Movement of a prawn: a Hidden Markov Model approach
Johannes Alneberg · KTH Publication Database DiVA (KTH Royal Institute of Technology) · 2011
This thesis presents the basic theory for hidden Markov models and standard inference algorithms for these. The performance for these algorithms are examined with the help of a small test problem and the algorithms are found to operate properly. An application of hidden Markov models for modelling the movements of a prawn is implemented. This model is shown to be significantly better to use for the problem than a more simple model, where all movements are independent. The model also proves useful for finding the time points for where the prawn changed direction. As a brief example of how this model can be used, logistic regression is used to find that the distance to neighbouring prawns have an effect on the probability of changing direction.