System modeling and data mining techniques have considerable potential for use in public health informatics. In this paper, we propose a methodology of constructing a parsimonious Takagi-Sugeno (TS) rule model from data in predicting the impacts of child deprivation on education according to the ones of deprivations on health and socioeconomic conditions at a small-area level. In order to identify the most influential TS fuzzy rules, some novel indices for TS fuzzy rule ranking are suggested and named as L-values and ω-values of TS rules respectively. The L-values of TS rules are suggested to identify important rules based on the effects of the local linear models in rule consequent parts. The ω-values of TS rules are defined by considering both the rule base structure and the contribution of the local linear models. The experimental results have shown that by using the proposed indices, the most influential TS fuzzy rules can be effectively selected for a parsimonious TS fuzzy model in predicting the impacts of child deprivation on education at a satisfactory level.