Machine Learning Model to Determine Dominant Features in Palm Kernel Cake Quality

This title has been presented on Friday, December 15, 2023 at 13.25-13.35 GMT+7.

Abstract

This title has been presented on Friday, December 15, 2023 at 13.25-13.35 GMT+7.

Pre-processing is an important stage in data classification as it can address the complexity of object-related problems and generate high-quality and accurate patterns. For instance, in determining the quality of palm kernel cake in concentrate formulation. The current issue is the limited interest in using palm kernel cake due to the presence of shell contaminants in the palm kernel cake. Analyzing the quality of palm kernel cake requires time-consuming and costly laboratory tests. Testing and identifying the dominant features that significantly affect the quality of palm kernel cake is expected to minimize costs and improve time efficiency. One commonly used data pre-processing technique to determine the most influential attributes in a dataset is feature selection. The data used in this research is the proximate test results of palm kernel cake in the laboratory. This study employs the correlation-based feature selection method. The objective is to identify the most relevant and influential attributes in the dataset for predicting the dominant features in determining the quality of palm kernel cake. The study includes five attributes: dry matter, ash, crude protein, crude fiber, and crude fat. The results will reveal the order and degree of their relationships with each other.