How use HMM correctly for Transition and observation probablities

1 次查看（过去 30 天）
Masood Abbasi
Masood Abbasi 2018-5-2
I want to compute Transition probabilities (Matrix) and emission probabilities (Matrix) from observations of 40 seconds in the form 'g' 'b' 'g' 'b' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' with 'g' as good condition and 'b' as bad condition respectively. i ran hmmtrain with seq= ['g' 'b' 'g' 'b' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g' 'g'] and guessed transition and emission probabilities but it gives error for seq as it should be numeric then i put 1 for 'g' and '2' for bad observations and it ran without error. Now questions are: 1. Is conversion for g to 1 and b to 2 correct or it could be any number or should it be 0 and 1?
2. What initial Transition and Emission Matrix should be when i know nothing about a system?
3. When Should i Stop recursively calling hmmtrain(). Is there any other method best fit?
4. How can i train my model through this?

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Markov Chain Models 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by