INDUSTRY NEWS

Frontline Systems releases XLMiner.com platform for advanced analytics

Frontline Systems, developer of the Solver in desktop Microsoft Excel, recently released XLMiner.com, a SaaS (software as a service) platform for data mining, text mining, forecasting and predictive analytics using only a browser. XLMiner for the Web offers business analysts point-and-click tools to create predictive analytics models themselves, without being expert data scientists or programmers.

Simultaneously, Frontline Systems announced that more than 100,000 users have adopted its advanced analytics add-ins for Excel Online and Google Sheets, according to data from Microsoft and Google. The add-ins include Solver for linear and nonlinear optimization; Risk Solver for Monte Carlo simulation and risk analysis; XLMiner Analysis ToolPak for linear and logistic regression; and XLMiner Data Visualization for Excel Online.
“Rapidly growing use of advanced analytics by spreadsheet-savvy business analysts, both desktop and cloud, is a ‘stealth trend’ that many industry observers have missed,” says Daniel Fylstra, Frontline’s president and CEO. “Other vendors are talking about ‘democratizing analytics’ through easier to use tools, but Frontline has been delivering on this vision for years.”

Frontline’s Solver and XLMiner brand names have long histories, related to Microsoft Excel, but today they represent advanced prescriptive and predictive analytics tools, offered in desktop and cloud versions for spreadsheet users, application users and developers.

Frontline’s Solver for desktop Excel, first introduced with Excel 3.0 in 1990, has been included in more than 1.2 billion copies of Microsoft Office and Excel, but now Frontline’s advanced Solvers work in desktop and cloud spreadsheets, desktop and cloud versions of its RASON modeling language, and desktop and cloud Solver SDK tools for developers. XLMiner, originally an Excel add-in created by Cytel Software and marketed by Statistics.com, was acquired by Frontline in 2011, rewritten from the ground up to use Web UI technologies and state-of-the-art analytics algorithms, and is now available for use on the Web with spreadsheets or independently without them.

The new XLMiner SaaS application offers point-and-click access to data in uploaded spreadsheets, CSV or text files, Azure SQL databases and Apache Spark Big Data clusters – including Frontline’s own Spark cluster running 24x7 on Amazon Web Services, which has been in use by Frontline’s academic customers since June 2015. An earlier case study shows how XLMiner users can build predictive models from big data that match the accuracy of a similar HortonWorks case study using a public FAA airline data set but without any of the complexity of Hadoop, Pig, Python and other developer tools.

For years, XLMiner has been the most popular way to introduce predictive analytics, and Frontline’s Solvers have been the most popular way to introduce prescriptive analytics in MBA and undergraduate business courses. In the last seven years, more than 500,000 students have learned analytics using Frontline’s software tools. XLMiner for the Web significantly expands access to data mining, text mining, forecasting and predictive analytics tools for students (and commercial users) with MacBooks, iPad tablets, and similar devices.

While XLMiner emphasizes ease of use, especially for first-time users of predictive analytics, on its higher-level subscription plans it can handle large data sets and very challenging problems – for example finding “best subsets” of many variables via exhaustive search in multiple linear regression. It performs operations such as clustering, principal components analysis, text analysis and latent semantic indexing, and training of ensembles of neural networks or classification and regression trees on Frontline’s RASON Server – a cloud analytics platform on Microsoft Azure that also handles challenging optimization and simulation models for Frontline’s customers.

AIMMS 4.14 offers many performance improvements

AIMMS recently released AIMMS 4.14, which includes the following performance improvements when solving MIP/LP problems with CPLEX and various AIMMS WebUI features to improve the user experience:
Parallel solving automatically enabled for CPLEX. The default values of the CPLEX options “Parallel mode” and “Global thread limit” have been changed for CPLEX 12.6 and higher. By default, CPLEX will now use the deterministic mode and all available threads for solving MIP problems and LP problems if the barrier algorithm is used. This can lead to significant performance improvements when solving MIP or LP problems.

Improved identifier selection for widgets. Selecting identifiers to determine the contents of a new widget, or editing an existing widget, have been improved by offering a more extended wizard. These new features make life easier for the app developer and for the end user who is changing widget contents.

‘No Changes Allowed’ feature for WebUI Apps. We added a “No Changes Allowed” feature for AIMMS WebUI Apps, preventing the end user from making changes to the setup of widgets or pages (end users can of course still change data). This feature is set through AIMMS PRO roles and thus can be differentiated per user.

Improved line up widgets. AIMMS improved the way widgets and items, such as buttons, line up to help achieve higher data densities and make Apps look better.

Shift-Click in Multi-Select Widget. Users can now use shift-click in the Multi-Select Widget to select a range of elements shown, at once.

Control save/restore WebUI State. By default an AIMMS WebUI App will save its user state (i.e., view) upon closing, and restore its state upon starting the App. As a developer, you can now have more control of this behavior, e.g., override it to not save any user state, or only save part of the user state.

Download table to .csv file. The Table widget now offers users the possibility to download its current contents to a .csv file on your local machine. This allows you to further process your data in Excel.