How to use neural networks for spatial prediction ?

Hello Everyone I have data of n different location spread all over my study area, each location has a number of independent variables and only one target variable. Some of the independent variables are static like the latitude, longitude, and (mean, min, and max) altitude, while the others are time series variables like the precipitation and temperature, the target variable is also a time series variable. My question is: how to arrange the independent variables (static and time series) to be used within the neural networks for the application of spatial prediction? is the structure in the attached image is right, where I have repeated the values of the static variables for each time series variable within each location?

My second question is how to normalize the input variables of the same type as the mean, min, and max altitude? Should I need to normalize them independently like the other variables or I should normalize them like a single variable?

Answer 1

[Greg Heath]

Normalize each input independently of the others.

However, it is always wise to first check the input variable correlation coefficient matrix for high cross-correlations.

Hope this helps

Thank you for formally accepting my answer

Greg