首页>
外国专利>
Factorial hidden Markov model with discrete observations
Factorial hidden Markov model with discrete observations
展开▼
机译:离散观测的阶乘隐马尔可夫模型
展开▼
页面导航
摘要
著录项
相似文献
摘要
A method for analyzing hidden dynamics, includes acquiring discrete observations, each discrete observation having an observed value selected from two or more allowed discrete values. A factorial hidden Markov model (FHMM) relating the discrete observations with a plurality of hidden dynamics is constructed. A contribution of the state of each hidden dynamic to the discrete observation may be represented in the FHMM as a parameter of a nominal distribution which is scaled by a function of the state of the hidden dynamic. States of the hidden dynamics are inferred from the discrete observations based on the FHMM. Information corresponding to at least one inferred state of at least one of the hidden dynamics is output. The parameters of the contribution of each dynamic to the hidden states may be learnt from a large number of observations. An example of a networked printing system is used to demonstrate the applicability of the method.
展开▼