Need data to classify fitness
- Dear All:My name is M. Fatih Akay, and I am an Assistant Prof. at Electrical and
Electronics Engineering Dept. of Cukurova University , Turkey . My
colleague and I are currently working on classifying whether or not a
person (or an athlete) is aerobically fit by using a statistical
learning method. Our project requires a data set of people (or athletes)
who have taken aerobic fitness test on a treadmill or a cycle ergometer. The
data set we need can contain parameters of each person (or an
athlete), such as: age, height, weight, test time, maximal oxygen
uptake, velocity at maximal oxygen uptake, heart rate at maximal oxygen
uptake, lactate threshold, anaerobic threshold, the time constant for
oxygen uptake kinetics, work efficiency and any other parameters
required for aerobic fitness assessment.
Please let me know if you can provide us with such a data set. Also, if
you would like, we can collaborate on this work, which may yield
preparation of a project proposal and/or publications.
I look forward to hearing your ideas/comments.
M. Fatih Akay
Yahoo! oneSearch: Finally, mobile search that gives answers, not web links.
- For those interested in this topic, I rejected Fatih's message initially and contacted him, as below. His reply and my reply follow.
My first message to Fatih:
You haven't quite got the right idea, Fatih. If you are going to
predict fitness, you need a criterion measure of fitness. You can't
just say the person is fit. You need something to base that on.
VO2max and so on are measures that tell you. One thing you have left
out is recent training, which you can also use to define fitness.
But cross-sectional relationships between training and whatever
measures of performance are not interesting any more. We did all that
stuff years ago.
If you still want to post your message, send it again and I will add
the above as a note.
Thank you for your kind response. However, I am afraid I was not able to follow your comments. We would like to predict fitness based on some criteria such as age, height, weight, test time, maximal oxygen uptake, velocity at maximal oxygen uptake, heart rate at maximal oxygen uptake, anaerobic threshold, and any other parameters required for aerobic fitness prediction. We will not just say the person is fit. Our prediction will be based on the parameters I mentioned.
We will apply a new method called support vector machines, for fitness prediction and evaluate the performance of the method.
What is wrong with this approach?
By the way, I am not interested in defining fitness through recent training.
Let's clarify whether or not there is a problem in our approach.
OK, you appear to be pursuing a kind of factor analytic approach. Post the message, I will add a comment, and we'll see what others say.
Will G Hopkins, PhD FACSM
Institute of Sport and Recreation Research
Private Bag 92006, Auckland 0627, New Zealand
Work +64 9 921 9793, Fax +64 9 921 9960
Home +64 9 376 0198, Cell +64 27 427 2518
Be creative: break rules