In this research, we have modeled the relationships between leaf nutrient concentrations and the yields of avocado trees with the aim of developing decision support tools for improved fertilization and nutrient management to increase avocado fruit yields. Using a data base of ~3500 observations in which nutrient concentration profiles and yields of individual trees were examined over several harvest seasons, we now present in our final report a refined model that predicts nutrient-yield relationships based on all possible combinations for the 11 elements that are monitored by leaf analyses. The research conducted represents an intellectual evolution of ideas on how to mine leaf and soil analysis data sets using different filters and mathematical methods to extract relationships between selected variables such as yield, fruit quality, and chloride toxicity (patent application filed, UCR 2016) Our modeling approaches involved a range of complimentary and independent methods that included artificial neural networks, Kohonen self-organizing maps, frontier/envelope analysis, and quantile regression. The ultimate product from our research is the translation of these data into working equations and lookup tables with macros that can be employed in sophisticated software for use as a decision support tool.