Latent state recognition by an enhanced hidden Markov model
作者：
Highlights：
• The hidden Markov model is extended to relax two primary assumptions.
• The observation is modelled as a vector autoregressive process.
• The hidden state evolution follows a semi-Markov chain.
• The autoregressive and covariance matrices are regularized with penalty.
• The proposed model shows a promising performance in latent state recognition.
摘要
•The hidden Markov model is extended to relax two primary assumptions.•The observation is modelled as a vector autoregressive process.•The hidden state evolution follows a semi-Markov chain.•The autoregressive and covariance matrices are regularized with penalty.•The proposed model shows a promising performance in latent state recognition.
论文关键词：LASSO,Vector-autoregressive model,Hidden Markov model
论文评审过程：Received 10 September 2019, Revised 8 June 2020, Accepted 3 July 2020, Available online 17 July 2020, Version of Record 25 July 2020.
论文官网地址：https://doi.org/10.1016/j.eswa.2020.113722
