1) A Data Mining project report should be closer to a scientific or engineer document than a business document. It should be objective and supported by quantitative results. In many cases a DM project report will be a portion of or an appendix to a broader business report (e.g. Marketing Plan, Program Roll-out Plan, Fraud Analysis and Prevention Report). For this reason, the report should contain the following sections:
Problem statement (MUST INCLUDE SUCCESS CRITERIA)
· Background
· Method and Approach
· Project Plan (your approach to solving the problem)
· Solution (alternatives and chosen solution)
· Experimental design (method of testing/comparing models)
· Results
· Discussion
· Conclusion
The closer you are to the prescribed layout, the easier it will be to mark … thus the better your chances of a good mark.
The Project Plan subsection should tell me your strategy for tackling the KDD project so as to solve your problem and NOT a chronological narrative that mixes a history of your actions and results.
The Solution subsection needs to clearly describe the final solution and possibly alternative solutions you may have considered.
The Experimental Design subsection needs to clearly indicate your method of evaluating and comparing the predictive models.
The Results section presents intermediate and final numbers and graphs based on the previously described approach and experimental design
The Discussion section discusses the results in the context of the problem. How well did you meet the success criteria? Discuss things that went well and not so well.
And the Conclusion is a summary of the report, final conclusions on success of project, recommendations and final remarks (such as what would you do next time).