Installation

From Eigenvector Research Documentation Wiki
Revision as of 15:48, 8 May 2012 by imported>Jeremy
Jump to navigation Jump to search

Installing PLS_Toolbox

PLS_Toolbox is delivered as a Windows Installer (.exe for Windows only) or as a compressed ZIP file (all platforms). Installation is a two-part process of copying and decompressing files into a local folder and then adding those files to your MATLAB path. You will need your License Code (obtained at the time of purchase).

System Requirements

  • MATLAB 6.5 or greater.
  • 20 MB of disk space.

Matlab Compatibility

Although we try to maintain inter-compatibility with Matlab add-on Toolboxes to the greatest extent possible some incompatibilities do exist. Most of these cases originate from situations where PLS_Toolbox has an established function and TMW (The MathWorks) introduces a new function with the same name and is unwilling to rename it. Because we don't have access to the entire TMW product family this list may not include all incompatible functions:

  • Incompatible Functions
    • cluster.m (Stats Toolbox)
    • crossval.m (Stats Toolbox)
    • regression.m (Neural Net Toolbox)
    • ridge.m (Stats Toolbox)
  • Intentional overloads:
    • boxplot.m (Stats Toolbox because of DataSet Object conflict)
  • Compatible Functions
    • dendrogram.m (Stats Toolbox)
    • range.m (Stats Toolbox)
    • kstest.m (Stats Toolbox)


To manage which function is being called, relocate PLS_Toolbox on your Matlab path. Putting PLS_Toolbox at the bottom of the path will allow the use of Matlab functions but will often make PLS_Toolbox inoperable until it's moved back to the top of the path. Within Browse you can select Help>Move Path Folders and indicate the location (Top or Bottom). The evrimovepath function can be used from the command line.

Using the Windows EXE Installer

It is recommended that Windows users use the PLS_Toolbox Windows Installer.

  1. Copy the .exe file to your Desktop (or other suitable local location).
  2. Double click the icon and follow the instructions. Be sure to verify the location of the installation folder; by default it will be the "toolbox" folder of your current MATLAB installation. Window 7 users: may need to choose a folder within your personal Documents folder in order to avoid permissions problems caused by Window 7 security.
  3. When the Installer has completed, it will prompt you to start MATLAB and run an install script.
  4. If MATLAB doesn't start see Step 2 below.

Manually Installing from ZIP File (all platforms)

  1. File Extraction
    Decompress the PLS_Toolbox ZIP file and move it to the MATLAB "toolbox" folder (e.g. C:\Program Files\MATLAB\R2006a\toolbox).
    Note: For those who already have a version of PLS_Toolbox installed, be sure the folders do not have the same name (by default they should not). Rename your existing copy of PLS_Toolbox if necessary.
    Note: Windows 7 users may find that they do not have sufficient system permissions to install in a folder under the C:\Program Files\ folder or that, after installation, they get warnings due to the inability to set program configuration. In these cases, it will be necessary to install PLS_Toolbox into a folder such as your personal Documents folder.
  2. Run evriinstall
    For MATLAB to find PLS_Toolbox, the directories that contain the toolbox files must be added to the MATLAB search path (except the @DATASET folder).
    Note: If you are familiar with installing toolboxes purchased from The MathWorks, this process is slightly different. TMW installers typically are able to set the path automatically.
    Start MATLAB and follow these steps:
    • Start MATLAB and use the Current Directory toolbar to navigate to the PLS_Toolbox directory created in Step One.
    • Once PLS_Toolbox is the current directory type evriinstall at the command line (in the COMMAND WINDOW):
      >> evriinstall
      A dialog box will appear:

      Install.png

      Change the settings as needed then click the Install button. EVRIINSTALL will add PLS_TOOLBOX to the MATLAB path, check for any installation problems, and output results to the command window.
    • See the Reference Topics page for information about PLS_Toolbox functions. Continue reading the Software User Guide for information using Solo and PLS_Toolbox.

Configuring Java

In some older versions of Matlab, including version 7.0.4 (R14SP2), it is necessary to manually add the jar files included with PLS_Toolbox to the Matlab java class path.

  1. Open Matlab.
  2. Navigate to the "..\PLS_Toolbox_65\extensions\javatools" folder (where "PLS_Toolbox_65" is the folder containing the version of PLS_Toolbox you are installing. The "65" will change for each version)
  3. Edit the Java class path file:
    >> edit classpath.txt
  4. At the bottom of the classpath.txt file add lines for each .jar file in the extensions folder using the correct path. In the following example, PLS_Toolbox version 6.5 (named "PLS_Toolbox_65") is installed in the "c:\Toolbox\" folder so the corresponding entries look like:
c:\Toolbox\PLS_Toolbox_65\extensions\javatools\evriJavaClasses.jar
c:\Toolbox\PLS_Toolbox_65\extensions\javatools\rsyntaxtextarea.jar
c:\Toolbox\PLS_Toolbox_65\extensions\javatools\mysql.jar
c:\Toolbox\PLS_Toolbox_65\extensions\javatools\libsvm_jre15.jar
c:\Toolbox\PLS_Toolbox_65\extensions\javatools\derby.jar
c:\Toolbox\PLS_Toolbox_65\extensions\javatools\commons-dbutils-1.3.jar
c:\Toolbox\PLS_Toolbox_65\extensions\javatools\jsoup.jar
  1. Save the file and restart Matlab.
  2. If MIA_Toolbox is installed, any .jar files in the "..\MIA_Toolbox\utilities" folder should be added using the same procedure as above.

Installing Solo

System Requirements

Windows

  • Windows 2000 or newer (Windows XP or later recommended).
  • 500 MB of disk space.
  • 2GB RAM (more RAM may be necessary for large DataSets).

MAC

  • OS X 10.5 or newer.
  • Intel Mac (64 bit).
  • 500 MB of disk space.
  • 2GB RAM (more RAM may be necessary for large DataSets).
  • X11 (X Window System) see here for installation.

Installation

Windows

  1. Copy the .exe file to your Desktop (or other suitable local location).
  2. If installing with a license file (evrilicense.lic) or license server file (evrilm.xml), move these files into the same folder as the .exe installer.
  3. Double click the icon and follow the instructions. Be sure to verify the location of the installation folder; by default it will be in the "EVRI" folder of your program files folder.
  4. Identify if you want desktop and Start menu shortcuts created as well as if you want to associate .mat files with Solo.
  5. Identify if you want to add Solo to the system path. Doing so will assure 3rd-party programs will be able to start up Solo (where the 3rd-party program offers this feature.)
  6. When the Installer has completed, you will be asked if you want to start Solo now. The first time you run Solo, it will need to unpack several libraries and make take a minute or so to complete installation. Subsequent startups will take less time. Note that when installed in the standard Program Files folder, this step needs to be done as a user with administrative privileges so it is best to do from the installer.

NOTE: If installing Solo with a license server, see the Evri_license_server documentation.

Troubleshooting: If having problems downloading, installing or running Solo, see the Solo Troubleshooting page.

Mac

Requires Intel Mac and OS X 10.5 or newer

  1. Image Mounting
    Open the Solo.dmg (or Solo+MIA.dmg) disk image file by double-clicking on it. The disk image will mount on your computer.
  2. Copy Solo APP
    Copy the Solo.app (or Solo+MIA.app) bundle to your Applications folder (or another local folder).
  3. Launching The Applicatioin
    Double-click Solo.app to start the application. NOTE: The first time you run Solo, it will need to unpack several libraries and make take several minutes (or longer depending on system resources) to complete installation. Subsequent startups will take less time.

Troubleshooting: If having problems downloading, installing or running Solo, see the Solo Troubleshooting page.

Linux

  1. Unzip/UnTar
    Copy the tar.gz file into a target folder (create one using the mkdir command) in any convenient location (e.g. /usr/local or any accessible location) and use the following command to unzip/untar the application:
     tar -xzf solo.tar.gz
    (note: archive name will be depend on the exact application you are installing)
  2. Launching The Applicatioin
    In a console window, change the working directory to the target folder and use the command:
     ./solo
    (application name will depend on the exact application you purchased). If starting from a different working directory, use the full path to the solo application.
You can also create a shortcut which performs this operation. NOTE: The first time you run Solo, it will need to unpack several libraries and make take several minutes (or longer depending on system resources) to complete installation. Subsequent startups will take less time.

Troubleshooting: If having problems downloading, installing or running Solo, see the Solo Troubleshooting page.

Troubleshooting Solo

If having problems downloading, installing or running Solo, see the Solo Troubleshooting page.

Updating

The most up-to-date version of PLS_Toolbox and Solo will be made available to users on our web site, www.eigenvector.com. Registered users can download new copies of the toolbox that incorporate the latest bug fixes. A username and password is required for access to the upgrades. Usernames and passwords are supplied when licenses are purchased (older version users should have received this information by registering their software). If you have forgotten your username and/or password, please write to helpdesk@eigenvector.com or use the lost username/password page accessible from the download page on the Eigenvector Research website.

To check for updates select Check for Updates from the Help menu of the Workspace Browser or Analysis interface, or, for PLS_Toolbox users, use the command:

>> evriupdate 

at the Matlab command line. This will bring up a small dialog box indicating the status of your installed products.

Checking Your Installation

Solo or PLS_Toolbox users can get product information from the Help > About menu in the Workspace Browser or Analysis window.

PLS_Toolbox users can use the following commands to verify a new or existing installation. We request the output from these commands when you submit a request to helpdesk@eigenvector.com.

  • ver - Matlab command show a list of installed products.
  • path - Matlab command showing the current folders on the Matlab path. ALL PLS_Toolbox folders (and sub-folders) should be on the Matlab path.
  • evridebug - PLS_Toolbox command that checks for common installation problems.

Also check our FAQ page, there are several more detailed issues including trouble with the Matlab path discussed there.

Getting Help

Start by checking our FAQ Pages. For Solo: also check the Solo Troubleshooting Pages

If you cannot find the information you need there, send problem reports to helpdesk@eigenvector.com

Please send your message with the following information:

  • Product License/Registration Number
  • Computer type, model and operating system version
  • Exact errors you are receiving (if any)
  • Hardware and memory configuration (if reporting a memory issue)
  • Your contact information including either your Eigenvector Username or the following information
First Name
Last Name
Company Name
Phone Country Code
Phone Area Code
Phone Main
Phone Extension