Hello Mohammad, Please go through herewith the attached code in matlab for ann based landslide mapping, i thought that this may helpful to you for further work.
I would suggest you do not pay attention to the content of ANN.zip at all. It looks like some random downloaded codes which are mostly useless (http://www.mathworks.com/matlabcentral/fileexchange/32905-neural-network-simple-programs-for-beginners/content/neural1.m ). They seem to demostrate training of simple perceptrons and do not have any practical value for you.
For working with NN, you can look at Matlab NN Toolbox the http://www.mathworks.com/help/nnet/index.html.
You should, most importantly, properly define what you would like the network to learn and predict and you should realize what data you have available.
What is the input you have? Maps? Rainfall measurements? Geological data? Do you have a large amount of historical data or data assesed by humans?
You have to realize that NN can only do things which you show them how to. E.g. if you want them to segment satelite images, you have to have a large dataset of pixel-labeled images. NN are useless without data.
ok thanks , i use 10 kind of data as e.g (topography, inventory landslide, distances from fault ,river,etc) for raster map, but i have never work & input on matlab, i want learn how i can frist import our data in matlab, with which command
In Matlab, you can use variety of formats for which import (read) function are already implemented. The best place to start is http://www.mathworks.com/help/matlab/data-import-and-export.html .
If you can produce a text file with your data in a matrix, such as:
1 2 0.5 0.4
1 5 0.1 2
...
You can simply type load( fileName) which will load the file as a Matlab matrix.
Function dlmread() is usefull as well and uiimport() enables you to specify how to import the data in graphical interface.
For images imread().
You can also read files the same way as in C with low-leve function, such as fopen(), fscanf(), fread() http://www.mathworks.com/help/matlab/low-level-file-i-o.html .