engleski

Discovering the Pattern of Life Expectancy not Achieving (LENA)

Life expectancy as a measure of longevity was estimated within those died due to IHD or sudden death as a cause of death in a 20-year follow-up period, using data mining strategy. The measure of longevity means (1) if a person achieves his/her life expectancy then he/she has reached the goal – longevity, he/she has lived long, (2) if a person does not achieve his/her life expectancy then he/she has not reached the goal – he/she has died before. The methods included were the following: (1) segmentation of the sample, (2) descriptive statistics of the segments, (3) finding predictor-candidates for LEnA by applying decision tree generating See5 algorithm, and (4) detection and description of the relationships between predictor-candidates. Ranking of predictors for earlier deaths (than expected) was shown that diastolic blood pressure appeared at the first place, followed by age, physical activity and animal fat consumption in the middle, and smoking and BMI at the end of the scale. The rules and obtained patterns in this study showed full epidemiological meaning. Diastolic blood pressure showed the greatest differences between LEA and LEnA. The age attribute showed that there obviously were some latent variables linked to age since the younger age group did not achieve the LE. Physical activity, high animal fat consumption, intensive smoking and higher body mass index seemed to be very trustworthy and suited well with the epidemiological experience.

life expectancy; pattern; knowledge discovery

nije evidentirano