Reference no: EM132385491
ITC573 Data and Knowledge Engineering Assignment - Decision Support Systems and Privacy Issues, Charles Sturt University, Australia
Task -
Australian Magazine Subscriptions (AMS) is a medium-sized company that manages the subscription services of a variety of Australian periodical publications. The company maintains an online website where customers can subscribe to magazines. They also obtain subscriber lists from their client magazines and manage the sending-out of published works to these subscribers.
Recently, AMS has been noticing a steep decline in new subscriptions and a considerable number of clients cancelling their subscriptions. AMS stores a variety of customer data which they feel may be useful in discovering how and why people decide to cancel their subscriptions. This data includes:
- Customer Age.
- Customer Address.
- Customer Gender.
- Customer Interests (inferred from the magazines they subscribe to).
- Date when subscription was initiated.
- Number of subscriptions.
- Length of subscriptions.
- Whether the customer has had a subscription that has lapsed in the past (True/False).
- Whether subscriptions were made online with AMS or through the individual magazine.
You have been hired as a consultant and charged with the design of a decision support system (DSS) which meets the following objectives:
1. The DSS should allow AMS to identify which customers are at risk of cancellation so that they can be targeted with marketing and special offers to attempt to dissuade them from cancelling.
2. The DSS should allow AMS to identify the individual patterns of cancellation for a particular customer.
3. The DSS should be able to identify distinct groups of customers and provide insights into their subscription habits.
4. The DSS should display changes in subscription habits of customer groups over time.
5. The DSS needs to be able to provide all of this information in an attractive way, which is understandable to marketing department at AMS.
TASK 1: Discuss the dataset and challenges in designing such a DSS as you understand it, particularly focusing on issues as follows.
- What privacy concerns need to be accounted for in the development of this system?
- How will they be accounted for?
- What privacy concerns could arise had AMS decided to share their customer data with their client magazines?
- How could these concerns be accounted for?
- For the DSS you need to design, what data pre-processing techniques might you need to use and why?
- What kinds of knowledge do you feel you may be able to extract from this data?
- What additional data do you think would be useful for this decision support system, that AMS would be able to collect?
TASK 2: Based on your understanding of the problem, identify the data mining algorithms you will need to use to achieve these business goals. Justify each of your answers with reference to the objectives outlined above. Discuss any transformations you intend to perform on the data, and the way the data will be used in the context of these algorithms.
TASK 3: Design the user interface and reporting components of the Decision Support System, with the knowledge that the end-users of the system will not be data scientists. Present these designs in the form of interface mock-ups. These may be prepared in any graphics editing software, but must be clear, precise, and understandable. Discuss any automation elements within the system in conjunction with your diagrams. Justify all decisions made in the design in reference to the outlined objectives and the scenario described above. Designs should incorporate the discussed data mining algorithms from Task 2.
TASK 4: This task is not related to the DSS for AMS.
Use the bank-full.arff dataset that you used in Assignment 2 to evaluate the usefulness of data cleansing and pre-processing in discovering useful knowledge for critical business decisions.
In this task first create a version of the bank-full.arff dataset with various attribute values being missing.
Apply a missing value imputation technique to impute the missing values. Demonstrate the impact of the missing value imputation on data mining activities such as classification and prediction by comparing the data mining results from the original dataset, dataset with missing values and imputed dataset.
Learning Outcomes - This assessment task will assess the following learning outcome/s:
- be able to compare and evaluate various knowledge discovery techniques.
- be able to identify and design approaches for knowledge discovery from data for making critical business decision.
- be able to compare and critique various data pre-processing techniques.
- be able to evaluate the usefulness of data cleansing and pre-processing in discovering useful knowledge necessary for critical business decision.
- be able to critically analyse privacy preservation in data mining, data publishing and data sharing.
Attachment:- Data and Knowledge Engineering Assignment File.rar