On Tuesday, 7/24, we had our second round of data collection. This round went better as there were fewer glitches with the eye trackers. However, the Microsoft Kinect virtual pointer caused a considerable lag in the video feed needed to perform the surgery. We got through 3 participants as opposed to 2 during the same time period however.

On Thursday, I met with Dr. Kleinsmith and using the machine learning application Weka, we ran K- means clustering (k=2 and k=3) on the 15 distinct pairs with the 5 synchrony features. K=2 showed the clearest distinction between the EDA of the paramedic trainee pairs. Once we identified the most discriminative EDA features for clustering (Euclidean distance of the 3 difference measures and Pearson correlation), I began writing up my final poster and developing figures, graphs and charts of my results, since the poster is due next week.