Molegro Data Modeller
Modelling, Mining and Visualization

• Regression and Classification (MLR, PLS, KNN, Support Vector Machines, and Neural Networks.)
• 2D Molecule Depiction (SDF and SMILES support).
• Principal Component Analysis, Clustering, and Outlier Detection.
• Feature selection and automated parameter tuning.
• Histogram, 2D, and 3D plots. High-dimensional data visualization using Spring-Mass Maps.

Molegro Data Modeller is a cross-platform application for Data Mining, Data Modelling, and Data Visualization.

The highly interactive user interface is ideal for fast and intuitive data exploration, as opposed to complex workflow based solutions or command-driven statistical products.

January 16th, 2009
Molegro Data Modeller Version 2.0.0 Released
This major update introduced several new features.

New features in version 2.0.0:

• Classification models:
  MDM now supports the following multi-class classification models: K-Nearest Neighbor and Support Vector Machines with Linear, Radial, Sigmoid, and Polynomial kernel functions. The classification models support automatic optimization of model parameters and feature selection.



• Chemistry module:
  MDM is now able to show 2D depictions of molecules. Molecules may be imported in SDF format or as SMILES strings. Molecules may be depicted inside the spreadsheet cells or in a separate grid window. It is also possible to visualize the molecules directly in the 2D plotter and optionally cluster overlapping molecules.



• A new Bivariate Analysis dialog for analyzing classification results (confusion matrix, precision / recall, F-measure, and several other measures).



• A new threshold-based clustering method.



• Minor user interface improvements and bug fixes

Features at a glance:

• Regression: Multiple Linear Regression, Partial Least Squares, Support Vector Machines, and Neural Networks.



• Classification: K-Nearest-Neighbors and Support Vector Machines.



• Chemistry: 2D depictions in spreadsheets and plots. SDF and SMILES support.



• Feature selection and cross-validation is simple to set up and use (using built-in wizards).
  


• Automated fine-tuning of regression model parameters (using grid-based search).



• Principal Component Analysis (PCA).
  


• Visualization: Histograms, 2D scatter plots, 3D plots and Spring-Mass Maps.
  


• Clustering: K-means clustering and density-based clustering.
  


• Built-in algebraic data transformation tool.
  


• Outlier Detection.
  


• Similarity Browser.



• Correlation Matrix dialog including descriptor pruning.



• Sophisticated subset creation: create diverse subsets by sampling from n-dimensional grids.
  


• Cross-platform:
  Windows: Vista, XP, and 2000.
  Linux: Most major Linux distributions.
  Macintosh: Mac OS X 10.4.0 or later (Intel and PowerPC) is supported.

More information about Molegro Data Modeller features...