Software

Bayesia announces the release of the version 4.3 of BayesiaLab, the leading software in the market of Bayesian networks and Data Mining, and offers new reduced fares for academic institutions (available from 140 Euros).

The main new features of this release are:

• Data Type: in order to ease the Bayesian networks evaluation by preventing the reloading of the databases after each evaluation, it is possible to automatically tagged learning and test lines in the database;
• Maximum Weight Spanning Tree: this learning algorithm is by far the quickest unsupervised learning algorithm. Even if the resulting network is not optimal, it can be used for a first imputation of the missing values, as the initial network before using Taboo or EQ, and it can also be used for variable clustering when there are many variables;
• Fixed Arcs for the Multiple Clustering: fixed arcs are used to prevent the data clustering algorithm to focus only on the strong (structural) relationships (e.g. the variables that have simultaneous filtered values);
• Stratification: modification of the probability distribution of the target variable with a very weak representation (as fraud for example) in order to be able to learn a network structurally more complex;
• Semi-supervised Clustering: a weight variable editor allows to guide the clustering. A weight greater than 1 will increase the importance of that variable during the clustering. A zero weight will make the variable purely illustrative.
• Global Performance: that evaluation is based on the log-likelihood of the network with respect to each data point. Besides the mean and the standard deviation of the measure, the density function and the repartition function are also available.
• and, as usual, increased ergonomics of the interface and performance improvements.

You can consult the newsletter that describes more in details the new functionalities of BayesiaLab 4.3, download BayesiaLab 4.3 for a new 30 days evaluation period, and you can also have a look at the dynamic and static presentations that illustrate successful applications of BayesiaLab.

What's this?