Modeling and Prediction of Human Behavior Based on KD-CNN Model

A person's feelings and thoughts about a communication robot have a significant impact on how the two connect. Robots that are both effective and sympathetic must first have a firm grasp of these emotional dynamics. The complex relationship between people's feelings, thoughts, and actions in relation to robots is investigated in this empirical study. Preprocessing input, selecting features, and sequencing are the three main steps in training a model. The Iliou preprocessing method improves the quality of the original data by creating a refined dataset from it. Feature selection makes use of principal component analysis (PCA), which finds important features by analyzing class-feature covariance values. In order to train KDCNN models to operate at their best, feature selection must be done effectively. With an impressive accuracy rate of 96.55%, the suggested method outperforms two state-of-the-art algorithms, KD and CNN, according to the comparative research. More intuitive and responsive robotic systems can be created as a result of this research, which adds to our knowledge of the dynamics of human-robot interaction.