How does time series work with multiple time series data sets on the same index?

For example, suppose I were a utilities company. Suppose I have the electricity usage of two homes, each indexed for the same time period, daily from January 2018 - December 2018.

If I wanted to train a model that could predict the daily t+1 energy usage based on data from January 2018 to time t ∀𝑡>12/31/2018, how would I accomplish this?

How would I format my input data matrix for the various time series forecasting models or neural networks like LSTM or RNNs?

House 1: 50kwh ...

House 2: 15kwh ...

<-----|-----|-----|-----|-----|-----|-----|-----> To be clear, I am not asking how to predict the aggregated sum total of energy usage. I am interested in predicting the next day usage of each house.

Would I need to have two separate models, or could the weights of one single model accomplish this task? If the latter, how do I format my dataframe?

Specifically, I tasked with applying a time series forecasts on financial option data (which I know is likely not going to result in good predictions but I am tasked nonetheless).

So I have a variety of historical options with varying lifespans. I am planning to use the last 100 days until expiry for financial option data as my training set. This way, I can use options throughout time. The only problem is, I want my model to be able to train off of both systematic and nonsystematic fluctuations. How can I achieve this?


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.