Helping Make the Optimal Decision
“Now what?”. “What should we do about it?”… These are questions we often ask when something has happened or we predict that something will happen with a level of certainty. How do we take prescriptive actions or make the optimal decision?
IBM Decision Optimization (DO) for Data Science Experience (DSX) is a prescriptive analytics offering that helps bring the power of prescriptive analytics within reach of data science teams that want to drive operational efficiency and business impact across the organization and complete their AI / ML lifecycle with decision optimization.
IBM DO for DSX, integrated with IBM Data Science Experience, helps data science teams combine optimization and machine learning techniques with model management and deployment capabilities — and other data science capabilities, to help develop innovative solutions and potentially improve operational efficiency. DO for DSX contains multiple examples that show how ML and optimization can be combined to help data science teams learn and adopt the discipline of prescriptive analytics efficiently and effectively. It can help data scientists build prescriptive models, analyze scenarios and solve optimization problems in the same platform and with similar tools that they currently use to build their ML models.
Data scientists can use IBM DO for DSX to help speed the end to end process of building and deploying optimization models. Through the use of the platform features teams can connect to the same data sources that were used to create the ML models.
IBM DO for DSX incorporates IBM CPLEX® solvers to help prescribe optimization models with millions of constraints and variables. This can help organizations expedite solving a wide range of optimization models by taking advantage of IBM CPLEX solvers, which organizations across multiple industries are using to run their mission-critical decision-making applications.
Decision Optimization Integral to the AI / ML lifecycle
DO for DSX is part of the A.I. and machine learning lifecycle as shown in figure #1. It uses machine learning to help organizations take the most appropriate (optimized) prescriptive action. It’s an integral part of the decision-making process and that’s why it has been integrated into the IBM DSX offering.
Typically the process or workflow for Decision Optimization is as follows:
An organization imports their data and selects what they want to apply the prescriptive model to. Once that step is done they create an optimization model. At this point the data science team has the ability to explore the solution either using key performance indicators (KPIs), or visually, using the dashboard as shown in figure #2 below.
The dashboard requires no-coding. Users can configure it with an editor simply by pointing and clicking. There are several tabs which show different types of graphics along with a selection of widgets such as tables and charts to assist the data science teams.
Essential to prescriptive model building is the what-if analysis. Organizations need to compare, different scenarios: What if my budget is increased? What if I finally have less resources available, than initially expected…
Organizations have the ability to duplicate scenarios, or create new ones, using the same data, or different ones. The dashboard enables users to visualize, within the same widget (chart / table…), the data from several scenarios, enabling the visual comparison of different options.
Finally, when the data science team is satisfied with a scenario, they can click the option to ‘deploy” it within the DSX cluster.
DSX contains a Github repository, that enables organizations to tag their models, then the IT person under the ‘admin’ profile, within model management and deployment, deploys the model.
Application developers are presented with an API end point, and a token, that can be integrated as a micro-service call, within the cognitive application they are developing.
It really can be that simple.
Your Next Steps
Whether you are a coder or someone that prefers a more intuitive graphic user interface, IBM Decision Optimization for the Data Science Experience can help organizations embrace decision optimization without having to be an advanced data scientist. If you want to see Decision Optimization with DSX in action watch this video here. For more information on the IBM DO for DSX click this link.
Steven Astorino, Vice President of Development, Private Cloud Platform and z Analytics
Follow me on twitter @astorino_steven
