Growth Depends on Some factors as:
- Distance between two cities
- Distance from Roads
- Elevation of City
- Near to outer area of city
- Trend of change from Agriculture to Urban, or Forest to Urban
- The actual landuse
- Rate of change
We can predict future landuse in two ways, Using Regression model or By CA Markov Method with Machine learning approach to solve complex problems.
So CA Markov is best to predict with MLP and Neural Network approach. This way we train our model with 50% learning and 50% testing approach. But any learning accuracy between 40% to 70% will provide accurate results. It depends upon how we are getting output results. Each time after modification check results by comparing to real landuse if it matches at least 70% then continue to predict for future. All task are too easy some work need to do in ArcGIS and some in Terrset Model. Both are NOT open source.
If you want to see in Detail Click Here to Know More
See in Video Below How it works.
