Despite the drawbacks there is a lot to be said for

  • No School
  • AA 1
  • 13

This preview shows page 4 - 7 out of 13 pages.

Despite the drawbacks, there is a lot to be said for methods that don’t force us to make assumptions about the nature of the relationship between variables before we begin our inquiry. This is not a trivial issue. Most of us are trained to believe that theory must originate in the human mind based on prior theory and data is then gathered to demonstrate the validity of the theory. Machine Learning turns this process around. Given a large trove of data, the computer taunts us by saying “If only you knew what question to ask me, I would give you some very interesting a nswers based on the data!” The reality is that we often don’t know what question to ask. For example, consider a healthcare database of individuals who have been using the healthcare system for many years, where a group has been diagnosed with Type 2 diabetes and some subset of this group has developed complications. We would like to know whether there is any pattern to complications and whether the probability of complication can be predicted and therefore acted upon.
Image of page 4

Subscribe to view the full document.

What could the data from the healthcar e system look like? Essentially, it would consist of “transactions,” that is, points of contact over time of a patient with the healthcare system. The system records service rendered by a healthcare provider or medication dispensed on a particular date. Notes and observations could be part of such a record. Figure 2 shows what the raw data would look like for 5 individuals, where the data are separated into a “clean period” which captures history prior to diagnosis, the red bar which represents the “diagnosis” and the “outcome period” which consists of costs and other outcomes such as the occurrence of complications. Each colored bar in the clean period represents a medication, showing that the first individual was on three different medications prior to diagnosis, the second individual was on two, and the last three were on a single medication. The last two individuals were the costliest to treat and had complications represented by the downward pointing red arrows whereas the first three individuals had no complications. Figure 2: Healthcare Use Database Snippet It is non-trivial to extract the interesting patterns from a large temporal database of the type above. For starters, the raw data across individuals typically needs to be aggregated into some sort of canonical form before useful patterns can be discovered. For example, suppose we count the number of prescriptions an individual is on at every point in time without regard to the specifics of each prescription as one approximation of the “health” of an individual prior to diagnosis. Another identifier might be the specific medications involved, where green and blue might be “severe” medications.
Image of page 5
From the above data, a “complications database” might be synthesized from the raw data. This might include demographic information such as a patient’s age and their medical history including a list of current medications aggregated into a count in which case we get a summary table of the type below. A
Image of page 6

Subscribe to view the full document.

Image of page 7

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern

Ask Expert Tutors You can ask 0 bonus questions You can ask 0 questions (0 expire soon) You can ask 0 questions (will expire )
Answers in as fast as 15 minutes