I am trying to model Sales which present at various levels (like Product-Category, Product etc.). Time is also one of them. Is there any way by which we can identify whether this is a time series data or not.(I read somewhere that Time series process is a stochastic one ??). I need this to decide on which techniques to use (ARIMA,Regression etc.) ?
I’ll suggest you to go through this article on AV for time series : https://www.analyticsvidhya.com/blog/2015/12/complete-tutorial-time-series-modeling/
Hope it helps.
It all Depends on your approach how you want to model your data, Time Series problem which has features can be modeled by using Regression, Trees, SVM or Neural Net Regressor.
ARIMA is not suitable if you want to use features and trying to do a long term forecasting because the long term forecasts in ARIMA tends towards mean value.