A Quantum Computing Approach to Human Behavior Prediction

Abstract

As quantum computing technologies become more mature, their applicability increases. One of the main challenges in intelligent environments is to correctly model and ascertain the users' behavior in order to react to it and cater to their needs. One of the main challenges in human behavior modeling is predicting the users' next actions. In this paper we propose using two different quantum computing algorithms in order to predict human behavior: Quantum Kernel Alignment and Quantum Support Vector Machines. Our experiments show that those algorithms outperform other traditional machine learning algorithms in this task. We also present a study that analyzes the influence of qubit noise in the performance of the quantum approach. This helps to understand how the accuracy of the quantum computing algorithms will increase as the underlying hardware matures and qubit noise is reduced.