<![CDATA[The oil content of oil palm fruit is a crucial parameter that must be determined before harvest, as it directly impacts crude palm oil (CPO) quality and processing efficiency. The conventional chemical method for oil content determination is costly and time consuming. This study aims to develop a non-destructive, accurate, and rapid method for predicting oil content in oil palm fruit based on its electrical properties. Measurements of electrical properties were taken across frequencies of 50 Hz to 5 MHz. Oil content of samples were determined by chemical method. Some pre-treatments of electrical properties were carried out and the pre-treated electrical properties were calibrated with reference oil content using simple linear regression and partial least squares regression. Linear regression model showed moderate accuracy (r = 0.61–0.81), with RMSE values between 9.54% and 12.99%. PLS regression models using admittance (r = 0.99, R² = 0.98, SEP 2.20%, RPD 7.99), resistance (r = 0.98, R² = 0.97, SEP 2.62%, RPD 5.56), and impedance (r = 0.98, R² = 0.95, SEP 3.16%, RPD 4.68) produced high prediction accuracy. The results confirm that electrical properties can be used to predict oil content in oil palm fruit rapidly and non destructively. Keywords: Electrical properties, Linear regression, Oil content, PLS.]]>
Copyrights © 2025
