# Can LSTM be used to predict value as regression problem?

I have time-series data as shown below. Which model is generally preferred if grig_id is needed to be predicted? Is it possible to use LSTM with a sigmoid activation function at the final layer for such a task?