Data Mining and Time-Series Analysis as Two Complementary Approaches to Study Body Temperature in Obesity

Abstract

Obesity is becoming a pandemic worldwide but the mechanisms that cause obesity are not well understood. One possibility are metabolic differences between lean and obese people, for which body temperature may offer a proxy which is relatively easy to measure. In the present contribution, we present results from two complementary methodological approaches to measure skin temperature as a function of body weight: in the first study temperature at the axilla and anthropometric measures were collected at a single time point in 1,073 male and female employees of all ages of the Universidad Nacional Autónoma de México (UNAM), whereas in the second study a 1-week continuous monitoring was realized of the skin temperature of the non-dominant wrist of 22 male young adults. In spite of the methodological differences, both studies indicate a higher mean temperature of the obese with respect to the lean subjects, possibly reflecting how obese people offset excess calorie intake by a higher heat transfer to the environment. On the other hand, with respect to the variance of the temperature over groups of underweight, normal weight, overweight and obese subjects, the first study that was realized in controlled circumstances did not detect any differences between groups, whereas the differences that were detected in the second study probably indicate behavioural differences between groups such as the level of physical activity.