PASW Text Analytics

PASW Text Analytics* is a text mining workbench that enables you to extract key concepts, sentiments, and relationships from textual or "unstructured" data and convert them to a structured format that can be used to create predictive models. You can access PASW Text Analytics directly from the interface of SPSS Inc.'s industry-leading data mining workbench, PASW Modeler*. The powerful combination of these two products transforms PASW Modeler into a fully integrated data and text mining workbench.

PASW Text Analytics uses a proven natural language processing (NLP) linguistic extraction process. With this software, you can access and process a wide variety of unstructured data, including text contained in:

• Microsoft® Office files
• Call center notes
• Survey text responses
• Customer e-mails
• Web forms
• RSS feeds, such as blogs and news feeds

PASW Text Analytics simplifies the categorization and extraction process of a wide variety of text. After concepts are extracted from text, data mining techniques are applied to them. This has been shown to improve the "lift" or accuracy of predictive data models and significantly improve results.

For example, Swiss cable television company Cablecom has used PASW Text Analytics to improve customer retention. "Our organization has been able to better understand and quantify reasons for customer churn through the use of predictive text mining software from SPSS, an IBM Company. We've uncovered concepts and relationships in text that would be too costly—or even impossible—to detect by any other methods," said Federico.

PASW Text Analytics offers you the ability to create and share text analysis packages (TAPs), which are categories and linguistic resources that control the extraction process. Three pre-built TAPs (English-language only) are included for customer, employee, and product satisfaction analysis—or you can also build your own TAPs.

When you are able to analyze the words people use to express their thoughts and feelings, you can draw more reliable conclusions about their future behavior and use that predictive insight to meet their needs more successfully.

Multi-lingual sentiment analysis helps you capture the attitudes, thoughts, and behaviors of more people across more markets and channels than ever before. The result? You'll be able to build more accurate predictive models that will help maximize customer loyalty, satisfaction, and profitability both locally and globally.

PASW Text Analytics is available for processing Dutch, English, French, German, Italian, Portuguese, or Spanish text, and it also supports Language Weaver™, a provider of software that automates human language translation.

Since PASW Text Analytics integrates with the data mining workbench from SPSS, an IBM Company, you may want to read more about the capabilities of PASW Modeler, and about data mining in general.

* PASW Text Analytics and PASW Modeler, formerly known as Text Mining for Clementine® and Clementine, are part of SPSS Inc.'s Predictive Analytics Software portfolio.

Easy access and customization

PASW Text Analytics has an intuitive, graphical interface that enables analysts and business users to perform text mining without help from a linguist. You can access text data directly and deploy streams and models to decision makers or for use in other applications.

You can customize concept dictionaries for specific domain areas by using the included Resource Editor. This easy-to-use resource for managing extracted text helps you discover relevant concepts and associations more quickly. Additionally, you can use the Resource Editor to edit the dictionaries included with PASW Text Analytics.

Deploy text mining results throughout your organization

PASW Text Analytics' open system works with many types of operational databases and applications, including SPSS Inc.'s predictive applications.

If your organization employs an enterprise feedback management (EFM) solution, PASW Text Analytics can help you understand the opinions, attitudes, and preferences of your customers, employees, citizens, business partners, and others.

Additionally, use PASW Collaboration and Deployment Services to help manage your analytical assets and analytical processes.

Note: PASW Text Analytics is a Windows-only client, but can connect to a remote server running Microsoft Windows, Sun® Solaris™, Linux® Red Hat®, HP-UX®, or IBMAIX®.

• Client
  • Operating system: Microsoft® Windows® Vista® Business and Enterprise (64-bit code running a 32-bit application), or Windows® XP Professional with Service Pack 2 (32- and 64-bit)
  • Hardware:
    • Intel® Pentium®-compatible processor or faster; x64 (AMD 64 and Intel EM64T) processor family (for 64-bit Windows) running at 1GHz or faster
    • Memory: 512 RAM minimum; a minimum of 1GB is recommended if you are working with a large volume of data
    • Minimum free drive space: 1GB on Windows PASW Modeler* client machine, plus additional space for data created
    • Monitor with 1024 x 768 resolution or higher
    • A CD-ROM drive is required if you are installing from a CD-ROM
  • Software: PASW Modeler 13.0. Microsoft Internet Explorer® 6.0 or later to access online help

• Server Windows
  • Operating system: Microsoft Windows Server 2003 Standard and Enterprise (32- and 64-bit) or Windows Server 2003 R2 Standard and Enterprise (32- and 64-bit)
  • Hardware:
    • Pentium or Pentium-class processor or faster for 32-bit Windows; AMD 64 or EM64T for 64-bit Windows. Multiple processor machines are recommended
    • Memory: 4GB RAM or more recommended on PASW Text Analytics server machine
    • Minimum free drive space: 1GB is recommended, plus additional space for databases on PASW Modeler server machine
    • DVD-ROM drive
    • A network adaptor running TCP/IP network protocol
  • Software: The client software must be at the same release level as the PASW Text Analytics Server software

• Server UNIX
  • Operating system: Red Hat® Enterprise Linux® ES4 and 5 (64-bit), Sun™ Solaris™ 9.x and 10 (SPARC® 64-bit), HP-UX® 11iV3 and HP-UX11i23 (64-bit), HP-UX11.23 and HP-UX 11.31 (64-bit), or IBM® AIX® 5L 5.2 and 5.3 (64-bit)
  • Hardware:
    • Pentium or Pentium-class processor or faster for 32-bit LINUX; AMD64 or Intel EM64T for 64-bit LINUX; UltraSPARCII or higher for Solaris™
    • Memory: 4GB RAM or more recommended
    • Minimum free drive space: 1GB is recommended, plus additional space for databases on PASW Modeler server machine
    • DVD-ROM drive
    • A network adaptor running TCP/IP network protocol
  • Software: The client software must be at the same release level as the PASW Text Analytics Server software

PASW Text Analytics

Harness the power of predictive analytics by enrolling in a training program from SPSS Inc. You have three options to choose from:

• OnSite Education (5 days)
• Live e-learning: train online and in real time with an SPSS instructor (i-WBT)
• Live instructor training: group or individual one-on-one sessions available (Live)