Modeling and Optimization of Signals Using Machine Learning Techniques

19Computing in Cognitive Science Using Ensemble Learning

Om Prakash Singh

Department of Computer Science and Engineering, Vidya Vihar Institute of Technology, Purnea, Bihar, India

Abstract

Cognitive science is deeply ingrained in artificial intelligence (AI). The phenomenon of thinking and acting either humanly or logically is how it is defined as an AI. Many academics have been concentrating on human activity recognition using ensemble learning in recent years. The dataset of human activities is unbalanced. This collection is a reflection of all human thoughts and behaviors that effectively forbids the coexistence of cognitive science and AI. The term “unbalanced dataset” refers to a set of data where at least one class has more instances than the other classes combined. The goal of machine learning with such an unbalanced data collection is to continuously learn from the skewed data distribution. Many common machine learning techniques, such as linear regression, k-nearest neighbor, and decision trees, perform poorly in terms of accuracy with this kind of dataset (77%, 62%, and 89% accuracy, respectively). Boosting is a general machine learning technique for raising any learning algorithm’s current accuracy. It is built on the simple idea that a group of simple classifiers working together can perform significantly better than a single classifier working alone. The foundation of ensemble learning is accelerated learning techniques. Simple classifiers are weak learners, while ...