[Skip Global Navigation]

OEMing SPSS Technology

Pyramid - Components

Embeddable Components

Embeddable components are the building blocks for your predictive analytics applications. They are programmable components that do one thing, and do it well. SPSS applications and tools such as SPSS PredictiveMarketing, Clementine, and SPSS, and applications created by OEM partners such as PeopleSoft, Aspen Technology, and Alterian, use SPSS components with demonstrable marketplace success.

The benefits of SPSS components

Much like an audiophile builds a high-end sound system using premium stereo components, enterprise application architects use SPSS components to assemble domain-specific, high-performance predictive analytics applications that are tailored to their customers' needs. SPSS components provide the following important benefits to OEM partners:

The components process

Predictive analytics applications are created either by embedding one or more components into an existing application, or building an applications around components. Data management is the key to success in both cases. The application must feed data into the component. An alternative is to embed predictive analytics that provide higher levels of functionality, such as those in SPSS tools or analytic applications.

SPSS currently offers the following components for OEM use:

Predictive Analytic Components

Predictive analytic components are used for building models that solve classification, segmentation, prediction, time series forecasting, and association problems. Many of these components are used in our analytic tools, SPSS® and Clementine®. As such, the analytical components have been proven in the market to successfully solve new and existing business problems.

Classification Components

Classification components are used to determine what differentiates a group of entities. The resulting models enable classification of entities based on their profiles. For example, classification algorithms can be used to determine the profile of an individual that is likely to respond to a marketing campaign, leave for a competitor, or perpetrate fraud.

SPSS currently offers the following classification components:

Segmentation and Data Reduction Components

Segmentation components are typically used to discover naturally occurring groups. The most common application of discovering naturally occurring groups is in the area of customer segmentation.The goal of data reduction is to take a large number of attributes and discover a more manageable number of attributes that capture the essence of the information represented by the larger set of attributes.

SPSS currently offers the following segmentation and data reduction components:

Prediction Components

Prediction components are often used to predict values based on profiles of entities such as customers. For example, prediction components can be used to model customer spending for the next quarter, given a customer's past purchase behavior and demographic profile.

SPSS currently offers the following prediction components:

Time Series Forecasting Component

The SPSS Time-Series Component is used to predict future values, given a historical set of data. This component is unique in that it enables users without statistical knowledge to leverage multiple forecasting methods. Using SPSS expert modeling technology, the SPSS Time-Series Component selects the best forecasting method for a given problem. This component is often used in demand planning and supply chain management to accurately forecast needs for hundreds of thousands of store/SKU combinations. The SPSS Time-Series Component is able to manage a range of forecasting problems, from simple series to complex issues involving predictors, events, interventions, or intermittent demand. Learn how Peoplesoft/J.D. Edwards were able to bring sophisticated statistical forecasting to the mid-market.

Association Components

Association components are used for discovering events or attributes that tend to occur with one another. The power of the association components lies in their ability to discover associations that aren't intuitive. The classic example of using association components is from the convenience store industry, which discovered an unusually high association between beer and diapers.

SPSS currently offers the following association components:

Scoring Components

The SPSS Scoring Engine enables OEM partners to deploy the value of predictive analytics to their customers, without requiring their customers to build models. For example, a fraud management system can use the SPSS Scoring Engine to score fraud cases from models refined during installation and periodically refreshed by the vendor. The SPSS Scoring Engine enables you to score models created by embeddable components, SPSS tools, and third-party applications that generate PMML representations of models. The SPSS Scoring Engine is a server that supports high volumes of data in both batch and real-time environments.