work that merits a good grade?

6. Identify a set of parameters which are required to be adjusted within Decision Tree techniques in order to optimise the solution in terms of prediction accuracy

7. Explain how the parameters of Decision Tree technique influence the prediction accuracy

8. Run experiments in order to verify the solution on the given data set

9. Analyse and compare the results of the experiments in a group and with the known from the literature

10. For A-grades (>72%), students will make a 5-min demonstration of the developed artefact

How does assignment relate to what we are doing in scheduled sessions?

Data Mining, Decision Trees and use cases developed in R Cloud will be considered on lectures and tutorials.

How will my assignment be marked?

Your assignment be marked according to the threshold expectations and the criteria on the following page.

You can use them to evaluate your own work and estimate your grade before you submit.