A Decision Tree Knowledge-based System for Reviewing Research Ethics Protocol

Abstract

Knowledge base systems have undergone many developments in providing similar results to experts. This can help increase time effectiveness in determining decisions and analysis results. Several methods have given good results in determining decisions based on knowledge, one of which is using a decision tree. In this study, the researchers applied decision tree modeling to determine the results of the review on ethical research protocols. Our target is to classify ethical protocols into one of three decisions: Exempted, Expedited, or Full Board. Three decision tree models are used in this research to evaluate the best results that can predict the ethical review protocol results according to the expert's dataset. The experiments showed that all models showed the same result, with an accuracy value of 0,91, precision of 0,93, and recall of 0,91. However, manual checking showed that the second model with Gini criteria parameters and class weight balance resulted in 10 data correctly predicted based on the dataset used.