﻿
/***************************************************************************

Software: Predict version 0.1

Type: A QGIS plugin

Role: Predict urban evolution
                             
-------------------
        
begin                : 2012-12-17
        
copyright            : (C) 2012 by Asma GHARBI
        
email                : asma.gharbiii@gmail.com
 **************************************************************************

/*************************************************************************
                                                                      * *      
This program is free software; you can redistribute it and/or modify  *
 *   
it under the terms of the GNU General Public License as published by  *
 *   
the Free Software Foundation; either version 2 of the License, or     *
 *   
(at your option) any later version.                                   *
 *                                    ***************************************************************************

Installation notes

------------------

-Install a compatible Qgis version under a linux distribution:recommanded Qgis 1.8 on Ubuntu 12.04

-Place Predict file under the .qgis/python/plugins folder in your home directory
.
-The folder structure should look like this:





.qgis
├── python

       │
       └── plugins

         │        
         └──plugin folder

         │        
         └──plugin folder

         │        
         └──plugin folder






Use the plugin manager of Qgis, found under the plugins menu to activate and start using Predict.




Usage notes

-----------

-

To run the prediction process:

-launch Qgis as the administrator.
-click on the Predict icon in Qgis.
-select the semi-Prediction case
.
-select the learning file (arff).

-press the OK button to perform the prediction

-To display the predicted land uses, click on the menu on "Layer"->"Labelling". Then you can choose the field name FFun (the predicted land uses).





Interpretation of Prediction results:

------------------------------------


The prediction process consists in giving all the possible land use evolutions based on current land uses.
Thus, to evaluate the prediction, we deﬁned the prediction Precision degree (P) as the proportion of correctly predicted changes (tp) among all the predicted ones:  True predictions (tp)
and false predictions (fp).

P = tp/tp+fp







