2021/02/14 by Taha Mansouri, Mansouri, Taha, Mohamadreza Sadeghimoghadam +3
Computer Science · Mathematics · #Algorithm #Anomaly Detection Techniques and Applications #Artificial intelligence #Benchmark (surveying) #Computer science #FOS: Computer and information sciences #FOS: Mathematics #Hidden Markov model #Machine Learning (cs.LG) #Machine Learning and Algorithms #Machine learning #Optimization and Control (math.OC) #Robustness (evolution) #Time Series Analysis and Forecasting #cs.LG #math.OC
paper · pdf · doi:10.48550/arxiv.2102.07112
published in arXiv (Cornell University) (Cornell University) · arXiv admin note: text overlap with arXiv:0811.4413 by other authors
arxiv created 2021/02/14 · openalex publication_date 2021/02/14 · arxiv updated 2021/02/16 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/06
This research focuses on the algorithms and approaches for learning Hidden Markov Models (HMMs) and compares HMM learning methods and algorithms. HMM is a statistical Markov model in which the system being modeled is assumed to be a Markov process. One of the essential characteristics of HMMs is their learning capabilities. Learning algorithms are introduced to overcome this inconvenience. One of the main problems of the newly proposed algorithms is their validation. This research aims by using the theoretical and experimental analysis to 1) compare HMMs learning algorithms proposed in the literature, 2) provide a validation tool for new HMM learning algorithms, and 3) present a new algorithm called Asexual Reproduction Optimization (ARO) with one of its extensions - Modified ARO (MARO) - as a novel HMM learning algorithm to use the validation tool proposed. According to the literature findings, it seems that populationbased algorithms perform better among HMMs learning approaches than other algorithms. Also, the testing was done in nine benchmark datasets. The results show that MARO outperforms different algorithms in objective functions in terms of accuracy and robustness.