APLIKASI MODEL REGRESI STEP WISE DALAM PENENTUAN HASIL JAGUNG PUTIH
No Thumbnail Available
Date
Journal Title
Journal ISSN
Volume Title
Publisher
Sekretariat Badan Penelitian dan Pengembangan Pertanian
Abstract
Description
White corn is a source of functional food, particularly in dry land and dry climate environments. Indonesian Agency for Agricultural Research and Development (IAARD) has released six white corn varieties with various characteristics with wide adaptation area. Breeding programs for developing high yielding white maize varieties is conducted by a series of steps including evaluation of agronomic characteristics of lines/varieties. The objective of the research was to test the applicability of step wise regression model to determine/select agronomic variables that significantly contributed to the yield of white corn. Research was conducted in rainy season 2012 at Muneng experimental station, East Java Province. Maize parental plants used were the result of crossing CML140 x CML264Q lines namely Bima Putih. A total of 14 variables were tested for their significances to maize yield. SPSS and Microstat software were used to calculate the best variables that contributed to the yield significantly. The results showed that among 14 variables involved in the calculation, there were five variables that contributed to the yield, namely: weight of cob at harvest (x5), ratio of the weight of dry grain to the total cob weight (x9), the seed moisture content (x6), and number of ears (x4). Regression models generated from step wise screening was y=-4,33+0,763x5+0,1009x11+0,104x9-1,22x6+0,016x4. The coefficient of determination (R2 ) of the model was 0.99, indicating the ability of the regression model to fit the data. This showed that the five independent variables included in the model were the main variables in determining the outcome of hybrid seed Bima Putih-1. This result could be further used as reference to conduct parameter screening to produce high yielding white maize
Keywords
white corn; model; step wise; yield