You need to have JavaScript enabled in order to access this site.

This assignment does not count toward the final grade.

PA4: kNN and AUC

Due No Due Date
Points 1
Submitting a file upload
File Types zip

See this document Links to an external site. for the PA instructions. Upload your code files as a zip.

Rubric

Title:

Keep in mind that 7 students have already been assessed using this rubric. Changing it will affect their evaluations.

PA4: kNN & AUC Rubric

Criteria

Ratings

Edit criterion description

The program compiles and/or runs without error

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The source code is well organized and commented

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

Each source code file includes a header indicating who wrote or contributed to the code

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

Code from any external sources, or code that relied on outside influence, includes a citation in a comment just above the corresponding code

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The code is built off of the codebase we built in class

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

There is an implemented, working kNN program

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The kNN program supports three distance measures: Euclidean, Manhattan, and Cosine

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The kNN program requires that the user specify the positive class, e.g., through a command line argument

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

In the output of the kNN program , the "predicted_label" column form our class example is now a "predicted_positive" column and includes a number between 0 and 1 indicating the confidence that an observation belongs to the positive class

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The kNN program has an accurate "help" message if the wrong number of required command line arguments are provided

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

There is an implemented, working AUC evaluation program

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The AUC evaluation program takes as input the output from the kNN program

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The AUC evaluation program requires that the user provide the positive class, e.g., through a command line argument

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The AUC evaluation program outputs the AUC in a nice format and includes the label of the positive class

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The AUC evaluation program has an accurate "help" message if the wrong number of required command line arguments are provided

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The submission includes a README

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The README describes how to compile (if Java) and run your program

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The README describes what command line arguments, if any, are available

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The README includes a section that states who worked on the project

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing

Edit criterion description

The README includes a section that describes the AUC evaluation of running kNN over the SMS spam training and development data sets with k=10 and spam as the positive class using each of Euclidean, Manhattan, and Cosine distance

Edit rating Delete rating

Passing

Edit rating Delete rating

Not passing