• SALFORD PREDICTIVE MODELER®

    SALFORD PREDICTIVE MODELER®

    Faster. More Comprehensive Machine Learning.
  • SPM Version 8.0!

    SPM Version 8.0!

    More Automation. Better Results. Take A Giant Step Forward In Your Data Science Productivity With SPM 8
On Demand Introductory Videos
Download Now Instant Evaluation
Get Price Quote

SPM® Testimonials

Rawa Shamroukh, Union Bank, Vice President / Senior Strategic Modeling Manager, Predictive Modeling Department

Using Salfords’ data mining Software TreeNet and CART for the last 8 years, Union Bank has developed multiple in-house statistical models for the various business units within Retail Banking. These tools have greatly expanded our ability to synthesize massive amounts of data quickly, reduce our modeling turn around-times, and mainly produce the most robust predictive models in terms of accuracy and performance. The main benefits from these tools from the bank’s perspective are time savings, Productivity increase, superior analytics, and affordable tools. Treenet and CART provided significant advantages and established a new paradigm for certain modeling challenges within our department. Benefits were obtained along each step of the overall data mining process from data processing to model development and implementation. It demonstrates remarkable performance for both regression and classification models. When I first started using Treenet for logistic Regression Modeling, I always used to develop a parallel model using basic SAS. Treenet kept outperforming SAS logistic regression models over and over. In my opinion, the most amazing features in these tools are their automatic handling of variables selection, collinearity, data transformation, outliers, and many more. In addition, you can have on-the-spot model assessment by generating immediate Gain / Lift charts, ROC measures, Train/ test comparison, Misclassification / prediction success, etc. Using these tools, you will have the options to deploy your final model in a SAS environment or a SQL environment since they automatically translates your model into SAS codes, C codes, or PMML codes. I have used Treenet in various applications. To name few: Attrition models, Behavior scoring models, Operation losses models, Overdraft models, Uncollected funds models, Fraud models, etc.

Rawa Shamroukh
Union Bank
Vice President / Senior Strategic Modeling Manager
Predictive Modeling Department


Sassoon Kosian, Sr. Assistant Vice President and Head of Methodology at EXL Service

In the Decision Analytics practice of EXL Service we build many predictive models for our clients representing a wide range of industries, levels of complexity and application environments. CART and MARS have been standard tools that our resources are trained on and apply in model development. I personally have been using Salford Systems products for nearly 8 years and I believe will continue using them for the foreseeable future. I have mostly used CART and sometimes MARS, and in recent years also tried Random Forests and TreeNet. All of the products offer an intuitive GUI interface and versatility but for me personally CART would be the winner.
When it comes to decision trees there are other options in the market but none has come close to CART in terms of rich functionality, intuitive ease of use and affordability. These are reasons why I continue recommending it to our modeling resources. There several ways we take advantage of CART features. First of all, it is quite easy and quick to build a CART tree with decent predictive power and which you can also explain to business users - something that's very important to our clients. Besides building tree models we often use CART to quickly derive insights about the data by building a preliminary tree and exploring the variable importance list, important nodes and making sense of individual variables. I find the data exploration functionality incredibly valuable especially when you have hundreds or thousands of variables. CART and MARS have also helped us to create composite variables that we have used in regression models - another very useful feature we frequently use. CART also comes in handy to quickly identify segmentation strategies in complex scenarios. All of these features make CART a very important tool in our model development toolkit and help us bring value to our clients.

Sassoon Kosian, Sr. Assistant Vice President and Head of Methodology at EXL Service


Xu Jie, Nanjing University of Information Science & Technology

I am conducting a project about GIS, in which many data analysis are needed. Lacking useful tools, our project made slow progress. In an accidental chance I got a TreeNet trial version and it shocked me with its powerful capabilities of data analyzing, friendly user interface and most important of all, accuracy. After using it we got many benefits from it during our research and our project had gone much faster.
In many features of TreeNet, we like most is plots which offers graphs displaying after building model. This feature is especially useful to us which provides the most visual and easy way to find shortcomings and make improvement of the model. We like the multiple model setting up ways as well, it’s flexible and covers most aspects of our research.
What attracts me most is the amazing speed of TreeNet. We have used some other software before using TreeNet. None of them could build a model in such a short time. What’s more, using TreeNet, the painstaking procedures of data preprocessing are saved, it greatly accelerated our research. Since our data contains over 100 million of variables, using common software, it takes weeks to get a result of analyze. However, by using TreeNet, it takes just a couple of days.
In addition, the most important is the accuracy. Due to the defects of sampling stage, there are some noisy variables in our data, which brings instability to our model. By using TreeNet, we got more stable model of our research than by other tools. What’s more, the model we built by TreeNet could be repeated and verified.
As to ROI, I cannot say how much money we have saved by using this tool, but we did consider to buy a high performance computer(about 5000 US dollars)to assist our research. After using the software, we decided to postpone that purchasing.
We are getting to use TreeNet just a few months , but we really impressed by its powerfulness. We know we are using just a few common features, a lot of powerful features are still waiting for us to learn.

Xu Jie, Nanjing University of Information Science & Technology


Megan Sun, Data Mining Analyst, Marketing Department at Genworth Financial

I have 6 years using SAS and other statistical software to conduct academic and business projects. I started using SPM to build predictive models in May 2014. Our team mainly uses SPM TreeNet to build models for direct mail campaigns. I think the SPM software (Salford Predictive Modeler) is S.P.M. - SMART, PRODUCTIVE and MANAGEABLE.

Read full testimonial

Megan Sun, Data Mining Analyst,
Marketing Department at Genworth Financial


Brian Griner, Chief Methodologist at Quintiles

The Salford Predictive Modeler Software Suite
Great product! Very easy to test different models, compare results and export code to score a database.

 Brian Griner, Chief Methodologist at Quintiles
New York, USA


Jim Kenyon, Director of Operations for Optimization Group.

We use SPM because it lets us quickly and easily build predictive models that produce useful and usable results for our clients.

 Jim Kenyon, Director of Operations at Optimization Group
Ann Arbor, MI, USA


[J#64:1602]

Get In Touch With Us

Contact Us

9685 Via Excelencia, Suite 208, San Diego, CA 92126
Ph: 619-543-8880
Fax: 619-543-8888
info (at) salford-systems (dot) com