Project for Chapter 3
Choose either
Option A or Option B to submit through Class Moodle site on the Project 3
link.
Option A
Austin Peay State, a university near
Nashville, Tennessee, is applying a data-mining approach to higher education.
Before students register for classes, a robot looks at their profiles and
transcripts and recommends courses in which they are likely to be successful or
have higher chances of success. The software takes an approach similar to the
ones Netflix, eHarmony, and Amazon use to make their recommendations. It
compares a student’s transcripts with those of past students who had similar
grades and SAT scores. When a student logs in, the program offers 10 “Course
Suggestions for You.” This recommendation is based on the student’s major and
other information related to that student. The goal is to steer students toward
courses in which they will make better grades. According to Tristan Denley, a
former programmer turned math professor turned provost, students who follow the
recommendations do substantially better. In the fall of 2011, 45 percent of the
classes that students were taking had been on their top 10 recommendations
list. This data-mining concept is catching on. Three other Tennessee colleges
now use Denley’s software. Institutions outside the state are developing their
own versions of the idea.
Answer the following
questions in a two-page paper submitted through Class Moodle site on link Project
3.
1.
Which other companies are using
approaches similar to the one used by Austin Peay State?
2.
Based on which data does the
system make a course recommendation to a student?
3.
What are the benefits and
drawbacks of this approach to course recommendations?
4.
Are there any data that should or
should not be included in data mining for this purpose? Why or why not?
OR
Option B
With more than 200 million registered users, Pandora Radio is a
personalized Internet radio service that helps you find new music based on your
past and current favorites. (The service is also available to mobile devices as
an app.) The success of Pandora Radio’s business model derives from applying
data-mining tools to the Music Genome Project, which is a vast database of
songs that a team of experts has broken down into their various components:
melody, rhythm, vocals, lyrics, and so on. Listeners begin by entering their
favorite songs, artists, or genres, creating customized “stations.” Then, Pandora
Radio mines its database to find songs that are similar. Another data-mining
tool that Pandora uses is the like/dislike (thumbs up/thumbs down) option that
accompanies each song the site suggests. These responses are also factored into
which songs the Web site decides to play for the user.
Answer the following
questions in a two-page paper submitted through Class Moodle site on link Project
3.
1.
How does Pandora Radio recommend
music to its listeners?
2.
How are the listeners of Pandora
Radio able to create their customized stations?
3.
What are some variables that
Pandora Radio uses to recommend a song?