| Abstrakt: |
The use of different criteria for the definition of the reference population, combined with measures of accuracy of nutritional diagnoses, can enable the definition of the best set of Integrated Diagnosis and Recommendation System (DRIS) standards for sugarcane. In this context, the objective was to use the accuracy measures proposed by Wadt and Lemos to define the best criterion for defining the reference population for the DRIS method in sugarcane, based on the nutritional status for phosphorus. The database was created based on productivity and leaf contents of N, P, K, Ca, Mg, S, B, Cu, Fe, Mn and Zn for sugar cane, considered in two agricultural years (2013 and 2014), making a total of 1,224 samples, composed of five reference populations that showed productivity above 172; 193; 214; 235 and 256 t ha−1. The Average Nutritional Balance Index (NBIm), DRIS index and productivity were studied; degree of agreement between standards and measures of accuracy of nutritional diagnoses for phosphorus. The measures of global accuracy, excess and nutritional deficiency, showed diagnostic performance below 50% in the 2013 and 201 harvests. The accuracy for balance, sufficiency and nutritional toxicity showed high performance of the nutritional diagnoses regardless of the harvest. The analysis of the accuracy of the nutritional diagnoses for phosphorus identified low performance of all sets of DRIS standards, however, with better performance of the standards derived from reference populations of average productivity + 1.0 standard deviation in the 2013 and 2014 harvests. [ABSTRACT FROM AUTHOR] |
