Here's just a small sampling of the concepts we're tackling.
What are the real factors that measure risk in cardiovascular heart disease? Was it the base categories identified in the Framingham Heart Study? Or is there better measurements of risk from CHD as observed in unstructured data characteristic like stress, social assistance & care dependency, racial/demographics, and economics that were missed in the base population of Framingham, MA?
“There has been 2,400 studies written using the Framingham data. During the years, many other risk factors were evaluated. Obesity, exercise, psychological, and social issues. In fact, the Texas Heart Institute Journal named the Framingham Heart Study as the top 10 cardiology advance of the 20th century.”
For the last year I've been working on a Big Data & Advanced Analytics Cloud based solution that would help healthcare specialists measure, understand patient care, & predict the risk of readmissions due to Cardiovascular Heart Disease using Big Data & Advanced Analytics, and now we've expanded our solution to include IBM Watson.
Who's to blame? Was there an epic mistake in play calls or a great defensive move made by a single individual that changed the game. The answer is... none of the above, because decisions were made based on emotion not facts.
Super Bowl XLIX - The play that changed the game
While its easy to sit in our arm chairs and make the call, we all tend to agree, that was probably biggest bonehead play call of Super Bowl history. What was the coach thinking when he opted to pass rather than use a run play from one yard out with 24 seconds left on the clock? What was the defensive position that ultimately changed the game? But from our perspective, would Watson have made the same decision?
The one image every woman dreads to find... a breast cancer cell magnified.
Detecting Change in medical images
The process takes time, for which every moment seems to slip away while we wait for results. The anxiety, the emotions, the fear is unbelievable for any one that has gone through a breast cancer patient's journey. Disease awareness if fraught with misguided information, yet we seek out answers on the internet, from Uncle Dave (he's a med tech after all), the magazines in the doctor's office with headlines promising miracle cures and weight loss without diet or exercise. Then there's the next phase, Diagnosis. For those that have been told their lives now have a finite amount of time the information is crushing, but often what we're given is a series of professional opinions based upon someone's dedicated time to the medical institution. The undeniable fact is - it depends; and that's what we're working on - gathering, analyzing, and providing the best facts, so that no one, no matter where they live, has to wait in a fluorescent lit sterile room wondering, was that the best opinion or should I seek another?
Would you Draft this man
In 2010, the Denver Broncos drafted as their number one pick, Timothy (Tim) Richard Tebow from the University of Florida. As a member of the Denver Broncos, he started the last three games of his rookie season and became the team's full-time starting quarterback beginning in the sixth game of 2011. The Broncos were 1–4 before he became the starter, but began winning with him on the field, often coming from behind late in the fourth quarter, until they won their first AFC West title and first playoff game since 2005, defeating the Pittsburgh Steelers in overtime. Tebow was traded to the New York Jets during the offseason after the Broncos acquired free agent quarterback Peyton Manning. Tebow received little playing time for the Jets and on April 29, 2013, the Jets released Tebow. So, what was Tim's value to the Denver Broncos as a number one draft pick? Should they have gone a different direction? What would you have done? What would advanced image analytics have said - "This quarterback has a slower than normal release during a pass play, requiring more than 4 seconds to setup a play, thus play must be optimized to allow movement outside of the box."
Predicting
Every time you text on your smart phone, an application in the background attempts to "Predict" your next word. The application uses n-gram Markov model for prediction. An n-gram model is a type of probabilistic language model used for predicting the next item in a sequence, basically it works on a term in the form of a (n − 1)–order Markov model. An n-gram of size 1 is referred to as a "unigram"; size 2 is a "bigram" (or, less commonly, a "digram"); size 3 is a "trigram". To try out my Predict the Next Word application click on "Predict."