Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Achieve Better Insight and Prediction with Data Mining


Published on

  • Be the first to comment

Achieve Better Insight and Prediction with Data Mining

  1. 1. Clementine® 12.0 – Specifications Achieve Better Insight and Prediction with Data Mining Data mining provides organizations with a clearer view Clementine is popular worldwide with data miners of current conditions and deeper insight into future events. and business users alike because it enables you to: With Clementine from SPSS Inc., your organization can n Easily access, prepare, and integrate structured conduct data mining that incorporates many types of data, data and also text, Web, and survey data resulting in more in-depth knowledge of every aspect of n Rapidly build and validate models, using the most your operations—including more complete analysis and advanced statistical and machine-learning techniques understanding of your customers and constituents. available n Efficiently deploy insight and predictive models on Clementine enables your organization to strengthen a scheduled basis or in real time to the people that performance in a number of areas. For example, you could: make decisions and recommendations, and the n Improve customer acquisition and retention systems that support them n Increase customer lifetime value n Detect and minimize risk and fraud Clementine’s many unique capabilities make it an ideal n Reduce cycle time while maintaining quality in choice for addressing the business applications found product development in today’s data-rich organizations. n Support scientific research Use the predictive insight gained through Clementine to guide customer interactions in real time and share that insight throughout your organization. With Clementine’s powerful data preparation, visualization, and predictive modeling capabilities, you can solve business problems faster.
  2. 2. Streamline the data mining process Choose from an unparalleled breadth of techniques Clementine’s intuitive graphical interface enables analysts Clementine offers a broad range of data mining techniques to visualize every step of the data mining process as that are designed to meet the needs of every data mining part of a “stream.” By interacting with streams, analysts application. You can choose from a number of algorithms and business users can collaborate in adding business for prediction, clustering, and association, including knowledge to the data mining process. Because data survival analysis, neural networks, support vector machines, miners can focus on knowledge discovery rather than and graphical probabilistic modeling. on technical tasks like writing code, they can pursue “train-of-thought” analysis, explore the data more deeply, Optimize your current information technologies and uncover additional hidden relationships. Clementine is an open, standards-based solution. It integrates with your organization’s existing information From this visual interface, you can easily access and systems, both when accessing data and when deploying integrate data from textual sources, Web logs, SPSS’ results. You don’t need to move data into and out of a Dimensions survey research products, as well as data in ™ proprietary format. This helps you conserve resources, virtually any type of database, spreadsheet, or flat file— deliver results faster, and reduce infrastructure costs. including SPSS, SAS , and Microsoft Excel files. No ® ® ® other data mining solution offers this versatility. Business Data Understanding Understanding Clementine’s powerful automation tools make it easy for data miners to find the best model, based on hidden Data Preparation patterns in their data, and quickly produce consistent Deployment and accurate results. Modeling Data Leverage all your data for improved models Only with Clementine can you directly and easily access Evaluation text, Web, and survey data, and integrate these additional types of data in your predictive models. SPSS customers have found that using additional types of data increases The CRISP-DM process, as shown in this diagram, enables data miners to implement efficient data mining projects that yield the “lift” or accuracy of predictive models, leading to more measurable business results. useful recommendations and improved outcomes. Follow a proven, repeatable process With the fully integrated Text Mining for Clementine ® During every phase of the data mining process, Clementine module, you can extract concepts and opinions from any supports the de facto industry standard, the CRoss- type of text—such as internal reports, call center notes, Industry Standard Process for Data Mining (CRISP-DM). customer e-mails, media or journal articles, blogs, and more. This means your company can focus on solving business And, with Web Mining for Clementine , you can discover ® problems through data mining, rather than on reinventing patterns in the behavior of visitors to your Web site. a new process for every project. Individual Clementine projects can be efficiently organized using the CRISP-DM Direct access to survey data in Dimensions products project manager. enables you to include demographic, attitudinal, and behavioral information in your models—rounding out your understanding of the people or organizations you serve.
  3. 3. Add enterprise-level capabilities Unified customer analysis techniques Clementine can efficiently analyze the amounts of data Gain a complete customer analysis toolkit that unifies typically generated by small to mid-sized organizations. both advanced and traditional techniques in a single If your data mining needs grow in volume or complexity, framework. SPSS makes it easy for you to move to our enterprise- n Identify your best customers using the RFM customer level offering, Clementine Server. value segmentation scoring technique n Accurately estimate customer attrition using Cox Using a client/server architecture, Clementine Server regression for survival analysis enables multiple data analysts to work simultaneously n Easily combine these traditional techniques with without straining computing resources. You can take advanced data mining prediction and segmentation advantage of in-database data mining on leading algorithms information platforms and efficiently process large amounts of data. Clementine Server also offers additional Automated modeling deployment options—helping you to extend the benefits Enhance productivity and achieve faster time-to-solution of data mining across geographic or functional lines and using Clementine’s powerful automated modeling. put results in the hands of decision makers. n Find the best models for predicting both binary (“yes/ no”) and numeric outcomes using automated modeling You can further optimize analytical assets throughout operations that create and evaluate many different your organization by using Clementine with SPSS Predictive models in a single step Enterprise Services , which enables you to centralize the ™ n Select the best model more easily with visual evaluation storage and management of data mining models and all information associated processes. With this platform, you can control n Gain more detailed control over automated modeling the versioning of your predictive models, audit who uses using new frequency, weights, and misclassification and modifies them, provide full user authentication, costs features automate the process of updating your models, and schedule model execution. As a result, your predictive Enhanced analytics models become real business assets and your Leverage new analytical techniques to improve predictive organization gains the highest possible return on your accuracy and achieve better results. data mining investment. n Obtain more accurate predictions by combining two or more models with the new Ensemble node What’s new in Clementine 12.0 n Employ advanced sampling techniques, including With this release, SPSS continues its commitment to stratified sampling to obtain more representative delivering a data mining solution that offers the greatest results, and clustered sampling to ensure samples possible efficiency and flexibility in the development include all related items and deployment of predictive models. n Compare scores easily across different algorithms and technologies using propensity scores Clementine features extensive enhancements in several key areas to help your organization improve performance, productivity, and return on your data mining investment.
  4. 4. Wider range of algorithms The cornerstone of the Predictive Enterprise Solve more business problems and predict outcomes Clementine makes data predictive and facilitates the with greater accuracy using several new algorithms. delivery of predictive insight to the people in your n Make more accurate predictions when working with organization who make decisions and the systems wide data sets using the new Support Vector Machine that support daily customer interactions. (SVM) node n Obtain richer insight from graphical Bayesian network If your organization has a great deal of data stored in text models form or in Web logs, you can gain additional value from this data by using Text Mining for Clementine or Web Mining Improved visualization and reporting for Clementine. And you can understand the attitudes and Use Clementine’s new reporting and visualization beliefs that lie behind behavior—why customers make the capabilities to achieve better insight and more effective choices they do—by incorporating survey data from any of communication. Produce the right graph or table type, SPSS’ Dimensions survey research products. export analysis to third-party tools, and communicate results clearly across the organization. Thanks to its integration with SPSS predictive applications n Create more compelling graphs using a wizard-like and other information systems, Clementine enables you interface that guides you to the most appropriate chart to guide daily decisions and recommendations, as well types for your data as long-range planning, with insight into current and n Interact with graphs via rich data selection tools, future conditions. You can accomplish this securely resulting in more accurately focused analysis and efficiently, across your entire enterprise, with SPSS n Improve your reporting by including easy-to-create Predictive Enterprise Services. custom tables. Nest, stack, or layer variables in multiple dimensions to display summaries for multiple statistics Clementine’s extensive capabilities are supported by the and display multiple response sets (requires SPSS most advanced statistical and machine learning techniques Tables 16.0). ™ available. For greater value, it is open to operation and n Discover the best models and get better predictive integration with your current information systems. insight with variable importance charts, which rank variables according to their relative importance Improved scalability and integration “We are extremely impressed with the customer-centric With Clementine 12.0, you gain both desktop tools and response from SPSS in advancing their data mining an upgrade path to a highly scalable, highly integrated, product to help organizations, such as EarthLink, better meet their customer marketing needs.” and easily managed data mining and predictive analytics platform for the enterprise. – Atique Shah n Leverage extended database optimization for improved Vice President of Direct Marketing and Consumer Insights performance EarthLink, Inc. n Gain enhanced real-time scoring capabilities n Achieve a deeper level of integration with custom or third-party components
  5. 5. Features nSpecify worksheets and data ranges – Access data management and Clementine’s main features are described when accessing data in Excel transformations performed in below in terms of the CRISP-DM process. – Unstructured (textual) data SPSS directly from Clementine n Automatically extract concepts from n RFM scoring: aggregate customer Business understanding any type of text by using Text Mining transactions to provide Recency, Frequency, Clementine’s visual interface makes it easy for Clementine* and Monetary value scores and combine for your organization to apply business – Web site data these to produce a complete RFM analysis. knowledge to data mining projects. In n Automatically extract Web site events addition, optional business-specific from Web logs using Web Mining for Modeling Clementine Application Templates (CATs) Clementine* Employ a wide range of data mining algorithms are available to help you get results faster. – Survey data with many advanced features to get the best CATs ship with sample data so that you n Directly access data stored in the use possible results from your data. can easily see the details of best-practice Dimensions Data Model™ or in the n Use interactive model and equation techniques. data files of Dimensions* products browsers and view advanced statistical n CRM CAT* – Data export output n Telco CAT* n Export data to delimited text files, n Show relative impact of different data n Fraud CAT* Microsoft Excel, SPSS, SAS 6, 7, 8, attributes on predicted outcomes with n Microarray CAT* and 9 files, and many databases variable importance graphs n Web Mining CAT* (requires the purchase n Export in XLS format through the n Combine models through meta-modeling of Web Mining for Clementine) Excel Output Node – Multiple models can be combined, or n Export data to Dimensions* for survey one model can be used to analyze a Data understanding applications second model n Obtain a comprehensive first look at your n Choose from various data-cleaning options – Ensemble node combines predictions data using Clementine’s data audit node – Remove or replace invalid data automatically for improved accuracy n View data quickly through graphs, summary – Use predictive modeling to n Import PMML models from other tools such statistics, or an assessment of data quality automatically impute missing values as AnswerTree® and SPSS n Create basic graph types, such as histograms, – Automatically generate operations for n Use the Clementine Extension Framework distributions, line plots, and point plots the detection and treatment of outliers (CLEF)* for custom algorithms n Create a wide range of basic and advanced and extremes graphs with automatic assistance using the n Manipulate data Clementine’s data mining algorithms are Graphboard node – Work with complete record and field organized into a “base” module and optional n Create complex tabular reports easily using operations, including: additional algorithm modules. The base the Custom Table node (requires SPSS n Field filtering, naming, derivation, module includes: Tables 16.0) binning, re-categorization, value n C&RT, CHAID & QUEST—Decision tree n Edit your graphs to communicate results replacement, and field reordering algorithms including interactive tree more clearly n Record selection, sampling (including building n Use visual link analysis to see the clustered and stratified sampling), n Decision List—Interactive rule-building associations in your data merging (including inner joins, full algorithm enables you to incorporate n Interact with data by selecting regions or outer joins, partial outer joins, and business knowledge into a predictive model items on a graph and view the selected anti-joins), and concatenation; n K-means—Clustering information; or use it in a later stage of sorting, aggregation, and balancing n GRI—Generalized rule induction association your analysis n Data restructuring, including discovery algorithm n Access SPSS statistics, graphs, and transposition n Factor/PCA—Data reduction using factor reporting tools directly from Clementine n Binning numerical attributes into analysis and principal component analysis sub-ranges optimized for prediction n Linear Regression—Best-fit linear equation Data preparation n Extensive string functions: string modeling n Access a wide range of data creation, substitution, search and – Structured (tabular) data matching, whitespace removal, The Clementine Classification Module* n ODBC-compliant data sources and truncation includes: with the SPSS Data Access Pack. n Preparing data for time-series analysis n Binary classifier and numeric predictor— Drivers in this middleware pack – Partition data into training, test, and Automate the creation and evaluation of support IBM DB2®, Oracle®, Microsoft validation datasets multiple models SQL Server™, Informix®, and Sybase® – Transform data automatically for n Self-learning response model—Bayesian databases. multiple variables model with incremental learning n Import delimited and fixed-width text n Visualization of standard n Time-series—Generate and automatically files, any SPSS file, and SAS 6, 7, 8, transformations select time-series forecasting models and 9 files Features subject to change based on final product release. Symbol indicates a new feature. * Separately priced modules
  6. 6. n C5.0 decision tree and rule set algorithm Deployment Analysis Services) and leverage high- n Neural Networks—Multi-layer perceptrons Clementine offers a choice of deployment performance database implementations with back-propagation learning, and radial capabilities to meet your organization’s needs. – Vendor-supplied algorithms included basis function networks n Export models using SQL or PMML (the for IBM DB2 DWE, Oracle Data Mining, n Support Vector Machines—Advanced XML-based standard format for predictive and Microsoft SQL Server 2005 algorithm with accurate performance for models) n Leverage high-performance hardware, wide datasets n Clementine Solution Publisher Runtime™ experience quicker time-to solution, and n Bayesian Networks—graphical probabilistic (optional*) achieve greater ROI through parallel models – Automate the export of all operations, execution of streams and multiple models n Cox regression—calculate likely time to including data access, data manipulation, n Transmit sensitive data securely between an event text mining, model scoring—including Clementine Client and Clementine Server n Binomial and multinomial logistic combinations of models—and post- through secure sockets layer (SSL) encryption regression processing n Discriminant analysis – Use a runtime environment for executing SPSS Predictive Enterprise Services (optional*) n General linear models (GLM) image files on target platforms SPSS Predictive Enterprise Services is an n Automatically export Clementine streams enterprise-level platform that enables you The Clementine Segmentation Module* to SPSS predictive analytics applications to manage and automate your analytical includes: – Combine exported Clementine streams processes and easily deploy results across n Kohonen Network—Clustering neural network and predictive models with business your organization to increase productivity n TwoStep Clustering—Select the right rules and exclusions to optimize and increase the value of your analytical number of clusters automatically customer interactions investment. SPSS Predictive Enterprise Base ™ n Anomaly Detection—Detect unusual records n Cleo (optional*) Services helps you to: through the use of a cluster-based algorithm – Implement a Web-based solution for n Centralize and manage analytical assets rapid model deployment to leverage organizational knowledge and The Clementine Association Module* includes: – Enable multiple users to simultaneously provide powerful change management n Apriori—Popular association discovery access and immediately score single capabilities to help with auditability and algorithm with advanced evaluation records, multiple records, or an entire compliance functions database, through a customizable n Automate your analytical processes to n CARMA—Association algorithm which browser-based interface increase productivity and ensure reliable, supports multiple consequents n Scripting consistent, accurate results n Sequence—Sequential association – Use scripts to automate complex n Deploy analytical results by delivering algorithm for order-sensitive analyses repetitive tasks output through customizable, browser- based end-user interfaces or integrating Evaluation Clementine Server (optional*) with your existing applications using n Easily evaluate models using lift, gains, Clementine Server, SPSS’ enterprise-level standardized Web services interfaces profit, and response graphs offering, provides all of the data mining – Use a one-step process that shortens capabilities of the client version plus SPSS Predictive Enterprise Services project time when evaluating multiple increased performance and other Clementine Client Adapter enables analysts models functionality. For a complete list of to interact directly with the Base services – Define hit conditions and scoring features, refer to the Clementine Server to store, retrieve, browse, and search for expressions to interpret model specifications sheet. Key capabilities enable analytical assets. SPSS Predictive Enterprise performance you to: Services Clementine Server Adapter enables n Use propensity scores for consistent n Employ in-database mining to leverage the Process Manager to control Clementine deployment and easy comparison between high-performance database implementations tasks. For more information about how your diverse model types n Use in-database modeling to build models organization can benefit by using Clementine in the database using leading database Server with SPSS Predictive Enterprise technologies (IBM® DB2® Enterprise Edition Services, see the SPSS Predictive Enterprise 8.2, Oracle 10g, and Microsoft SQL Server Services brochure. Features subject to change based on final product release. Symbol indicates a new feature. * Separately priced modules To learn more, please visit For SPSS office locations and telephone numbers, go to SPSS is a registered trademark and the other SPSS products named are trademarks of SPSS Inc. All other names are trademarks of their respective owners. © 2008 SPSS Inc. All rights reserved. CLM12SPC-0108