How to train selforgmap with trajectories?

I have a data with trajectories(x,y points) of moving people and i want to classify trajectories if its normal or unusual behavior using SOM but i dont know how to do it and i cant find any similar examples of this problem.Don't know even how to set input as both files with X and Y points.

Answer 1

[Yash]

Hi,

To train a selforgmap with trajectories, you need to reshape your trajectory data into a suitable format for training. The selforgmap function in MATLAB's Deep Learning Toolbox is designed for vector data, so the trajectory data is required to be reshaped into a cell array of sequences, this can be done using the con2seq function. Each sequence will then represent a trajectory, where each element of the sequence corresponds to a time step.

To know more about the selforgmap and con2seq functions, you can refer to the documentations here:

1. https://in.mathworks.com/help/deeplearning/ref/selforgmap.html
2. https://in.mathworks.com/help/deeplearning/ref/con2seq.html

Hope this helps