event_entropy_of_lengths(EVENT_DATA) This function takes a sequence of eye fixations (cevent) and treats each instance in a cevent as a distinct category. In this way, the whole cevent is viewed as a distribution and event lengths are viewed a probability. The function then calculates the entropy of the whole cevent. The idea is to capture the dynamics of cevent, both the number of instances and the lengths of those instances.

0001 function e = event_entropy_of_lengths(data) 0002 % event_entropy_of_lengths(EVENT_DATA) 0003 % 0004 % This function takes a sequence of eye fixations (cevent) and treats each 0005 % instance in a cevent as a distinct category. In this way, the whole cevent is 0006 % viewed as a distribution and event lengths are viewed a probability. The 0007 % function then calculates the entropy of the whole cevent. The idea is to 0008 % capture the dynamics of cevent, both the number of instances and the lengths 0009 % of those instances. 0010 0011 0012 p = data(:,2) - data(:,1); 0013 p = p ./sum(p); 0014 e = -sum(log2(p) .* p); 0015

Generated on Wed 24-May-2017 00:00:56 by