I have been reviewing the data I have been gathering on myself for the BascBookReader and it turns out (no big surprise here) that how long a person spends reading a page directly reflects how interested they are. The following is a composite of the statistics I currently am gathering. This includes iris tracking, blood presser and heart rate, personal rating, time spent per page, activity spent per page (i.e. moving the mouse).
- Red represent long term interest, which is an average of interest over an extended timeframe.
- Blue represents short term interest, which is an averaging of interest over a shorter timeframe.
- Orange represents instantaneous instants in that instant.
As you can see there was a dip in the interest at around 135 seconds. This was when my girlfriend walked into the room and asked me a question. It is clear that this is somewhat accurate in determining interest or at least focus over time. If you then look specifically at how long I spent per page you would see the following:
This data seems to corroborate the previous interest graph. As I became more interested I would spend more on pages. The series of pages (20 – 24) where I the reading time was low, this was because those pages were rather simple and had a fair amount of white space. During this same time period I was using an algorithm to average and prune the data:
This graph required one thousandth of the space of the full EEG data, yet it seems to be accurate.
Further, I have compiled several different sessions over a ten minute period:
On this graph you can see exactly which times I exited and restarted the program, it clearly had an affect on the program (due to the averaging of the time frame). It would be wise to zero this value out to ensure the data is accurate. This data does not show you that much because of the stop-start nature of the test I was running today. However, I intend to gather 5 – 10 subjects tomorrow to paint a more accurate depiction of the data.
It does seem as though the largest indicator of whether a person enjoys a book is how much time they spend on a page. This would need to be determined for each individual, but I feel would be accurate enough to determine focus.
mfw
This seems counter intuitive. Subjectively, when I’m really interested in something, I can fly through it, while I’ll keep getting distracted from a boring page and it’ll take forever to get through.
Perhaps, however when I am bored, i’ll just go to “speed reading mode.” Essentially, I just gather facts and move on. I don’t try to visualize the information and add it to my mental palace so to speak.
I think it’s subjective, which is why other factors to determine interest are VERY useful. It helps with individual classification.
Amazing…
How did you track your iris movements?
I work in a research lab which has me working with computer vision. I figured out how to track circles, then I just calculate the dimensions of the face that gives me a “guesstimate” on where their eyes are on the page.