SPSS Inc. Predictive Analytics Solution Updated
PASW Statistics offers users wide array of open source statistical functions and procedures in R and Python.
- By James E. Powell
- August 24, 2009
SPSS Inc. has launched PASW Statistics 18 (formerly SPSS Statistics), a statistics software suite used by commercial, government, and academic organizations to solve business and research problems.
Part of the SPSS Predictive Analytics Software (PASW) portfolio of products, PASW Statistics 18 combines new and enhanced capabilities to support the entire analytic process -- preparing data, running analysis and creating and delivering reports. The latest version has innovative new features that automate time-consuming operations, new analysis and reporting capabilities that improve results, and expands the ability of users to create their own analytical functionality using open source -- R and Python -- programming languages.
Jason Verlen, chief product strategist for SPSS, said, “PASW Statistics 18 delivers incredibly robust functionality that increases the usability of advanced analytics for all audiences, enabling business users, analysts and statistical programmers to collaborate on a single platform. We are also excited to deliver more flexibility, greater choice and customization with PASW Statistics by welcoming open source users into our Predictive Analytics framework.”
Businesses continue to rely on PASW Statistics for marketing effectiveness; administrative research; human resources and resource planning; medical, scientific, clinical and social science research; planning and forecasting; quality improvement; reporting; and ad hoc decision making. PASW Statistics also remains the choice of academia worldwide that rely on it in classroom instruction as they prepare students for successful professional careers.
New PASW Statistics 18 functionality includes greater speed to results, enhanced analysis and reporting tools, and access to R and Python programming languages.
Greater Speed to Results for Business Users
Automation and analysis features in PASW Statistics 18 enable users of all skill levels to prepare and conduct analysis quickly and accurately and access advanced functionality for more finite control when needed.
PASW Direct Marketing gives marketing professionals a straightforward, easy way to perform analyses to plan and carry out more effective marketing campaigns. A single, intuitive interface guides marketers with six techniques to better understand customers: RFM (recency-frequency-monetary) analysis; cluster analysis; and prospect profiling to identify those with the greatest propensity to purchase; test the effectiveness of direct response packages; and, analyze response rate by ZIP or postal code.
Automated Data Preparation simplifies a previously tedious and repetitive, time-consuming process by automatically detecting and correcting quality errors, and entering missing values. It also provides a report with recommendations to ensure the quality of data used. For example, if large amounts of data are missing from a data set, it will respond with a “Do not use” recommendation.
Enhanced Analysis and Reporting Tools
PASW Statistics 18 provides skilled analysts with even more options to improve the reliability of data sets. For example, PASW Bootstrapping creates more dependable and stable estimates for predictive models to ensure the most reliable results. PASW Bootstrapping estimates the sampling distribution of an estimator by re-sampling with replacement from the original sample.
Nonparametric Testing enables users to make multiple comparisons within non-normal data. Most statistics assume a normal distribution -- bell-shaped curve -- but because not all business problems have data that meet this assumption, special algorithms are required. Non-normal data distributions occur in situations such as insurance claims where the majority of claims are for $0. At times, the incidence of claims is small compared to the number of customers who never make a claim, nonparametric methods are required to correctly assess differences in claim amounts between groups (such as different types of policies).
Interactive Model Viewer provides greater visualization of data and models, allowing users a simple way to explore and compare results, and determine the best algorithm for a particular data set. For instance, it will graphically show results of Automated Data Preparation, or of the non-parametric tests.
Increased Accessibility to Open Source
PASW Statistics 18 now allows access to R and Python programming languages with the introduction of PASW Statistics Developer. With PASW Statistics Developer, any R and Python package can be easily wrapped in PASW Statistics syntax so it takes on the appearance of a standard procedure, which is then easily invoked through the product. It can also be given a dialog box interface and R packages can produce standard tabular output, making both indistinguishable from PASW Statistics.
Available in 10 languages, PASW Statistics Developer is the ideal solution for users who want to employ thousands of statistical functions in R and/or Python that can be accessed and executed by anyone familiar with PASW Statistics. It includes the core functionality found in PASW Statistics Base -- the flagship module containing the most fundamental and essential analysis, such as data access and management capabilities, programmability options, Custom Dialog Builder, report creation and deployment.
Pricing and Availability
PASW Statistics 18 from is available now for Windows, Linux, and Mac platforms on the desktop and Windows, Solaris, Linux, AIX and HP/UX on the server. Pricing for PASW Statistics Base, all other PASW Statistics modules, and PASW Statistics Developer is available at the SPSS Web Store.
SPSS no longer requires PASW Statistics Base for installing and running related modules because capabilities such as data access and management as well as reporting have been added to every module. All PASW Statistics modules can now be installed and run separately, or in conjunction with any other module.
More information is available at http://www.spss.com
James E. Powell is the editorial director of the Business Intelligence Journal and BI This Week newsletter.