Multilinear Regression Approach in Predicting Osmo-Dehydration Processes of Apple, Banana and Potato

The potential for improving food quality through osmo-dehydration is tremendous but limited by quantitative data and methods. A Multiple Linear Regression (MLR) approach was developed for water loss and solid gain during osmo-dehydration of apple, banana and potato taking into account the effect of temperature, concentration, time of immersion, sample size, sample type and agitation.

Tortoe, C.; Orchard, J.; Beezer, A.

2011

English

The potential for improving food quality through osmo-dehydration is tremendous but limited by quantitative data and methods. A Multiple Linear Regression (MLR) approach was developed for water loss and solid gain during osmo-dehydration of apple, banana and potato taking into account the effect of temperature, concentration, time of immersion, sample size, sample type and agitation. Temperature was the most important factor influencing osmodehydration of the plant materials whereas agitation was the least. A regression coefficient of determination (R2 = 0.886) indicating a good correlation coefficient (r = 0.941) between experimental and predicted data was identified for water loss. However, the regression coefficient of determination (R2 = 0.305) for the solid gain did not show a good regression correlation coefficient (r = 0.552) between the experimental data and the predicted data. Prediction of water loss was more adequate than solid gain due to the variability of the pathways of water and solid diffusion into the different plant materials in favour of water loss.