Course Description: This course will introduce students to the theory and practice of data mining for knowledge            discovery. This includes methods developed in the fields of statistics, large-scale data analytics, machine learning and   artificial intelligence for automatic or semi-automatic analysis of large quantities of data to extract previously unknown interesting patterns. Topics include understanding varieties of data, classification, association rule analysis, cluster         analysis, and anomaly detection. We will use software packages for data mining, explaining the underlying algorithms   and their use and limitations. The course include laboratory exercises, with data mining case studies using data from     biological sequences and networks, social networks, linguistics, ecology, geo-spatial applications, marketing and             psychology

Course Objectives:

In this course students will:

1.    Be introduced to the concepts of supervised and unsupervised learning and how they are used to solve various data-centric problems.

2.    Learn the fundamental algorithms of supervised and unsupervised learning (e.g. regression, decision trees, k-means, k nearest neighbors, anomaly detection methods, etc).

3.    Learn the critical preprocessing steps and validation methods that are required to ensure such methods are implemented properly.

4.    Understand why programming languages are critical for data mining.

Learning Outcomes:

By the end of the semester students will be able to:

1.    Define what data mining is as a field and be able to identify how it applies to various use cases.

2.    Implement supervised and unsupervised statistical and machine learning methods to answer questions about real world datasets.

3.   Articulate the pros and cons of the various methods, models and techniques learned in class and make decisions about which one is best for a given scenario.

4.    Describe the technical steps that proceed the application of a given algorithm and justify why they are needed.

Online Course Format:

This course will use several different mediums for learning.  Most weekly content will be given in the form of a lesson I have created, R scripts, or chapters from the assigned textbook. I will also post youtube videos that cover and recap     content, but these are not meant to be stand-alone lessons but rather supplements to the written content provided.   Being an online course, you will need to be more self-directed in both your learning and solving problems.

There will be two main coding exercises each week.  The first is found in the lessons I provide. You are expected to code   through, modify, play with the code in the lessons created to show you how these methods work and relate to the            overall goal. The second is the actual homework assignments.  These will be where you apply the ideas and skills learned in the chapter to a new data set to make inference.

R and Getting Help:

This course will use the programming language R, which all of you should have had at least one course in. The first week is a refresher that will check and make sure you have the skills needed for this class. It’s fine if you don’t ace it, but if you are really struggling then you should chat with me.

One of the key skills to develop when programming is figuring out how to solve your own problems.  This can be one of    two ways.  First, google … google is a great and normal way to solve problems, but it’s also easy to get bogged down        with the thousands of hits one various stack overflow pages. Second, if googling your problem doesn’t work, jump over   to our Slack channel (Check out setup tutorial on D2L).  Slack is the go-to communication software for tech companies as it's faster and more direct than emailing.  It also preserves an archive of communication.  So, if you can’t solve an issue,    go post an informative question (or answer!) to the Slack.

Regarding emailing me for help - essentially, don’t do it if it’s related to course content!  If you are struggling with an        issue you should post it to Slack!  Chances are someone else is struggling as well, or someone has already solved the        issue.  I want you all to work together to solve problems.  Now, if I see wrong answers or something not being answered, I’ll jump in (eventually), but I will be giving credit for Slack participation (both on answering and asking).  Where this gets messy is that although I highly encourage you to help each other, This does not mean you can directly share code            associated with an assignment (this is a violation of UA’s Code of Academic Integrity). So, Slack help should be in the    form of providing suggestions, similar code samples, or direction as to where to find help (i.e. a website or where in a      lesson we covered something). I’m not going to be too harsh on this as I want you all to participate, though. So if you’re  unsure if you are giving too much code, just post.

A Few Words on Technology and Communication:

1.   You will have access to and will be required to retrieve all course materials from the course page in D2L. Training for D2L can be found online at: http://help.d2l.arizona.edu/students  .

2.   You will need to have R and R Studio installed and functioning week 1. Here’s a tutorial.

3.   Again Slack participation is critical!  If you are having a coding issue, first try and solve it on your own.  If you’re still struggling, then post it to the class Slack channel.  Essentially, if you are about to email me with a

homework/class/coding question, post it to slack first. I’m not doing this to save me time, but rather because Slack is the communication tool for most tech companies, and learning how to use it is really important.  Also, answering

questions on Slack also contributes to your participation score. THIS IS A LARGE, REALLY EASY PART OF YOUR GRADE.  IF YOU HAVE ISSUES PARTICIPATING ON SLACK TELL ME ASAP.

Readings:

Weekly lessons will come via R Markdown HTML files.  This will be the foundational content along with teaching you a lot of how to code for the class and problems.

We'll supplement my content with the book Introduction to Statistical Learning by Gareth James, Daniela Witten, Trevor Hastie and Robert Tibshirani.  You can download a free PDF from the book’s website, although I highly recommend          buying a copy. It’s both easier to work from a paper copy and is a truly excellent book that will give you a solid                  foundation for machine learning. This will provide some more conceptual depth that I don’t have room for in my book.

If you want some extra R help you can check out the book “R for Data Science” by Hadley Wickham and Garret    Grolemund. This book covers how to create full data science pipelines in R (more than we’ll be doing here) and is available free here: https://r4ds.had.co.nz/  .

Complete List of Assignments with Grade Breakdown:

Grade Distribution:

90-100%  = A “exemplary, far beyond reqs/expectations”

80-89%     = B “exceeds requirements/expectations”

70-79%     = C “meets requirements/expectations”

60-69%     = D “falls short of requirements/expectations”

< 60%       = E “repeat of course needed”

Course Schedule: Here is the schedule for the class in terms of topics and due dates.   The class is divided into 10 main topics of learning with homework assignments due at the end of each topic.

1.   Wednesday, January 11th: Say hi on Slack. Update R ! Make sure you understand course structure.

2.    Monday, January 16: Reviewing R, R data structures, and an overview of data mining

a.    Homework due Saturday, January 21

3.    Monday, January 23: Diving deep into linear regression + Understanding through graphing

a.    Homework due Saturday, January 28

4.    Monday, January 30: Manipulating and preprocessing your data + Features vs. targets

a.    Homework due Saturday, February 4

5.    Monday, February 6: Classification via parametric and nonparametric models

a.    Homework due Saturday, February 11

6.   Tuesday, February 14: Midterm

7.    Wednesday, February 15: Cross validation and understanding types of error + For loops

a.    Homework due Saturday, February 18

8.    Monday, February 20: Trees and forests

a.    Homework due Saturday, February 25

9.    Monday, February 27: Unsupervised Learning - Clustering

a.    Homework due Wednesday, March 1

10. Friday, March  3: Final exam

Exam & Final Format:

This course has a single midterm exam and a final exam, both worth the same amount of points and use the same           format.  For the exams you will be given three hours to complete the online exam.  You will be able to take this exam       anytime within the day stated on the syllabus.  Once you start your exam you must finish, so be sure to block off enough time to complete it.  Both exams are open book/notes, but you can’t use google or any other comparable search          engine to find answers (doing so will result in a zero for the question at a minimum). The exams will include both          conceptual questions, applied questions, as well as general coding skills questions.  University policy on final                     examinations can be found here: 

https://www.registrar.arizona.edu/courses/final-examination-regulations-and-information

Student Accommodations: