BayesiaLab
Mine your Data for extracting Useful Knowledge
If you need an integrated set of tools combining the best learning methods, providing a very easy way for integrating new models to ergonomic decision interface, and reasonably priced, then you need BayesiaLab!
BayesiaLab, the key features for optimal Data Mining
- Pre-process your data quickly and effectively by using the discretization and aggregation tools
- Manage the missing values rigorously with the E-M algorithm
- Discover all the direct probabilistic relations that hold between your variables. You will then get a global understanding of your data and save the time dedicated to the identification of the direct and indirect correlations
- Carry out unsupervised segmentation (typologies) by identifying similar groups of records using Bayesian unsupervised clustering
- Identify new concepts with the powerful variable clustering algorithm, and create the corresponding hidden variables with Bayesian unsupervised clustering. You will then be able the learn probabilistic relationships between these latent variables and the manifest variables (hierarchical models, fully automatic structural equations).
- Use our very effective selection algorithm that will allow you finding the minimal subset of the variables that are really pertinent with respect to your target variable, among thousands of variables
- Quickly set up a probabilistic profile of your target variable
- Gain insight into the obtained models graphically or by using all the power of the complete analysis toolbox
- Test "What-if" scenarii by exploiting the power of Bayesian Inference
- Use your Bayesian networks to carry out your imputation tasks rigorously
- Generate efficient adaptive questionnaires taking into account costs and information gains
- Use the Batch mode to classify records of a data base using an off-line process
- To ease the deployment of your Bayesian scoring functions, export the Markov Blanket of your target variable to external programming languages (SAS, PHP, ...)
Data Mining Application Examples
|
Presentation
Trial Version
