Objectifi joins the leaderboard in $3 Million Data Mining Contest

How can predictive modeling save lives?

More than 71 million individuals in the United States are admitted to hospitals each year[1], with $30 billion spent on unnecessary admissions. Dr. Richard Merkin, President and CEO of Heritage Provider Network, set out to improve the healthcare of all Americans while lowering the cost of care for the average family. Believing that incentivized competition is the best way to achieve the radical breakthroughs necessary to begin fixing America’s health care system, he created the Heritage Health Prize Competition and challenged industry leaders to build an algorithm that uses available patient data to predict and prevent unnecessary hospitalizations.

Of the 748 teams competing, Objectifi is proud to announce that we are currently holding 8th place. “We built and evaluated 1000+ predictive models, applied 50+ attributes transformations and examined 20+ advanced learning algorithms in just 6 months!” (Dr. Saed Sayad). Not even half-way through the two year competition, we have a long way to go yet, but are confident that we have some of the brightest minds in the industry on our team, led by Dr. Saed Sayad. To help solve real world problems like the those facing the Heritage Provider Network, Objectifi uses industry leading data mining tools including Oracle R, Oracle Real Time Decisions (RTD), Oracle Data Mining (ODM) and others.

If successful, the winning algorithm would help identify earlier those most at risk and ensure they get the treatment they need, by predicting how many days a patient will spend in a hospital in the next year. Health care providers can then develop new care plans and strategies to reach patients before emergencies occur, increasing the health of patients while decreasing the cost of care.

The $3 million prize is being awarded by Heritage Provider Network. As one of the largest networks of physicians in the world, Heritage Provider Network provides healthcare for approximately 700,000 Californians, by connecting patients with physicians, specialists, and hospitals.

Check back soon for an interview with Dr. Saed Sayad to learn more about the competition, and how Objectifi is changing the future of the American healthcare system.

To learn more about the contest, please visit http://www.heritagehealthprize.com.

To learn more about Objectifi, please visit Objectifi.com.

[1] According to the latest survey from the American Hospital Association, http://www.heritagehealthprize.com/c/hhp/Details/FAQ

Review of RTD Decision Manager

Oracle Real-Time Decisions is Oracle’s premiere decision management product, a powerful platform that combines both rules-based qualification and predictive analytics to power solutions for real-time enterprise decision management. Until now, however, users were left to manage their platform by either developing their own User Interface or by managing it directly from the robust back-end Decision Center. Oracle RTD Decision Manager provides an intuitive and easy to use business-level interface, allowing business users to manage and control their systems directly. This end-user oriented web-based tool is configured specifically to each organization’s own choices, and allows users access to select Decision Center analytic reports directly from the Decision Manager interface.

Decision Manager Overview

The major benefit of the Decision Manager is that it can be used easily and effectively by users with little technical know-how. Users can be assigned specific views and permissions, allowing different users to view different role-based perspectives. For example, Channel Managers can have a Placement (or Page)-based view while a Marketing Manager can have a campaign-based view. This allows different users to configure the tool in the way that works best for them, while ultimately managing the same set of objects.

The Decision Manager uses a folder-based structure specific to the individual installation. Creatives can be organized into Offers, and Offers into Campaigns, each with its own basic eligibility information (start and end dates, for example) as well as user-defined business rules for more targeted eligibility. Business rules can use any customer attribute, including session and contextual data.

The Deicison Manager allows users to organize their choices into Projects. Projects are defined groupings of choices within which elements can be bundled together, and then versioned, tested, deployed, and disabled as a unit, enabling business users to collaborate with content creators and technical users on the same project. Tags are another useful addition, allowing users to mark choices, then later pull up all offers with the same tag (Holiday, for example).


The Main screen of the Decision Manager is arranged into four sections: the Project Context box, the Panel box, the Detail Box, and a Search box. There is also a helpful Breadcrumb Line that appears in the Header when a choice is selected.

Decision Manager Main Screen

The Project Context Box

The Project Context Box

The Project Context box displays the context in which a user wants to operate, either within the Main Repository or within an active Project. The Main Repository is the database that contains all of the current choices in RTD. It is used for viewing choices and Projects, but users cannot create, edit, or delete choices; they must open the Project to make those changes.  The Project Context box is used primarily to inform the user of what Project they are working within.

The Panel Box

The Panel Box

The Panel box displays a hierarchical tree of the choices within the current Project, the changes in a Project, or the results of a search. The changes in a Project and Search Results generally appear as bar entries at the bottom of the Panel box that the user can toggle between.

The Detail Box

The Detail box is the main work area within the Decision Manager. When a user selects an existing choice or creates a new choice, the required fields and options will open within the Detail box. The contents of the Detail box will vary depending on the option selected: Create Project, View Project, Create Choice, or Choice Details.

The Decision Manager includes some nice usability features including customizable error and warning messages, tooltips to help users populate fields correctly, and a choice-level audit trail.


Oracle RTD Decision Manager provides an easy to use interface for Real-Time Decisions, allowing non-technical business users to easily create and manage all elements of their campaigns. Decision Manager allows seamless integration with content and campaign management systems while utilizing the powerful self-learning abilities that RTD is known for, all with a simple, intuitive, and easily learned interface.

For more information about Oracle Real-Time Decisions or the Decision Manager, call Objectifi today.

Debugging an Oracle Real-Time Decisions Project in Eclipse

When developing and debugging a new Inline Service (ILS) for Oracle RTD, programmers sometimes require more than just standard console logging functionality. Because an ILS is actually code running on the server, debugging an ILS requires the use of Java remote debugging technology. In this blog post, we will walk through the procedure for debugging an Inline Service in Eclipse, for Oracle Real-Time Decisions.
There are three major steps involved in debugging an ILS from within Eclipse:

  1. Enabling remote debugging on the server.
  2. Creating the remote debugging configuration in Eclipse.
  3. Setting up breakpoints and debugging using Eclipse Visual Debugger

Enabling remote debugging on the server

Enabling remote debugging on the server side is done using Java -X options. Because these are not standard options, you may need to verify how it will work for your application server/JVM platform.

On oc4j under Windows, simply edit C:\OracleBI\RTD\oc4j\bin\oc4j.cmd by adding the following line at an appropriate location:

set JVMARGS=%OC4J_JVM_ARGS% -server -Xdebug -Xrunjdwp:transport=dt_socket,server=y,suspend=n,address=4000

This will configure the application server to listen on port 4000 for the remote debugging session.

To enable remote debugging on port 4000 for Weblogic 11gR1 (that comes with Oracle Fusion on Windows, you would use the following:

SET JAVA_OPTIONS=-Xdebug -Xrunjdwp:transport=dt_socket,server=y,suspend=n,address=4000

When starting the server, you should see something like the following image to confirm that the server is correctly set up to listen on remote debugging session on port 4000:

Creating the remote debugging configuration in Eclipse

Now that the server is ready, you can create the remote debugging configuration in Eclipse. We will use the CrossSell sample that comes with Oracle Real-Time Decisions as the base example.

1. Deploy the latest version of your ILS in Development mode.

2. Switch to the Java perspective in Eclipse.

3. Create the debugging configuration by clicking the Debug button  on the menu bar.

4. Select Remote Java Application and click New.

5. Name the configuration and select the appropriate project. Set the port to 4000 and click Apply.

6. You are now ready to debug the ILS. Start the debugging session by clicking the Debug button in the previous modal window, or by selecting your configuration from the Debug drop-down menu.

7. When it has finished, you must end the debugging session. To do so, click the Disconnect button from the Debug view.

Setting up breakpoints and debugging using Eclipse Visual Debugger

The Eclipse ILS will generate code for all of the ILS objects (such as choices, decisions, or integration points). The generated code will include the various code snippets added in the properties of these ILS objects, for instance, in the Pre/Post selection logic panel of a decision object. Because of the generated code, setting a breakpoint is not trivial as it is to set a (normal) breakpoint in regular code but is not difficult to do, as the generated sources all follow a simple naming convention and are all generated in the same location. The following procedure details how to set up a breakpoint in the CrossSell ILS’s Decision named ‘OfferDecision’. This decision is used by an advisor named ‘OfferRequest’. So setting a breakpoint in this decision will make the program stop every time we call the ‘OfferRequest’ advisor.

The Decision Pre/Post Logic panel:

Setting up a breakpoint in your ILS and tracing your code using the Eclipse debugger:

1. Go to the Java perspective.

2. In the Package Explorer, navigate to gensrc/com.sigmadynamics.sdo/GENOfferDecision.java

3. Locate the execute(int) method. Inside this function you will find the code that was declared in the Pre and Post Selection Logic section. There are comments the clearly defines them (// SDCUSTOMCODESTART.GetBestCreative.PostSelectionBody.java.0 for post logic for instance). Set a breakpoint on the line where the decision post-selection logic begins.

4. We can now call the advisor using the test view of the ILS. Eclipse will switch to the Debug perspective at the breakpoint that you just set.

5. You may receive a ‘Source not found’ message. If this is the case follow step 5.1 to 5.5. Otherwise, continue to Step 6.

5.1 Click on the Edit Source Lookup Path button. The following modal window will open:

5.2 Click  Add, then select Workspace Folder.

5.3 In the next modal window, select the gensrc folder inside the CrossSell Project and click OK.

5.4 Click OK again when you are returned to the initial modal window:

6. We can now step in and step out through the Decision’s code, as well as browse through the variable values, or expand arrays as you can in any GUI debugger.

Contact Objectifi for more information! www.objectifi.com

Oracle RTD Case Study – Dell Computers

Learn how Dell transformed its Call Centers into Profit Centers with Oracle RTD and Objectifi.

Download the Case Study:

Using Personalization with social media

Every marketer knows the growing importance of maintaining an active social media campaign in any marketing plan. Social Media allows you to interact with your customers daily in a direct and meaningful way, to get feedback, and to engage in real-time dialogue. But other than maintaining your social media accounts and advertising opportunities on popular sites like Facebook and Foursquare, is there a way to use the growing popularity of social media to connect with your clients?

One under-used method is to create “Brand Advocates” and use personalization technoligies such as Oracle Real-Time Decisions to present special offers to them.  Brand Advocates are your most socially active customers; customers that tweet numerous times a day, have more Facebook friends than they could ever possibly list, and can easily be located by their active FourSquare accounts at any time. What makes these socially over-active customers so valuable is the size of their captive audience and the influence they have over them.

Recognizing these customers as Brand Advocates allows Oracle Real-Time Decisions to select special offers to present to them. For example, a customer logs into her cell phone provider’s website to pay her monthly balance. Oracle RTD recognizes her as a Brand Advocate and presents an offer for an exclusive new Smart Phone that isn’t yet available to the public. Considering how actively she uses social networking sites, she is eager to upgrade from her Talk-and-Text phone. She accepts the offer and receives her new phone shortly.  In no time at all, she is talking, tweeting, and blogging about this exclusive new phone to all of her friends, reaching potential new customers before the phone has even been released!

Oracle Real-Time Decisions can even be used to identify customers as “Brand Advocates”.  By tracking customers that refer offers to their friends and seeing what attributes are unique about them, RTD can model customers based on whether or not they have influence. For example, you can create an email campaign that offers a new cell phone to existing customers with a unique tracking code for each email. After sending the email out to customers, you can trace the tracking codes to see which codes generate clicks from more than just the original recipient.  This allows you to track which customers forwarded the email to their friends and how many of their friends clicked on the offer; determining that those customers are more influential.  In the future, Oracle RTD can treat those influencers uniquely to ensure that they always have a positive experience.

By recognizing Brand Advocates with Oracle Real-Time Decisions, you can create buzz around any new product or service. And because people trust comments and recommendations from their friends, recommendations from Brand Advocates can be a lot more effective than traditional advertising methods.  Use Personalization with Oracle Real-Time Decisions to transform your most socially active customers into champions of your brand!

Increasing sales through personalized in-bound marketing

As consumers become weary of the thousands of advertisements they see each day, more and more they start to look in the other direction, ignoring offers that they might otherwise be interested in. Traditional “Outbound Marketing Strategies” such as mailers, emails, and cold calls are proving to be less and less effective. Increasing subscriber opt-outs and growing “do not call” lists have made reaching customers harder than ever. As companies clamor to change their marketing strategies, the most successful marketers are looking at the tools they already have in place. With this turn away from out-bound marketing techniques, in-bound strategies are taking on a new importance.

Inbound Marketing utilizes services you are already offering your customers: your website, social media accounts, and blogs. Instead of intruding on consumers by interrupting them to advertise a product or service, smart marketers are taking advantage of situations where the consumer is already engaged. While inbound strategies are proving be the future of successful marketing, one of the biggest overlooked opportunities in inbound marketing is something most companies already have in place: call centers.

Presenting offers to customers calling into a call center is nothing new; traditional methods have agents referring to precomputed lists of appropriate offers. But these lists have serious limitations; they cannot account for the context of the interaction between the customer and the company. Personalization with Oracle Real-Time Decisions presents relevant, current offers to the agent and can take into account the reason for the call, recent transactions, exact state of the account, time of interaction, and call method (mobile device, computer, home phone, etc.). There are an almost infinite number of inputs available with Oracle RTD, and over 200 real-time inputs that offer the customer the most relevant experience possible.

Personalization even allows for this information to be shared between channels. Imagine a customer calling in to a call center for his telecommunications provider. The agent helps him sort out the Cable TV issue he called in about, and in the process discovers that the customer is on a competitor’s cell phone plan. While the customer is interested in bundling his services, his current cell phone contract does not expire for another 8 months. Personalization between channels allows an outbound campaign to be launched for this customer just prior to him renewing his contract with the competing service provider.

By personalizing offers from service agents, customer support lines and call centers can become a crucial part of your marketing strategy. Optimization with Oracle Real-Time Decisions allows your call center employees to engage in personalized dialogues, creating meaningful interactions that improve the customer’s experience and generate sales.

By understanding customer preferences, purchases, and past interactions, call center employees can suggest only the offers that the caller is most likely to respond to. Oracle RTD allows your call center agents to present the right offer to the right customer at the right time, making your call centers the front-line of your marketing strategy.

Copyright © Objectifi Inc.