You are required to undertake the following tasks:
1. Problem Identification
Download the dataset assigned to you from the module Blackboard site.
Read the data description file to learn some basic characteristics of the dataset. Make sure you have understood the nature of the data.
Perform simple data exploration to get to know: the total number of instances in the dataset, the number of attributes, the data type of each attribute, the basic statistics of each attribute (value range, skewness, and kurtosis), etc.
Identify and understand the business problems concerned with regard to the data.
Translate the business problem to a data mining problem, and identify the associated data mining tasks to be performed.
3. Data Preparation
Transform the dataset into the proper format to be used by SAS® in order to carry out the required data mining task.
Choose appropriate methods for data pre-processing, including dealing with missing values, tackling noisy data, conducting proper data transformation and normalisation, etc.
Divide the whole dataset into several subsets to be used for model training, test and validation.
4. Model Building
Perform the data mining task you have identified in the first task using the pre-processed dataset. Each task should be completed by applying at least two different algorithms. For classifier building, for example, you may choose decision trees and artificial network networks, or decision trees and nearest-neighbour based algorithm, etc.
In order to build the most appropriate and accurate models different combinations of the relevant model parameters should be considered for each of the selected algorithms.
5. Model Evaluation
Use the test and validation datasets created in the second task to evaluate the performance of the model produced from the data mining process. Compare the performance of different models in terms of accuracy, generalisation ability, simplicity and cost etc.
Discuss how the models created can be used to address the main business problems identified in the first task.
Final report
You final report should be well-formatted as a formal report containing Title page, Table of Contents, Abstract and References. The main content of the report must as a minimum include the following information:
A brief discussion on the methodology adopted for the data mining process.
A discussion on what pre-processing was carried out on the given dataset and why it should be conducted.
A discussion on each of the algorithms that were chosen and applied for the data mining task, and an explanation of the settings for the relevant nodes employed in SAS® Enterprise Miner.
A detailed analysis and sound interpretation of the models constructed, including the performance of each model, and their applicability to address the original business problems.
A reflective commentary and evaluation on the coursework.
Essential statistics, screen shots, and graphs.
The report should be submitted in a hard copy as well as an electronic copy.