Define a problem on the dataset

Reference no: EM132352626

Question: 1. Choose data mining problem and data set:

For this project, you must choose your own dataset. It can be one found from an on-line source, one of your own, or one of the ones from the UCI repository. A list of additional dataset sources is provided at the end of this document. If you would like to use data collection API to curate data, the attached document provides an example of using R to collect Twitter data.

Some rules/tips about choosing data sets:

a. Do not choose the datasets that we have already analyzed in class.

b. It should not be a small or made-up dataset. For this semester, "small" is defined as fewer than 1000 examples in the dataset.

c. Choose a data set that does not require excessive data preprocessing.

2. Experiment design:

Define a problem on the dataset and describe it in terms of its real-world organizational or business application. The complexity level of the problem should be at least comparable to one homework assignment. The problem may use at least TWO different types of data mining algorithms that we have studied this semester such as Classification, Clustering and Association Rules, in an investigation of the analytics solution to the problem.

This investigation must include some aspects of experimental comparison: depending on the problem, you may choose to experiment with different types of algorithms, e.g. different types of classifiers, and some experiments with tuning parameters of the algorithms. Alternatively, if your problem is suitable, you may use multiple algorithms (Clustering + Classification, etc.). If there are a larger number of attributes, you can try some type of feature selection to reduce the number of attributes. You may use summary statistics and visualization techniques to help you explain your findings.

3. Final project paper:

To complete this project, write a final report that conforms to general research paper format. See (Pang, Lee, and Vaithyanathan, 2002) as an example. Your report should be within 6 pages, 1 inch margin on all sides, and at least 12 point Arial or Times New Roman. Remember that your project paper serves as the tour guide for your readers to be able to repeat your data mining process and discover the same patterns as you did. It is very important to cite and paraphrase relevant work appropriately.

Reference no: EM132352626

Questions Cloud

Describe signs that may indicate an individual is distressed : L/601/8143-Support Individuals Who are Distressed-Explain how working with an individual who is distressed may impact on own well being.

Evaluate different types of research methods : Evaluate different types of research methods. Compare and contrast the differences between qualitative and quantitative research methodologies.

Describe how current legislation relates to assisting : K/502/7583-Understanding and Enabling Assisting and Moving Individuals- Describe how current legislation relates to assisting and moving individuals.

Determine how the principle impacts data security : Read five articles and discuss principle of least privilege(POLP promotes minimal user profile privileges on databases based on the idea that limiting user's.

Define a problem on the dataset : Define a problem on the dataset and describe it in terms of its real-world organizational or business application. The complexity level of the problem should be

How you plan on correcting the errors : Throughout the term, feedback on your milestones has been provided. What is important about feedback is not only do you need to read the feedback.

Examine risk management approaches the organization consider : Identify a SMB of your choice. In a 500-word paper, examine risk management approaches the organization may consider deploying for mobile device management and.

Distinguishing set of characteristics : Label each cell with the name, distinguishing set of characteristics, either total magnification or approximate size (in um), and its function.

Plot the curve for risk and expected returns : Case Motivation - PARTNERS HEALTHCARE - Plot the curve for risk and expected returns of the optimal portfolio combinations involving the 4 asset classes

User Account

All Pages