Showing posts with label Similarity. Show all posts
Showing posts with label Similarity. Show all posts

Friday, February 8, 2013

A Follow on Similarity and Association Rules Heart Monitor Example

 The last two lectures have been focused on Association Rules and Similarity. During the class, there seemed to be some confusion relating the theory behind Association Rules and finding the frequent pairs. On Thursday (02/07/13), Dr. Megahed introduced the Chapter 3 topic of finding similar items. He gave a real world Industrial Engineering application of similarity with a Coordinate Measuring Machine (CMM) that measures a car door for defects. The CMM measures the car door at 11 distinct places to determine if the gap between the door and car was sufficient and within spec. He then broke down the differences in Similarity and Association Rules when inspecting the car door. The Association Rules included a 3D scanner to identify each doors unique association. With this blog I will attempt to give another real work example to hopefully clarify how Association Rules exist in the Industrial Engineering Community.
Throughout the years, people had to visit the doctor or special clinics with special equipment to monitor their heart rate, blood pressure, and other heart related metrics. These visits were usually scheduled and preplanned. Most visits returned no abnormalities and everything was ok, which simulates the similarity example.
With the routine visit to the doctor, one can only monitor if there is a fault or abnormality at that specific time. You can't verify your hearts activity since the last visit or until the next visit.

             With the advancement in smart phone technology, AliveCor has come out with an iPhone case that offers real-time EKG readings. The user must open up the app and place the case in their hands or on their chest for 30-second intervals. There are also continous monitoring capabilities out there from Everist Genomics. A wrist band that attaches electrodes to a patient's skin offers around-the-clock arrhythmia detection. The wrist band, with a built in Bluetooth device, collects the heart data, and the app on the smart phone analyzes the information and sends it to your physician for constant monitoring.

Association Rules
This continous monitoring is a good example of Association. With the smart phone, the doctor could detect any variation patterns and hopefully eliminate critical defects and abnormalities. After several months, the physician could use the Association Rules data to better predict when regular scheduled checkups should be planned.