Satisfaction means fulfillment with what one has, and the bottom-up structural approach states that life satisfaction is an assessment of one's life as a whole, and not an assessment of the current state in which an individual finds themselves. This means a general assessment of life satisfaction using separate but interconnected units that contribute to individual life satisfaction in different ways. The aim of this paper is to investigate the relative contribution of individual indicators in explaining life satisfaction. Five Vanderweele units were used to operationalize the bottom-up approach: religiosity, family, work, health and community. Each unit contains the corresponding variables from the tenth wave of the European Social Survey, which was conducted from September 2020 to August 2022 through face-to-face interviews in 22 participating countries. For the purposes of this work, the countries were divided according to the EUROVOC classification into the geographical areas of Western, Central-Eastern, Northern and Southern Europe, and for each of the mentioned geographical units, a Random Forest ensemble model was created, which showed the classification errors for each of the five classes of life satisfaction and estimated the classification error on data sets outside the training sets. The basic parameters of each of the four models were graphically presented and explained. The results show that it is possible to observe a certain pattern: the most significant variables for each of the classification models belong to the domains of health and work, followed by variables from the domains of religiosity and community. Variables belonging to the domain of family are not significant for any of the models.