HMMs can be used very well to model processes which consist of different stages that occur in definite (or typical) orders.
If, for example, you want to model the behavior of a technical system that first boots, then operates, then enters sleep mode, and iteratively changes between sleep and operation later on, you might need three states (boot, operate, sleep) and can use this process model to find out what's going on in the system at any one time - using fairly effective algorithms that even fulfill reasonable optimality criteria.
HMMs can be used very well to model processes which consist of different stages that occur in definite (or typical) orders.
If, for example, you want to model the behavior of a technical system that first boots, then operates, then enters sleep mode, and iteratively changes between sleep and operation later on, you might need three states (boot, operate, sleep) and can use this process model to find out what's going on in the system at any one time - using fairly effective algorithms that even fulfill reasonable optimality criteria.
HMM is used when you have a state-machine system and you don't know the states (hidden states), but you know the observations that produced from that states. In this case HMM uses the observations to get the states.
It can be used for time series classification if the given time series data is an output of a state-machine system.
HMM is a powerful modeling technique when the system states are partially observable and the behavior of the system is considered as autonomous. Although with being partially observable, POMDP could be an effective alternative but the nature of autonomy of the modeled system will determine the ultimate selection of Markov models. Generally system autonomy is dependent on the agent decision making ability with concern of human intervention. HMMs are variant of FSMs where the flexibility of decision process could be perfectly implemented and the selection of output is basically could be described as real system outputs and to reach for efficient performance.
can i use HMM to find some rule from observable geometry design. i have generated data of images (200 synthetic images) and i would like to extract the rules that allow to regenerate these images or any other with same rules.
HMM is suitable to be used in application that dealing with recognizing something based on sequence of feature. The decision making is done based on sequence of information. Same as DTW or edit distance. However, for kNN, ANN or SVM, there are under same category which is one-off decision making. Meaning that the decision is made after one static features was given.
HMM is widely used in computational biology and Bioinformatics.
However HMM's are yet to be explored for agricultural data. Application of HMM requires that the phenomenon under study can be characterized by a parametric stochastic process whose parameters can be estimated precisely.