Prediction Using Knowledge Growing System: A Cognitive Artificial Intelligence Approach
Arwin Datumaya Wahyudi Sumari, Rosa Andrie Asmara, Dimas Rossiawan Hendra Putra, Ika Noer Syamsiana · 2021
In the Artificial Intelligence (AI) world, prediction is a means for recognizing a phenomenon. Three approaches have been used for this task, namely supervised, semi-supervised, and unsupervised. Data annotation or labeling is a must in the first two methods, while the last method clusters the inputted data without knowing the labels. The most basic for those methods is a training data set has to be provided at first and a massive amount of data for the supervised one. This paper used a new means called Cognitive Artificial Intelligence (CAI) Knowledge Growing System (KGS) for the prediction task. KGS is characterized by its cognitive learning capability to learn by interacting directly with the phenomenon. Therefore, it is prospective to overcome the existing methods in the context of past data providing and data recognizing through just-in-time labeling. We also show that KGS succeeds in predicting some data with 100% perfect accuracy.