Hello,
I want to predict land-use changes for the distant future (like 2050 or beyond). my data collection is for 1995-2005-2015. So:
1- How to predict for the future? Is creating a model based on for example 1995 to 2015 and predicting based on 2015 information suffice? (and then maybe predicting based on the result of 2015 prediction). Is this type of stepwise prediction the only way?
2- Is there any research paper you could suggest that predicts landuse change for distant future using machine learning techniques?
thanks.
Conference Paper Predict Land Covers with Transition Modeling and Incremental Learning
https://www.ingentaconnect.com/contentone/asprs/pers/2003/00000069/00000007/art00005?crawler=true
http://dare.ubvu.vu.nl/bitstream/handle/1871/46586/270022.pdf?sequence=1
https://www.cc.gatech.edu/gvu/people/faculty/greg.turk/my_papers/predicting_future_land_cover.pdf
Hi Mojtaba Asgharzadeh
The prediction of LULC based on ML using historic growth trends alone wont yield better results. The reason is below:
# The land use system is a complex adaptive system, prediction of any complex system is not possible, since they linked with many internal and external factors. and including all into a model is not possible.
# Models are abstraction reality, and especially they meant to understand the land use system and to support scenario analysis.
I would suggest to go through the following paper for more details of scenario based modelling approach rather project based on only historical trends.
Article SUSM: a scenario-based urban growth simulation model using r...
In addition to ML, you may need to use CA or ABM to make your model spatially explicit.
Regards
Hi Mojtaba Asgharzadeh ,Hope you fine,
Because of very diffrent and multiple criteria that effect on land use, we couldn't predict it for long-term interval , and our prediction will be full of fault and mistak,
It's commont that predict land use chang for near future
Hello,
Geoprospective is a difficult discipline to manipulate, not because it deals with the future, but because humans are like the Nature they are from : unpredictable. So, there is a possible solution : using together the nomothetic (spatial analysis) which explains us what is permanent through time (and so, we can suppose it will still be the case in 50 or 100 years from now) but also axiology which deals with ethics, esthetic but also (that's a personnal opinion) with cognitive science. Axiology is not permanent, even if scientific methods were tried on it, ethics change through time and between societies, esthetic too and i don't talk even talk of minds (the cultural matter) ! ... The mixing of those two sides should produce something interesting, as long as there is a perfect logical, reasoning and objective work.
See you here or there.
Térence
- Badges
- Science topic