Due Date: Thursday, Oct. 28 @ 11:59 PM
Submission Contact: ptaele [at] gmail [dot] com
Submitting Your Code:
For this programming project, it is assumed that you are using Eclipse as your programming IDE. To submit your code, first find your Java workspace directory and put it in an archive file (e.g., ZIP, RAR). Afterwards, e-mail me this archived file. If the file size for the attachment is too big, then upload it to a file host (e.g., Dropbox, Sendspace, RapidShare) to send me the URL.
IMPORTANT: Before submitting your code, you must set your output to the following directory:
- Output (Result File): C:\pp2\output
Submitting Your Results:
You are required to submit two different text files for your project:
- README File: Briefly explain your results.
- Results File: Your accuracy results output in a specific format (see below)
NOTE: Marty has provided a Java class to help you generate your confusion matrix:
About confusion matrices
Confusion matrices allow the accuracy rates to be visualized in a convenient matrix. The columns (x-axis) represent the predicted values, and the columns (y-axis) represent the actual values.
For an example, suppose you have a domain that has 4 shapes: A, B, C, and D. Then in the confusion matrix example below, Row 1 represents the predicted values, and Column A represents the actual values. Looking at the example below, Cell C3 shows that Shape B was predicted correctly 70% of the time, while Cell C5 shows that Shape B was predicted as Shape D 10% of the time.
A visualization of the confusion matrix with colors, where the darker colors represent higher accuracy, is given below. The goal of the confusion matrix is to make the diagonal values be as dark as possible.
Formatting your output
Using the example confusion matrix above, you are to output your data in the following format below.
Below is the same data represented above, but formatted to make it easier for you to read. Be sure to submit this output into the specified file path mentioned earlier.
