ADAPA® Predictive Analytics Edition

Category Intelligent Software>Data Mining Systems/Tools and Intelligent Software>Neural Network Systems/Tools

Abstract The ADAPA Predictive Analytics Edition is a highly scalable decision engine that offers at its core batch and real-time scoring of predictive models that follow the Predictive Model Markup Language (PMML) standard.

It enables you to manage, deploy, and execute predictive models though a web-based user interface.

By using web service calls, models deployed remotely can be seamlessly used as an integral part of all your on-site decision strategies.

The ADAPA decision engine executes a multitude of pattern recognition models. It uses PMML to upload and execute these models. PMML, a vendor-driven, open standard for predictive models, is an XML- based language which provides a way for applications to define statistical and data mining models and to share these between PMML compliant applications.

Broad Support for Predictive Algorithms -- ADAPA supports an extensive collection of PMML elements, including the following predictive techniques:

1) Back-Propagation Neural Networks;

2) Radial-Basis Function Networks;

3) Neural-Gas Networks;

4) Support Vector Machines;

5) Binary Logistic Regression;

6) Multinomial Logistic Regression;

7) Ordinal Multinomial Regression;

8) Stepwise Polynomial Regression;

9) General Linear Models;

10) Generalized Linear Models; and

11) Decision Trees.

In addition, ADAPA supports several PMML elements which deal with pre- and post processing of input and predicted variables, respectively.

These elements include data dictionary, transformations, targets, functions, and more. ADAPA consumes model files that conform to PMML version 3.2.

If you exported your models from another statistical analysis tool, a PMML converter application is available the manufacturer's Web site to transform your models from previous PMML versions into a 3.2 compliant format.

Web-Based Model Management and Interactive Scoring --

Models are deployed through an intuitive, Web-based management console. It allows the user to manage the deployment of decision models in a central ADAPA repository.

The management console also includes a model validation test, allowing the user to verify that the model has been deployed correctly.

By executing a test file containing input data and expected results for a model, the engine reports deviations from expected results, greatly enhancing traceability of errors and debugging of model deployment issues.

The Web interface console also provides functionality to upload a data file and batch-score it against any of the deployed models.

Results are returned in Comma Separated Value (CSV) format and can be downloaded for further processing and visualization.

Simplified Integration via SOA -- Service Oriented Architecture (SOA) principles simplify integration with existing IT infrastructure.

To facilitate loosely-coupled integration with your in-house applications, ADAPA implements the Java Data Mining (JDM) specification for its Application Programming Interface (API), both at the Java and the Web Services level.

You can score data records against any of the models deployed in the ADAPA engine from within your environment by calling the ADAPA web service.

On-Demand Predictive Analytics Solution -- The ADAPA Predictive Analytics Edition is a fully hosted Software-as-a-Service (SaaS) solution.

You only pay for the service and the capacity that you really need, eliminating the necessity for expensive software licenses and in-house hardware resources.

At any given time, you may launch one or more instances using the ADAPA Control Center web interface. For each instance you can select the most appropriate instance type, “small”, “large”, or “extra large”.

Private Instance for Your Predictive Models -- Implementing single tenant architecture, the manufacturer's service is implemented as a private, dedicated instance of the ADAPA® Predictive Analytics Edition that is dedicated to your models.

Only you have access to your ADAPA engine instance(s) via HTTPS.

Your models and your data never share the same ADAPA engine with other clients. You launch and terminate your instances through the secure ADAPA Control Center web interface.

Trusted, Secure, Scalable Infrastructure -- Zementis leverages Amazon Web Services to provide the on-demand infrastructure for ADAPA.

The Amazon Elastic Compute Cloud (EC2) offers utility computing with virtually unlimited scalability. Billing and subscription management are handled through Amazon Payment Services.

System Requirements

Contact manufacturer.

Manufacturer

Manufacturer Web Site ADAPA

Price Contact manufacturer.

G6G Abstract Number 20288

G6G Manufacturer Number 103998