I would like to extract a representative pattern from a univariate time series data set

I am a novice user of Matlab. I found out within stack overflow (https://stackoverflow.com/questions/11752727/pattern-recognition-in-time-series), that people have used HMM (Hidden Markov Models) to extract the shapelet for time series data. Since Matlab, already has this feature in its Machine Learning Toolbox, what would be the ideal way to implement it in order to obtain the desired output.

Answer 1

[Jaynik]

Hi Deepak,

A Hidden Markov Model (HMM) is a model in which you observe a sequence of emissions, but do not know the sequence of states the model went through to generate the emissions.

The following documentation provides an example of a Markov model with two states and six possible emissions, with transitions determined by weighted coins and emissions by dice:

• https://www.mathworks.com/help/stats/hidden-markov-models-hmm.html

The Statistics and Machine Learning Toolbox provides several functions related to Hidden Markov Models like hmmgenerate, hmmestimate, hmmtrain, etc. The above documentation provides a step-by-step guide on how to use these functions.

Hope this helps!