Contact Sales: (619) 330-0780

Key Benefits

Key Benefits of Real Time Big Data Processing

Overall

  • Accelerates time-to-market for enhancing intelligent decision making via predictive data
    • Allows for the clear separation of tasks: Model development vs. model deployment
    • Allows predictive models to easily flow from the scientist’s desktop to production where they can be put to work right away
    • Simplifies and accelerates deployment of predictive models within the broader organization
  • Reduces cost and complexity of predictive analytics for big data
    • Utilizes open standards
    • Is compatible with all leading commercial and open-source data mining tools (IBM SPSS, SAS, KNIME, R)
  • Enhances adoption and encourages utilization of predictive analytics for real time big data processing
    • Operates via an intuitive user interface, irrespective of deployment option
    • ADAPA: Web console or Excel add-in tool
    • UPPI: embedded function within leading analytics databases, within Datameer (an end-to-end business intelligence solution) or within Hive (a data warehouse system for Hadoop)
    • Allows analytical results to be utilized either within enterprise applications or to be exported to Excel
  • Ensures scalability and reliability of model execution via uniform deployment platform
    • Automates real-time scoring and big data processing
    • Offers enterprise-grade performance and stability by utilizing application servers or data warehouse appliance platforms
    • Utilizes multi-threaded execution, multi-server load balancing or massively parallel execution to deliver extreme scalability
  • Supports highly dynamic and complex data environments
    • Allows analytics teams to utilize multiple models in PMML concurrently
    • Offers conditional logic and extensive pre and post-processing capabilities
    • Enables organizations to create solutions that are tailored to different customers (in case of people data) or sensor types (in case of machine data).
    • Facilitates implementation of different decision strategies based on different data attributes (e.g. score, customer location)

Data Scientists

  • Enables the exchange of predictive models and solutions between different applications and vendors
    • Utilizes the PMML (Predictive Model Markup Language) industry standard file format
    • PMML enables the import, deployment and execution of predictive models
  • Consumes model files conforming to multiple versions of PMML
    • Supports PMML versions 2.0 through 4.2
    • Automatically converts older versions of PMML into a 4.2-compliant format
  • Supports all of the leading commercial and open-source data mining tools
    • Supported tools include: IBM SPSS, SAS, KNIME, R
    • A model built on any of the supported tools can be automatically exported into a PMML file
  • Supports a wide range of predictive modeling techniques
    • Association Rules
    • Decision Trees for classification and regression
    • Neural Network Models: Back-Propagation, Radial-Basis Function, and Neural-Gas
    • Support Vector Machines for regression, binary and multi-class classification
    • Linear and Logistic Regression (binary and multinomial)
    • Naive Bayes Classifiers
    • General and Generalized Linear Models
    • Cox Regression Models
    • Rule Set Models (flat decision trees)
    • Restricted Boltzmann Machines
    • Clustering Models: Distribution-Based, Center-Based and 2-Step Clustering
    • Scorecards (including support for reason codes and point allocation for categorical, continuous and complex attributes)
    • Multiple Models: Model Composition, Segmentation, Chaining, Cascade and Ensemble, including Random Forest Models and Boosted Trees
  • Facilitates data management and audit activities
    • Defines a data dictionary
    • Includes rules for handling missing, outlier and invalid values
    • Implements multiple functions for data pre- and post-processing
  • Simplifies and accelerates deployment of predictive models within the broader organization
    • Reduces error rate and rework between analytics team and IT department
    • Facilitates IT department’s ability to rapidly operationalize predictive analytics capabilities for business users
    • Supports deployment of predictive models that the analytics team develops individually or in collaboration with the experienced Zementis team of data scientists
  • Helps organizations more readily incorporate predictive analytics into ongoing business operations
    • Offers a simple, intuitive and dynamic user interface
      • ADAPA Web-based console serves as a centralized point from which users can utilize predictive models and solutions
      • ADAPA Web services API allows user to score data either by single record or across multiple records (multiple-record processing support enables batch-mode or on-demand scoring)
    • Includes a model validation test, allowing the user to verify that the model has been deployed correctly
      • The engine executes 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 facilitating debugging of model deployment issues
      • Supports the Model Verification PMML element, complementing and enhancing the Zementis model validation functionality
    • Consumes and returns CSV files to simplify the scoring process
      • Allows users to upload a (compressed) CSV data file and run batch scoring against any deployed model
      • Returns results in the same format
      • Allows user to download results for further processing and visualization
  • Extends organization’s predictive analytics capacity and insight via multiple models
    • Supports analysis segmentation across multiple models in PMML, allowing different models to be executed depending on input condition (e.g. device type)
    • Allows multiple models to work together as a single predictive solution to generate an overall score
    • Enhances organization’s ability to generate predictive results in extremely dynamic, complex data environments
  • Systematically improves data quality
    • Logs missing or invalid data elements submitted to the engine during model execution
    • Allows analytics team to assess data quality and measure its potential impact on deployed models

IT Professionals

  • Simplifies and accelerates deployment of predictive models within the broader organization
    • Eliminates the need for custom code and proprietary model deployment solutions
    • Accelerates learning for DBA staff via simple SQL-based interface
    • Reduces error rate and rework between analytics team and IT department
    • Facilitates IT department’s ability to rapidly operationalize predictive analytics capabilities for business users
  • Efficiently utilizes IT architecture, optimizing computational performance, capital expense and operating expense
    • Facilitates massively parallel in-database execution
    • Eliminates the need for data movement
    • Leverages investment in existing powerful database platforms and appliances
  • Enables a vendor-neutral, cross-platform deployment of predictive analytics capabilities
    • Cloud: Amazon Web Services, FICO Analytic Cloud
    • On-site: Oracle WebLogic, IBM WebSphere, Red Hat, JBoss
    • In-database: Pivotal / EMC Greenplum, SAP Sybase IQ, IBM PureData / Netezza, Teradata, Aster
    • For Hadoop: Hive and Datameer
  • Integrates seamlessly with existing enterprise software and business processes
    • Easily deployable in most commercial or open source application servers (ADAPA is a Java Enterprise Edition application)
    • Intuitive user interface via simple Web console
    • Scoring capabilities automatically available as Web services
  • Provides technical and economic benefits of cloud architecture
    • Available versions:
      • Cloud version of ADAPA available as an alternative to on-site deployment
      • Public cloud options: Amazon Elastic Compute Cloud (EC2), FICO Analytic Cloud
      • Available as a private, secure cloud instance with a dedicated ADAPA engine
    • Key benefits:
      • Offers on-demand, scalable capacity that aligns cost with computing demand
      • Provides a cost-effective expansion path via multiple instances on virtual servers
      • Enables failover capability

Business Leaders

  • Extends predictive analytics capabilities throughout the organization
    • Allows data scoring in Excel via add-in tool
    • Enables employees to generate predictive analytics outputs instantly from desktops, laptops or tablet devices
  • Increases speed, accuracy and consistency of business decisions
    • PMML post-processing instantly transforms predictive scores into business decisions
    • PMML-based rules allow users to pair different score ranges with specific business decisions
  • Scales predictive analytics capabilities with the growth of the business
    • Scores thousands of data records per second
    • Delivers instant predictive results to enable real-time decision making at scale
  • Enhances accountability for data-driven decisions
    • Logs both data and decisions to support tracing and individual accountability
    • Enhances organization’s ability to comply with internal standards and external regulatory requirements

 

For more details or if you have any questions, please feel free to contact us!

Copyright © 2017 Zementis / Software AG. All Rights Reserved.