SISR based Hidden State Estimation of HMMs with transition density function
Longteng Li, Chengwen Zhu, Xiaoyan Cai, Chi Zhang, Chuizhen Zeng · Proceedings of the 2nd International Conference on Computer Science and Electronics Engineering (ICCSEE 2013) · 2013
Traditional Viterbi algorithm cannot be generally effective.Regarding the hidden state estimates of HMM as a Bayes filtering problem, the Sequential Importance Sampling with Resampling algorithm could get an approximate of its Bayes estimates.Its performance reached or even exceeds the Viterbi algorithm while lower dependence on the model, having a wider range of adaptation.