Caching in TimeSeries

Hi,

I have prepared a time series model. The model works well with few cases. Answering time raises extremly after processing the model with all cases.

In a book I read sth. about a cahing possibilty. But how can this be defined in Visual Studio ?

Thanks in advance
Achim

Since the time series algorithm does not do predictions on new input data (it uses the historical data that the model was trained with), it returns the same result each time you request a prediction for the next N steps, for a given N. You can take advantage of this by getting the results for the maximum number of steps you're likely to request in one query, saving the results on the client (in an in-memory list or on disk) and then using the saved results for future requests against that model instead of sending the prediction query to the server. You would only need to refresh the cached results if the model gets reprocessed with new data.|||

The algorithm has a certain built in mechanism for caching. For instance, if you ask for the next 50 forecasted values, the first query will be relatively slow. Subsequent queries for less than 50 values should be much faster. However, asking for the first time for 51 values will also take a longer time.