EXAMINATION OF THE RELATIONSHIP AMONG PLANT CHARACTERISTICS AFFECTING YIELD IN PEA PLANTS WITH MARS ALGORITHM
S. Celik1, A. Bakoglu2 and M. İ. Çatal3
1Department of Animal Science, Faculty of Agriculture, University of Bingol, Turkey
2Department of Plant and Animal Production, Vocational School of Pazar, University of Recep Tayyip Erdogan, Turkey.
3Department of Field Crops, Faculty of Agriculture and Natural Sciences, Recep Tayyip Erdogan University, Turkey
Corresponding author e-mail: firstname.lastname@example.org
This study was conducted to investigate the effects of some plant characteristics on green fodder and dry matter yield in pea plants. Plant characteristics; height (PH), pod number (PN), number of seeds in pods (SeedP), straw weight per plant (SeedW), straw yield (STY), seed yield (SeedY), Harvest Index (Hi) and 1000 Seed Weight (SeedWth) were evaluated for some yield traits: green fodder yield (GreHYi:), dry matter yield (DryHYi) . To estimate green fodder and dry matter yield in pea plant, two different MARS (Multivariate Adaptive Regression Splines) algorithms were performed. In both MARS models, a 2nd-degree interaction equation was obtained. To determine the suitability of the model, it has been considered that generalized cross-validation (GCV), root mean square error (RMSE), Akaike's Information Criterion (AIC) statistics to be minimum and coefficient of determination (R2) and adjusted coefficient of determination (Adj. R2) values to be maximum. In two separate MARS models formed to estimate green fodder and dry matter yield, R2 values were 0.998 and 0.998 respectively; Adj. R2 values were 0.999 and 0.998, RMSE values were 8.268 and 0.571, SDratio values were 0.037 and 0.019, and AIC values were 241 and 21. The greatest increase in green fodder yield in pea plants occurred when plant height was less than 42. The contribution of plant height to yield was 332 kg. The biggest increase in dry matter yield occurred when the harvest index was 20.5%. The contribution of harvest index to dry matter yield was 16.4. It has been noted that MARS is a good model in terms of predicting yield in pea plants.
Keywords: Pea, yield, MARS algorithm, Generalized Cross Validation (GCV)