MODEL DEVELOPMENT FOR ESTIMATION OF MOISTURE CONTENT OF BACCAUREA MOTLEYANA (RAMBAI) SEEDS
Determination of seed moisture content (MC) has conventionally been done by oven drying at 103°C for 16 h according to International Seed Testing Association. Modeling has recently been found useful for the prediction of MC of some recalcitrant seeds. In this study, predictive models were developed to estimate the MC of Baccaurea motleyana seeds. The seeds extracted from mature but unripe fruits were air dried at 25±2°C and 55±5% relative humidity (RH) for 1-5 days in a 24-hour air-conditioned laboratory. A total of 27 seeds were drawn randomly at pre-determined intervals. The weight (FW), length (L), width (W) and thickness (T) of each seed were recorded. Volume of seed, V, was calculated as L×W×T. MC of each seed was then determined by standard oven drying procedure. Two-thirds of the data of each desiccation period were used for model building. The relationships between response variables (MC; Y) versus predictor variables (FW, L, W and T; X) of seeds were first determined by scatter plot and correlation analyses. Then, regression models were developed for estimation of seed MC. The models were validated using the remaining one-third of the collected data. Among the seven candidate models developed, MC of B. motleyana seeds (Ŷ) can be estimated as Ŷ = 22.9 - 25.9L - 36.7T + 338FW in terms of simplicity and accuracy. The model is statistically significant (P<0.05) with R2 and SEE values of 0.84 and 5.88% respectively. In the validation procedure, I2 and RMSE were 0.89 and 4.87% respectively with this selected model. The good fit of selected model gives an indication of the potential of measured seed variables for estimating seed MC of B. motleyana. The method also provides an important basis for future research on recalcitrant and intermediate seeds of other species.
Hasmah, A., Tsan, F.Y., Suratman, M.N., Low , S.M. and Tajuddin, Z. (2012). MODEL DEVELOPMENT FOR ESTIMATION OF MOISTURE CONTENT OF BACCAUREA MOTLEYANA (RAMBAI) SEEDS. Acta Hortic. 932, 365-370
predictive model, seed, moisture content