한글 홈페이지
한글 CambridgeSoft | 한글 Systat Software  
Salford System   |  제품   |  CART   |  MARS   |  기술지원   |  White Papers   |  가격/주문   |  문의

· Leo Breiman Elected to National Academy of Sciences

· Salford Systems Wins International Web Mining Competition

· Japanese Deming Prize Committee Awards 1999 Nikkei Quality Control Literature Prize to Dan Steinberg, President of Salford Systems

· The Next Technology Advance in Data Mining Has Arrived -- Salford Systems Launches Exclusive Predictive Modeling Solution

· Salford Systems Introduces CART -- Robust Decision-Tree Software for Data Mining

Salford Systems Wins International Web Mining Competition
Developer of CART®, MARS, HotSpotDetector, and TreeNet Data Mining tools wins in the KDD Cup 2000 (Knowledge Discovery and Data Mining) competition.

August 30, 2000 - San Diego, CA., - Salford Systems has won an international web mining competition organized by the ACM (Association for Computing Machinery), the School of Electrical and Computer Engineering, Purdue University, and Blue Martini Software. The announcements and awards ceremony was held in Boston, MA at the 6th Annual ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, which attracted over 900 data mining experts from North America, Latin America, Europe, and Asia. The competition drew more than 170 participants including major firms in the data mining and business intelligence world. Only the names of winners and honorable mentions are released by the organizing committee.

The two-month long competition involved intensive analysis of the clickstream and customer data of an e-commerce retail web site. Contestants were provided data from February and March of this year and were asked to apply their models against the April clickstream. Models were designed to support web-site personalization and to improve the profitability of the site by increasing customer response. Salford Systems won in the category of identifying and characterizing the most valuable customers using Salford Systems CART, MARS, HotSpotDetector and TreeNet software.

About Salford Systems

Salford Systems provides data mining and web mining software and consultation services. Applications include CRM, market research, segmentation, direct marketing, fraud detection, credit scoring, risk management, bio-medical research and manufacturing quality control. Industries using Salford Systems products and services include banking, financial services, insurance, telecommunications, transportation, health care, manufacturing, retail and catalog sales, and education. Salford Systems software is installed at more than 4,500 sites worldwide, including 300 major universities. Salford has its headquarters in San Diego, Calif., and representation throughout North America, Europe and Australia.

For more information about KDDCup 2000 see http://www.ecn.purdue.edu/KDDCUP

Japanese Deming Prize Committee Awards 1999 Nikkei Quality Control Literature Prize
to Dan Steinberg, President of Salford Systems

November 1, 1999 - San Diego, Calif., - Dan Steinberg, President of Salford Systems, has won Japan's Nikkei Quality Control (QC) Literature Prize.  The award, announced last week, recognizes literature that contributes to the progress and further development of quality control and quality management.

The prize-winning book is a theoretical and practical introduction to CART® (Classification and Regression Trees) decision-tree methodology.  Titled in English, "Applied Tree-Based Methods Using CART," the book is co-authored by Yuji Horie and Atsushi Ootaki.   Currently available only in Japanese from JUSE Press Ltd. (Tokyo), the book provides a detailed road map for using CART as implemented in Salford Systems software.   A broad range of real world problems is discussed, including quality control, market research, credit risk management, and medical diagnosis.

The Nikkei QC Literature Prize, established in 1954 by the Nippon Keizai Shimbun (Japan Economic Journal), is awarded annually by the Deming Prize Committee, which also awards the Deming Prize for achievement of significant gains in performance through Total Quality Management (TQM).   The award ceremony will be held November  15th, 1999, in Tokyo.

About CART

CART is an automatic high-speed data analysis tool capable of discovering complex relationships in huge databases.  Results are displayed in flowcharts that are easy for non-specialists to understand.   CART is notably successful in quality control applications such as predicting assembly-line failures and conditions under which defects are most likely to occur.  In addition, the software is widely used in business intelligence applications such as direct mail targeting, managing credit risk, developing customer retention and acquisition strategies, and detecting credit card fraud.

About Dan Steinberg and Salford Systems

Dan Steinberg is President and Founder of Salford Systems, a San Diego-based company specializing in business intelligence software and consulting services. He received his Ph.D. in Economics from Harvard University and has 20 years of experience in the industry.
Salford Systems products and services are used in banking, financial services, insurance, telecommunications, transportation, health care, manufacturing, retail and catalog sales, and education. Key customers include AT&T Universal Card Services, Fireman's Fund Insurance, Pfizer Pharmaceuticals, General Motors, and Sears, Roebuck and Co.  Salford Systems software is installed at more than 4,500 sites worldwide, including 300 major universities.

The company is headquartered in San Diego, CA, with representation throughout North America, Japan, Europe and Australia.

The Next Technology Advance in Data Mining Has Arrived:
Salford Systems Launches Exclusive Predictive Modeling Solution

MARS® software automatically builds deployable business models
with superior predictive accuracy and speed

August 3, 1999 - San Diego, Calif., - Salford Systems today introduces MARS, a high-speed flexible regression tool for data mining and predictive modeling.  A natural alternative to Neural Networks and more conventional regression models, MARS was designed to solve problems such as how to predict credit card holder balances, insurance claim losses, and cell phone usage with superior forecasting accuracy.  MARS builds its models automatically, self-tests to ensure validity, and graphically displays the effect of each important variable on the outcome.

"MARS is the ideal modeling tool when an analyst needs to accurately predict a future outcome and also needs to understand the 'why' underlying the predictive model," says Dan Steinberg, president of Salford Systems.   "Prior to the advent of MARS, data miners had to choose between tools that optimized predictive accuracy or tools that told a story about the underlying data patterns and relationships.  Now, with MARS, an analyst no longer has to sacrifice intelligibility to obtain predictive accuracy.  Further, the time-consuming trial and error process of building accurate regression models is a thing of the past with the availability of MARS" highly-automated, fast analytical engine and unique model-building capability."

Herb Edelstein, President of Two Crows Data Mining Consultancy and a widely regarded industry analyst, predicts that "MARS will be one of the next hottest algorithms.  MARS addresses some of the shortcomings of decision trees, and it does so in a fairly elegant fashion."  Independent studies have demonstrated that MARS not only outperforms neural networks in a wide variety of settings, but is also hundreds of times faster.

Easy to Use Interface

The intuitive, Windows-based interface and intelligent default settings enable non-technical business users to run MARS easily.  The analyst provides MARS with a database and target variable; MARS then develops a model, self-tests to prevent over-fitting, and graphically displays the impact of each predictive factor on the outcome.  Predictions for new data can be obtained either through the MARS engine or via C-source code produced as part of the modeling process.
MARS is ideal for predictive modeling problems involving continuous outcomes, such as how much a customer will spend on a catalog order, the minutes a customer will use his cell phone, or how electricity production will change as generator inputs change.  MARS is also adept at predicting 'yes/no' binary responses such as whether a consumer will default on a loan, refinance a mortgage, or defect to a competitor.

Experienced data analysts also can use MARS as an exploratory tool to refine and improve more conventional linear and logistic regression models.  By automatically detecting needed variable transformations and interactions, MARS can radically reduce the time required to build a model and significantly improve its predictive accuracy.

"MARS is an essential tool for any data miner," says Thomas Brauch, Marketing Manager of the Data-Driven Marketing Department at Fireman's Fund Insurance.  "It finds significant effects in complex data structures where other methods simply fail.  I use it as both a stand alone solution and as a transformation tool for simpler modeling techniques."

What is MARS and What Does It Do?

MARS, an acronym for Multivariate Adaptive Regression Splines, is a non-parametric regression procedure.  It extends technology originally introduced in CART®, the industry's leading decision-tree software.  MARS automates all aspects of model development and deployment, including:
      • separating relevant from irrelevant predictors;
      • transforming predictors exhibiting nonlinear relationships with the target variable;
      • determining interactions between predictors;
      • handling missing values with new surrogate variable techniques; and
      • conducting self-validation tests to ensure the model holds up to new data.
MARS enables analysts to rapidly search through a large number of candidate models and quickly identify the 'optimal' solution - providing valuable business understanding in the process.   Predictive models can be deployed directly from within the software or exported as ready-to-run C and SAS® compatible source code.

*For more detail on the MARS methodology and its features and functionality, please refer to the enclosed "Frequently Asked Questions and Answers About MARS" document.

Adaptable Across Environments

MARS is supported in a variety of standalone and client/server operating environments.  Operating systems supported include Windows 95/98, Windows NT, UNIX, and IBM MVS and CMS.  Hardware platforms supported include Intel PCs, Sun, SGI, HP, Compaq Alpha, IBM RS6000, and IBM mainframes. MARS' scalability is limited only by a system's available RAM.  The software requires a minimum of 10 MB of disk storage and 64 megabytes of RAM.

MARS' data-translation engines convert data from more than 80 file formats, including popular statistical-analysis packages such as SAS® and SPSS, databases such as Oracle and Informix, and spreadsheets such as Microsoft Excel and Lotus.

About Salford Systems

Founded in 1983, Salford Systems specializes in data-mining and data-analysis products and training and consultation services for a broad range of industries.  Salford Systems also offers statistical consulting and custom programming services for the design and analysis of discrete choice experiments.  The company has its headquarters in San Diego, Calif., with representation throughout North America, Japan, Europe and Australia.

Salford Systems Introduces CART®:
Robust Decision Tree Software for Data Mining

Affordable, proven technology reveals hidden relationships;
automatically generates accurate predictive models

May 2, 1998 - San Diego, Calif.,- Salford Systems, a new-generation data-mining software developer and consultant, today introduces CART®, user-friendly classification-and-regression-tree software.  CART produces the most reliable classification and prediction models for applications such as profiling "best" customers, targeting direct mailings, detecting telecommunications and credit-card fraud, and managing credit risk.

"The most important data-mining business applications, such as classification and predictive modeling, can be accomplished using just CART," says Dan Steinberg, president of Salford Systems.  "Many businesses don't need to go overboard buying data-mining suites, which contain multiple data-analysis components, cost tens of thousands of dollars, and require a high level of expertise to operate.  As a cost-effective stand-alone package, CART gives beginning data miners a highly accurate, easy-to-use tool that does not require technical expertise."  For experienced data miners, CART is a high-performance, proven methodology that can be used as an independent data-mining system and as a companion tool that can significantly extend preprocessing capabilities, such as variable selection, for neural nets and other data-mining systems.

Accessible, Sophisticated Functionality

CART uses an intuitive, Windows-based interface that enables non-technical business users to create models quickly and interpret results easily.  The software uses historical data to discover patterns, trends and relationships, and it automatically generates high-performance predictive models that can be applied to new data.  This information facilitates better business decisions and increases profitability.  In addition, CART can grow to fit your business by easily expanding from single-desktop applications to enterprise-wide servers accessing data marts and data warehouses.

For experienced data analysts, CART provides the following advanced features* in a combination not available in any other decision-tree package:

      • multiple automatic self-validation procedures;
      • adjustable misclassification penalties;
      • intelligent surrogates for missing values;
      • eight choices for tree-growing criteria;
      • multiple-tree, committee-of-expert methods, or bootstrap aggregation; and
      • a complete programming language with flow control for on-the-fly data manipulation.

"CART's range of features make it an exceptional out-of-box data-analysis package for beginning and experienced data miners," says Steinberg.  "Even with 'dirty' datasets, or those with many missing values, CART will develop robust, stable models.  In addition, CART's default-setting performance rivals - and sometimes outperforms - neural nets."

*For more detail on CART and its advanced features, please refer to "Frequently Asked Questions and Answers About CART".

Branching Across Industries

Worldwide, CART has more than 1,000 users found in nearly all industry segments, including marketing, financial services, insurance, retail trade, health care, pharmaceutical, manufacturing, telecommunications, energy, agricultural, transportation and education.  In these data-intensive industries, CART is especially efficient in discovering multi-dimensional relationships within large, complex data warehouses - the data repositories that now drive mission-critical business decisions.

"CART is an important statistical-analysis tool that we use to segment our databases and predict risk factors for the Sears Card," says Steven Li, Sears, Roebuck and Co.'s senior manager of risk technology.  "The advantage of the decision-tree format is that our results are easy to interpret; especially with CART, we are able to see a great deal of detail about each of the nodes, such as the node's misclassification 'costs,' the count of data assigned to that node, and a display of the surrogate values substituted for the node."

"In addition to risk management, CART is notably successful in targeting direct mailings and improving response rates.  In the financial services industry, the software is used to retain customers by making preemptive offers to mortgage holders identified as most likely to refinance their homes.  In telecommunications, CART is used to identify households likely to "churn," or switch carriers in a given time frame.

CART is also used for detecting credit-card and insurance fraud; customer profiling and market segmentation; identifying cross- and up-selling opportunities; credit-card scoring; medical diagnostic-test development; predicting assembly-line failures; and myriad other business-intelligence needs.

Time-Tested, Proven Methodology

On the surface, CART's design automates several processes that traditionally require computer programming and statistical expertise.  Underlying the "easy" interface, however, is a mature theoretical foundation that distinguishes CART from other methodologies and other decision trees.  Salford Systems' CART is the only decision-tree system based on the original CART code developed by world-renowned Stanford University and University of California at Berkeley statisticians; this code now includes enhancements that were co-developed by Salford Systems and CART's originators.

"CART is the most accurate decision-tree software commercially available," says Steinberg.  "This tree method is the fruit of a decade of machine-learning and statistical research, and our software is the only complete implementation of the original, proven algorithm."

Adaptable Across Environments

CART supports a variety of desktop standalone and client/server operating environments.  Operating systems supported include Windows 3.x, Windows 95, Windows NT, Mac OS, UNIX, IBM MVS and CMS.  Hardware systems supported include Intel PCs, Sun, SGI, HP, Digital Alpha and VAX, and IBM RS6000 machines.  CART's scalability is limited only by a system's available RAM; at the least, it requires 10 MB of disk storage.  CART's data-translation engine supports data conversions for more than 70 file formats, including popular statistical-analysis packages, such as SAS® and SPSS; databases, such as Oracle and Informix; and spreadsheets, such as Microsoft Excel and Lotus.

Salford System  | 제품 | CART | MARS | 가격/주문 | White Papers | 기술지원 | 문의