Logo image
Using Hidden Markov Model to Predict Human Actions with Swarm Intelligence
Conference paper   Peer reviewed

Using Hidden Markov Model to Predict Human Actions with Swarm Intelligence

Zhicheng Lu, Yuk Ying Chung, Henry Wing Fung Yeung, Seid Miad Zandavi, Weiming Zhi and Wei-Chang Yeh
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol.10637 LNCS, pp.21-30
2017

Abstract

Hidden Markov Model Human activity prediction Particle Swarm Optimization Theoretical Computer Science Computer Science (all)
This paper proposed a novel algorithm which named Randomized Particle Swarm Optimization (RPSO) to optimize HMM for human activity prediction. The experiments designed in this paper are the classification of human activity using two data sets. The first testing data is from the TUM Kitchen Data Set and the other is the Human Activity Recognition using the Smartphone Data Set from UCI Machine Learning Repository. Based on the comparison of the accuracies for the conventional HMM and optimized HMM, a conclusion can be drawn that the proposed RPSO can help HMM to achieve higher accuracy for human action recognition. Our results show that RPSO-HMM can improve 15% accuracy in human activity recognition and prediction when compared to the traditional HMM.

Metrics

1 Record Views

Details

Logo image