 |
 |
Home NEWS SAMT SAMT- download Applications Developer |
| Institute of Landscape Systems Analysis | ||
| download programs |
|||||
SAMT3.1 Please make a directory "samt" under your home directory and change to it (cd samt). In samt make a directory "src" and unpack the source code there (cd src; tar zxvf samt_src.tgz). Please read the README.txt and the INSTALL.txt before you start to compile. The 3d part of SAMT is an add on (3d.tgz) and should be installed under samt/3d (cd samt; mkdir 3d; cd 3d; tar zxvf 3d.tgz) SAMT3.1 ------> samt_src.tgz 3d -------> 3d.tgz SADATO The SAMT data tool SADATO can be used as stand alone tool. Together with SAMT it should be installed under samt/sadato (cd samt; mkdir sadato; cd sadato; tar zxvf sadato_src.tgz SADATO -------> sadato_src.tgz |
|||||
SAMT-Fuzzy Fuzzy development toolbox including Fuzzy Generator |
|||||
| The Fuzzy Generator was developed as additional tool to SAMT-Fuzzy to allow the semiautomatic training of a fuzzy model using data. Both programs "samt_fuzzy" and "fuzzy_g" are combined to a fuzzy development toolbox which can be used together with SAMT or stand alone. It can be compiled under Linux (Debian Squeeze and Ubuntu 10.04 LTS,...). Please open the "samt_fuzzy.tgz" (tar zxvf samt_fuzzy.tgz), go to to the new directory "SAMT_FUZZY" and follow the instructions in INSTALL.txt samt_fuzzy.tgz |
|||||
SAMT Including Samt Data Tool (SADATO) |
|||||
The final version of SAMT 3.0 is out. It runs in the virtual box as a virtual machine with zero installation under all common operationsystems like Linux, Mac-OS and Windows. It only depends on the virtualbox from SUN: http://www.virtualbox.org/.The new version contains the Samt Data Tool Box (SADATO) anda new fuzzy generator for fuzzy rule training supported by statistics. It can be used to find rules, to find gabs in the knowledge, to find gabs in the data. This can be very helpful for environmental science. Please help to make the new version better and write anemail if you have any problems. This has big advantages: 1.) The user can use SAMT under Linux, Mac and Windows as host system 2.) The update to a new version of SAMT is simply a import. 3.) The exchange between host and guest is supported by the VirtualBox From point of view of the developer of SAMT the software maintenance is concentrated to one Version and not different (CD-Version, USB-STick-Version, etc.) This makes it easier to communicate with the user. Installation Please load all three files in a new Directory on your machine. - {ebb971d6-a7b3-4097-a975-91e655583d19}.vmdk - samt.mf - samt.ovf Install VirtualBox and import the loaded files. Then you can start new generated SAMT in VirtualBox. Here is a hint how to share data between the host and SAMT: When done with you shared folder(s) specification, you may now go ahead and actually mount these folders from Ubuntu. First you have to create a mount point, that is, a directory in your Ubuntu which will reflect the shared folder from Windows:
|
|||||
Including Samt Data Tool (SADATO) Download sadato.tgz SADATO Help |
|||||
| README is an integrated development platform for neural networks (NN) and Support Vector Machines (SVM). It also combines a data base engine (sqlite3) with the statistical library apophenia. This allows the efficient storage of data sets and the statistical analysis of the data. All is wrapped by a graphical user interface written in QT and developed under the operating system Linux. To compile it please install first:
The easiest way to do this is using a modern Ubuntu and select this software via the packet manager Synaptic The additional basis software for SADATO must be installed from source:
The installation is strait forward and should run smoothly. Is all installed then unpack the sadato.tgz: tar -zxvf sadato.tgz and go in the new created directory "sadato". cd sadato In this directory there is a file sadato.qwt which is used to create a Makefile. Please check the software and the path name and adapt that if necessary. To compile it type:
To use SADATO the environment variable: export SADATO=/home/ralf/sadato should be included in the .bashrc of your home directory. SADATO itself expects that all data set are read in as CSV-Files (exported by Open Office for example) and stored as data base table before it can be analyzed. This helps to organize all your data sets. The combination of multiple statistics, neural networks and support vector machines allows to compare the results. SADATO can be used to handle regression problems as well as for classification problems. The development of SADATO is not finished yet. We have some ideas like: 1.) include a graphical analysis for a selected column. This includes a simple graph and a continuous wavlet transformation to detect some cycles in the data set. 2.) add new radial basis function network and counter propagation network. It would be nice if we could get some response about SADATO. Please write a email to: Dr. Ralf Wieland: rwieland@zalf.de or Karin Groth: kgroth@zalf.de SADATO is published under GPL V3 http://www.gnu.de/documents/gpl.de.html |
|||||
| SAMT-
Source (about 30 MB) SAMT-Version 2.8.5 << part I >> <<part II >> install-procedure The source are packed in a tgz file samt.tgz please install it with: - tar -zxvf samt_data.tgz and compile it. The best location for samt is /usr/local/samt for the sources and binaries and ~/new_samt for the data. Please install the following libraries before compiling SAMT:
Remark: use the bashrc as example for your .bashrc |
SAMT-USB-Stick make_stick.txt knoppix-data.img knoppix.iso samt.tgz SAMT- CD (about 700 MB) SAMT-Version 2.8.5 - load the iso image to your computer - burn this image on CD This image contains a Knoppix system including SAMT Shutdown your computer, then boot from this CD - open a shell - start samt by typing: new_samt |
||||