SMALL DATA AS A PESONALIZED SOLUTION FOR FORMATION OF THE STRATEGY OF THE PERSON’S WEIGHT CORRECTION

Abstract

The article shows the handling with so-called universal data and problems during this, but which isn`t universal in real due to some hidden patterns between some attributes. Even when algoritmization and classification of tasks for only several big branches are possible and Big data exists – even now it’s not the best idea to always use universal methods for dealing with all problems. Such behavior can look strange a little but can be caused by coupling between some of attributes in the multiattributes samples of a data. And even more – context exists as well but big data doesn`t always very good to deal with it. So the aim of the article is to find some problem which shows this principle brightly and one of them is the task of normalizing the human’s weight. This problem is insidious due to causing not only the different effect during the same actions in quite similar people, but even between the different version of one people in different time periods. For using the conception of small data for so applied tasks, it will be useful to use some methods like multicriteria analysis, f.e. TOPSIS to deal with the most close to ideal solution while choosing between many criterias simultaneously. Different methods for ierarchy analysis as well so we can compare the alternatives (it’s about different strategies of weight’s correction for the same person). Also the methods of correlation analysis will be useful due to allowing to establish connection between different attributes from the multiattributes data samples (f.e. between type of sport activity, amount of possible kkal for burn, amount of desire to eat after that, and motivation to exercise with the help of this sport for concrete person in general). Scientific innovation here is in researchement of situation with the help of real empiric data from real world and finding out the impact on the person.