Business analysts gain new power and flexibility to build predictive and prescriptive analytic models.
Frontline Systems is shipping a new product line for desktop and cloud, Analytic Solver® 2017-R2 for Excel and enhancements to AnalyticSolver.com, its SaaS (software as a service) equivalent. Building on the February V2017 release that unified and simplified Frontline’s product and introduced Analytic Solver Basic, the V2017-R2 release features major performance improvements in optimization, new distribution and correlation functionality in simulation/risk analysis, and a new, simple way to convert Excel analytic models into Power BI® custom visuals, for distribution to business intelligence consumers throughout an organization.
“We’re not only making advanced analytics power tools and learning aids accessible to every business analyst,” said Daniel Fylstra, Frontline’s President and CEO. “We’re making the fruits of their efforts – working analytic models and their insights – readily accessible to everyone who needs them.”
All that has changed with Analytic Solver V2017-R2. Working in Excel with this tool, a business analyst user can turn his or her Excel-based optimization or simulation model into a Power BI Custom Visual, with just a few mouse clicks. Users simply select rows or columns of data to serve as changeable parameters, then choose Create App – Power BI, and save a custom visual file created by Analytic Solver. They click the Load Custom Visual icon in Power BI, and select the file just saved. The result isn’t just a chart – it’s a fully-functional optimization or simulation model, ready to accept Power BI data, run on demand on the Web, and display visual results in Power BI. Users simply need to drag and drop appropriate Power BI datasets into the “well” of inputs to match model parameters.
The Magic of RASON®
The secret behind the ability to turn desktop Excel models into Web-based Power BI custom visuals, without any programming, is Analytic Solver’s ability to translate an Excel model into RASON (RESTful Analytic Solver Object Notation), Frontline’s powerful modeling language embedded in JSON. While not every model can be automatically translated, a wide array of simulation/risk analysis and linear and nonlinear optimization models can be.
Analytic Solver then “wraps” Power BI custom visual code around the RASON model. The result is a radically simple way to move from traditional Excel to cloud deployment to business intelligence consumers – not available anywhere else.
Performance Gains for Large Models
For large optimization models with tens to hundreds of thousands of decision variables, the V2017-R2 release greatly improves the speed of “problem setup,” where a Jacobian matrix is built from many thousands of Excel formulas. Though individual results vary, Frontline has seen speed gains of 10 to 50 times on real customer models. These improvements benefit all of Frontline’s 12 different “Solver Engines” that handle problems ranging from linear programming and nonlinear optimization, to arbitrary models and stochastic optimization.
Also part of the V2017-R2 release are new, higher performance Solver Engines for optimization based on latest releases of the Gurobi Solver, FICO’s Xpress Solver, and Artelys Knitro Solver, as well as Frontline’s greatly-enhanced, high-performance Evolutionary Solver, which is focused on challenging non-smooth, non-convex models with integer variables. Free trials of the optional large-scale Solver Engines are “built in” and available to users at any time.
New Simulation/Risk Analysis Functionality
The V2017-R2 release supports a new family of probability distributions, called the metalog distributions, even more general than the Pearson distributions – members of this family can be chosen based directly on historical data (even just a few observations), without a distribution fitting process. The metalog distributions allow empirical data to “speak for itself” in simulation models.
Frontline’s V2016-R2 release introduced Gaussian, Student, and Archimedean copulas – popular in quantitative finance – to define correlation among uncertain variables with dissimilar distributions – complementing existing support for rank-order correlation. New in the V2017-R2 release is the ability to fit copula parameters to historical data – a complement to automatic distribution fitting sometimes called “correlation fitting.”
The new release builds on Frontline’s V2017 release earlier this year, which delivered a wide range of enhancements for data mining and predictive analytics including the ability to create ensembles of any mix of six different classification methods and four different prediction methods. Direct use of categorical variables in many predictive algorithms that, in other analytics software, can use only continuous variables is supported. It also has the ability to save and apply to new data an array of data transformation and unsupervised learning tools, such as PCA and clustering
Enhanced neural networks with any number of layers, and better scaling, training, and activation function options and support for industry-standard PMML (Predictive Model Markup Language), plus PMML extensions for text mining are incorporated.
New Release Available for Download
The new V2017-R2 release is available now at Frontline’s Website, Solver.com, and at its cloud version portal AnalyticSolver.com. In this release, Frontline is making available to all users Analytic Solver Basic, a new version of its Analytic Solver Platform for Education product, formerly offered only for university MBA programs, and used by hundreds of thousands of students. Analytic Solver Basic includes every feature of Frontline’s analytics software, but handles only limited-size models and datasets, suitable for learning. (All 100+ example models and datasets included with the software can be used.) A software license is just $250 per year and includes both the desktop version and access to AnalyticSolver.com. Paired with Solver Academy, Frontline’s new Open edX-based online learning platform, Analytic Solver Basic offers a way for business analysts to gain “advanced analytics ninja” skills.
Frontline Systems Inc. (www.solver.com) is the leader in analytics for spreadsheets and the Web, helping managers gain insights and make better decisions for an uncertain future. Its products integrate forecasting and data mining for predictive analytics, Monte Carlo simulation and risk analysis, and conventional and stochastic optimization for prescriptive analytics. Founded in 1987, Frontline is based in Incline Village, Nevada (775-831-0300).
Excel® and Power BI® are registered trademarks of Microsoft Corp. Knitro® is a registered trademark of Artelys/Ziena Optimization. Analytic Solver and RASON are registered trademarks of Frontline Systems, Inc.