Is there any algorithm used in machine learning which takes the cross-sectional impacts in data models?To give an example, the vibration in a factory is formed by the individual vibrations of separate machines. A further examination would be how similar may ML be with the panel data which is used extensively in econometrics?
Sorry the forum did not permit to answer so I write here.
I have 3 areas in my mind.
First is the vibration and other related KPI’s as temperature and humidity as we have deployed thingworx in Arçelik(One of the biggest electronic companies in Europe) and now they entered Thingwatcher beta program.
Second is a use case for a retailer named LCW(Probably the third biggest cloth retailer in Europe). We try to model a sales prediction model for their retail shops for them which will be prescriptive. The independent variables may be spatial(geographic locations, number of customers entering the shopping malls, number of items in collection etc)
Third, how can we use historical data for macroeconomics?Nowcasting is a kind of an interesting approach these days as traditional multi-regression or panel studies can not answer precisely.