machine learning

1

CSCI 4723*50 MACHINE LEARNING

Course Project

Group Project

Group size: 1-2 students. (1 means working individually.)

I. Project Description

The goal of the project is to identify and address a domain-specific problem on data analysis by applying machine learning algorithms or models. The topics can be on prediction, regression, or classification.

In this project, you will need to locate a specific dataset, define the problem(s) that you want to study, and implement at least one machine-learning algorithm to solve the problem. Data preprocessing should be conducted if needed. You are recommended to use Python and apply Scikit learn library. Repeating an analysis that was performed in another class is not allowed.

The Evaluation will be based on completeness and quality of problem definition, methodology introduction, implementation and presentation (project report writing).

II. Submission Project Report and Source Codes should be submitted by the due date. The source codes should be in a separate file or zipped folder. Only one group member should submit on behalf of the whole group. The group members will get the same grade unless the work distribution is extremely unbalanced.

The project report should include but not limited to the following components.

• Title and author(s). The title should capture your project question and main methodology.

2

• Abstract. Briefly summarize the questions that the project works on, the methods and the results.

• Background/Introduction. Introduce the background, understanding and analysis of the problem. Explain the motivation or the context.

• Problem Statement. Define the problem(s) that the project works on in detail.

• Methodology. Describe the organization and patterns of the data set. Describe methodology or algorithms using pseudocode or flow chart. Describe the experimental settings (e.g., size of training data and test data), data preprocessing (if any), and evaluation metrics of methods.

• Experimental Results. Depict experimental results using text, figures or tables. Patterns of the data should be described. Appropriate visualizations of data should be created to illustrate the experimental results. Interpret the findings from the experiment.

• Conclusions. Conclude your course project. Briefly describe what you have gained in this project.

The expected length of the project report can be 6 to 8 pages (single column, using the template provided by the instructor).

III. Resources Hints on datasets: Here are some possible data sources, but many more exist:

o UC Irvine Machine Learning Repository http://archive.ics.uci.edu/ml/ o The Home of Data Science & Machine Learning

https://www.kaggle.com/ o U.S. open government data. http://www.data.gov/ o Kaggle https://www.kaggle.com/

The above UC Irvine Machine Learning Repository and Kaggle website provide many featured data sets that are suitable for machine learning tasks such as classification and clustering. Some datasets require considerable knowledge to interpret, while others are easier to understand. In government data website, you can search by keyword say, ‘Education’ and then download a specific dataset. Hints on Problem Definition and Methodology: Course Modules in Canvas, Textbook.