I have some rated time-sequential data and I would like to test if an ANN can learn a correlation between my measurements and ratings.

I suspect I could just try a CNN where 1 Dimension is time or an LSTM/GRU and put the result through sigmoid, but is there any good literature on this? I have been trying to find information on datasets for the problem but it seems that Sequence regression is lacking any big official datasets, even though use-cases are there(e.g. learning personal music taste, try to predict rotten-tomato scores, etc..).

Looking for links to papers describing successful architectures or benchmarks where I can test my models.

  • $\begingroup$ A trajectory is equal to a sequence. It can be a spatial walk, but the trajectory can also become more abstract for example measured sensor readings, which are connected to a timestamp. The appropriate abstract model to explain a trajectory- generation, -prediction and -optimization problem is a simulation. This can be realized with a physics engine which mirrors the underlying system. $\endgroup$ – Manuel Rodriguez Oct 15 '19 at 8:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.