Why Solo?
Aug 30, 2023
I was asked the other day to provide a list of advantages that our Solo software has over its competitors for chemometrics and machine learning. Well, I don’t spend much time keeping track of what’s in other companies’ software. But I can tell you what is in Solo and why we think it’s a great value. (Apologies in advance for all the acronyms but I’ve included a guide below.)
Solo supports a very wide array of methods for data exploration, regression and classification. The standard PCA, PCR, PLS, MLR and CLS are included of course, but also MCR, SIMPLISMA/Purity, Robust PCA and PLS, O-PLS, PLS-DA, Gray CLS, PARAFAC, PARAFAC2, MPCA, batch data digester, batch maturity, LWR, ANN, Deep Learning ANNs, SVM, N-PLS, XGBoost, KNN, Logistic Regression, user specified Hierarchical Models, Automated Hierarchical Models, SIMCA, UMAP and t-SNE.
Solo has a sophisticated point and click user interface (see below) for graphical data editing and model building which users find very intuitive. Plots are easily customizable with class sets, color by properties, etc.
Solo supports the most extensive set of data preprocessing methods: centering, scaling, smoothing, derivatives, automatic WLS baseline, selected points baseline, Whitaker filter, EEM filtering, MSC, EMSC, detrend, EPO, GLSW, despiking, Gap-Segment derivatives, OSC, normalization, PQN, SNV, block scaling, class centering, Pareto and Poisson plus build your own math functions. User selectable preprocessing order and looping for more flexibility.
Solo does full cross-validation including all preprocessing steps, and includes a variety of cross validation methods as well as customizable specified data splits.
Solo supports a host of methods for variable selection: i-PLS, VIP, SR, Genetic Algorithms, r-PLS and stepwise.
Solo includes a full suite of methods for calibration transfer: DS, PDS, DW-PDS, SST, OSC, plus they can be inserted before or between preprocessing steps as required with the Model Centric Calibration Transfer Tool (MCCT).
Solo includes the Model Optimizer which can be used to create, calculate, compare and rank linear and non-linear models with a variety of preprocessing and selected variables
Solo is available on all three major platforms: Windows, MacOS and Linux. Models and data files are completely compatible between platforms.
Solo includes the Report Writer which makes models easy to document by creating PowerPoint or web pages from models. Solo maintains data history and includes model caching for preserving traceability.
Solo optional add-ons include Model_Exporter which allows users to export models as numerical recipes as well as Python and MATLAB code so they can be applied online and in handheld devices. Solo also works with Solo_Predictor, a stand-alone, configurable, prediction engine for online use. The MIA_Toolbox add-on allows users to seamlessly apply all the methods above to hyperspectral images.
Solo is completely compatible with PLS_Toolbox for use with MATLAB. PLS_Toolbox has all the features of Solo, including the point and click interfaces and graphical data editing, but also allows users to access all the functionality from the command line and incorporate these methods in user specified scripts and functions for ultimate flexibility. This allows users to work the way they want (command line or point and click) and still work together.
Solo has the widest array of training options available including our “EVRI-thing You Need to Know About” webinar series, Eigenvector University live classes, and EigenU Recorded courses as well as courses at conferences such as APACT, SCIX and EAS. Eigenvector teaches over 20 specific short course modules.
Solo gives users access to the Eigenvector HELPDESK, user support that is prompt and actually helpful. HELPDESK is manned by our developer staff, the people that actually write the software. Need more help on specific applications? Eigenvector offers consulting services.
Despite all of its advantages, Solo costs less than other major chemometric packages. We publish our price list so you can compare. We offer single-user and floating licenses that work great for small or large groups.
Finally, Solo is and always has been a product of Eigenvector Research, Inc, owned and operated by the same people for 29 years now.
Still have questions? You can try Solo yourself with our free demos; start by creating an account. Or you can always e-mail me, bmw@eigenvector.com.
Best regards,
BMW
- PCA == Principal Components Analysis
- PCR == Principal Components Regression
- PLS == Partial Least Squares Regression
- MLR == Multiple Linear Regression
- CLS == Classical Least Squares
- MCR == Multivariate Curve Resolution
- SIMPLISMA == SIMPLe to use Interactive Self-modeling Mixture Analysis
- Purity == Self-modeling mixture analysis via Pure Variables
- O-PLS == Orthogonal PLS
- PLS-DA == PLS Discriminant Analysis
- Gray CLS == CLS incorporating EPO and GLS filters
- PARAFAC == Parallel Factor Analysis
- PARAFAC2 == PARAFAC for uneven and shifted arrays
- MPCA == Multi-way PCA
- LWR == Locally Weighted Regression
- ANN == Artificial Neural Network
- SVM == Support Vector Machine
- n-PLS == PLS for n-way arrays
- XGBoost == Boosted classification and regression trees
- kNN == k-Nearest Neighbors classification
- SIMCA == Soft Independent Modeling of Class Analogy
- UMAP = Uniform Manifold Approximation and Projection
- t-SNE == t-distributed Stochastic Neighbor Embedding
- WLS == Weighted Least Squares
- EEM == Excitation Emission Matrix
- MSC == Multiplicative Scatter Correction
- EMSC == Extended MSC
- EPO == External Parameter Orthogonalization
- GLSW == Generalized Least Squares Weighting
- OSC == Orthogonal Signal Correction
- PQN == Probabilistic Quotient Normalization
- SNV == Standard Normal Variate
- i-PLS == interval PLS
- VIP == Variable Influence on Prediction
- SR == Selectivity Ratio
- r-PLS == recursive PLS
- DS == Direction Standardization
- PDS == Piecewise Direct Standardization
- DW-PDS == Double Window PDS
- SST == Subspace Standardization Transform