Forecasting a Multivariate Time-Series
Time-series forecasting is an active research topic in academia. Forecasting long-term trends is not only a fun challenge, but has important implications for strategic planning and operations research in real-world applications such as IT operations management, manufacturing, and cyber security.
A multivariate time-series has more than one dependent variable. This means that each dependent variable not only depends on its own past values, but also potentially on the past values of other variables. This introduces complexity such as colinearity, where the dependent variables are not independent, but rather correlated. Colinearity violates the assumptions of many linear models, and it is therefore even more appealing to resort to models that can capture feature interactions.
This figure shows an example of a multivariate time-series, COVID deaths in different countries (from the English Wikipedia article about the COVID-19 pandemic):
...
