Eigenvector Software Explained

Eigenvector Research produces a variety of software products for chemometrics and machine learning and we often get asked how they work together. Here’s the roadmap!

We have two main packages for modeling, our MATLAB® based PLS_Toolbox, and our stand alone Solo (with versions for Windows, macOS and Linux). Solo is the compiled version of PLS_Toolbox, so in practice they are nearly identical, the difference being that if you are using PLS_Toolbox under MATLAB you also have access to the command line versions of all the functionality. PLS_Toolbox and Solo are highly interfaced point and click programs and include PCA, PLS, PCR, MCR, ANNs, SVMs, PARAFAC, MPCA, PLS-DA, SIMCA, kNN, etc., a very wide array of preprocessing methods, and also tools for specific tasks such as calibration transfer/instrument standardization.

Eigenvector Software for Chemometrics and Machine Learning

	MATLAB®-based	Stand-alone
General Modeling & Analysis	PLS_Toolbox Our flagship product 30 years in the making. Point and click or command line access to the widest array of chemical data science tools and methods.	Solo The stand-alone version of PLS_Toolbox, available for Windows, macOS and Linux. Point and click chemometrics and machine learning.
Hyperspectral & Multivariate Image Analysis	MIA_Toolbox Add-on to PLS_Toolbox, allows seamless use of modeling methods on hyperspectral data plus additional image analysis tools.	Solo+MIA The stand-alone version of PLS_Toolbox plus MIA_Toolbox. Point and click modeling of hyperspectral images.
Model Export for Online Predictions	Model_Exporter Add-on to PLS_Toolbox, turns models into numerical recipes or code for application in other software or platforms.	Solo+Model_Exporter Solo with Model_Exporter built in.
Online Prediction Engine		Solo_Predictor Full featured online prediction engine applies any PLS_Toolbox or Solo model to new data, MIA_Toolbox compatible.

If you are doing hyperspectral imaging then you can add our MIA_Toolbox to PLS_Toolbox, or choose Solo+MIA. This allows use of all the above methods directly on hyperspectral images plus adds a few more image specific tools. 

If you want to automate model application (say you want to get a PLS model online and have it make predictions as new data comes in) there are two main routes. Solo_Predictor is a full featured stand alone prediction engine that can apply any model you make in PLS_Toolbox/Solo and there are a variety of ways to communicate with it, the most common being socket connections. Solo_Predictor is compatible with hardware from many of our Technology Partner spectrometer companies.

Model_Exporter, on the other hand, creates numerical recipes and code required to apply models to new data streams in a variety of languages including MATLAB and Python. These recipes can then be compiled into other programs or run on hand held devices (such as ThermoFisher’s TruScan or Si-Ware’s NeoSpectra). 

So what software should you buy? PLS_Toolbox or Solo?

Buy PLS_Toolbox if you …
— already have access to MATLAB, it’s less expensive than Solo
— know you want to automate pieces of your modeling process
— want to customize plots using MATLAB
— want to access additional functionality from other MATLAB toolboxes

Buy Solo if you …
— want to work only within visual interfaces
— don’t need to script or program
— prefer the lower cost of Solo compared to MATLAB + PLS_Toolbox

Still have questions? Write to sales@eigenvector.com. Happy modeling!

BMW

MATLAB is a registered trademark of The MathWorks, Inc., Natick, MA.