Board Enterprise Analytics Modeling (BEAM)
Make better decisions thanks to better Information and forecasts.
60-70% of companies’ painstakingly collected data goes unused1 – Board’s built-in Business Intelligence Solution BEAM (Board Enterprise Analytics Modelling) remedies the situation and
Enables you to become a data-driven organization. BEAM makes it easy to perform complex analyses, supporting your management decisions with correctly evaluated data from a single point of truth. This gives you valuable insights into your business and enables you to take appropriate action early – easier than ever before.
BEAM is a joint development of Board and IDSIA, a world-leading research institution for artificial intelligence, and solves two previously inadequately addressed requirements at once. First, with the new Advanced and Predictive Analytics module, you can use forecasting, data mining or Big data analytics, directly into your business processes and thus ensure the agility your company needs to needs to operate successfully in the fast-paced economy. Second, using these functions is so simple and intuitive that no in-depth knowledge of statistics is required.
In contrast to traditional data mining and analytics tools, where analytical models must be built manually by experts, business users can work independently with BEAM right from the start. BEAM is enormously flexible, powerful and natively integrated into Board. In combination with its user-friendliness it is a breakthrough solution to effectively embed the extraordinary power of predictive analytics into day-to-day business and decision making.
With its three modules – predictive analytics, clustering and analytic functions – BEAM covers a wide range of analytical disciplines. Alternatively, you can easily integrate external modeling software such as R into Board.
1: Source: Up to 73 Percent of Company Data Goes Unused for Analytics – 2019, Forbes / Barrett Digital
What is predictive analytics?
That’s how easy it is to integrate predictive analytics into analysis, planning and simulation applications!
How does predictive analytics work?
A time series that is identified as continuous goes through an automatic process that aims to model the trend, seasonality and covariate and then find the most applicable forecast model for each time series. The process consists of the following 4 steps:
Preparation of the data series
To improve the accuracy of the forecast, BEAM normalizes the data (“data trimming”). Based on a recursive statistical test, BEAM also removes trends and seasonality from the time series, if necessary.
Defining a training set and a test set
Now each time series is divided into two parts: A training set, to which possible models are applied, and a test set, with which the accuracy of these models is measured.
Selection of the model: the IDSI-ARX® (autoregressive, exogenous) algorithm
Now the forecast model is set. BEAM starts a proprietary automatic forecast procedure called IDSI-. ARX®which was developed by the Swiss research institute IDSIA. IDSI-ARX® is a linear, autoregressive Procedure. It goes from a linear Relationship between the future Values a Variables and the Observations in the past from.
The selection of the model structure is made in a kind of competition of possible models. There are persistent or seasonal-naïve predictors and linear autoregressive models of the ARIMA family. Finally, the model structure which shows the smallest deviation when minimizing the MASE (Mean Absolute Scaled Error).
Detect and exclude outliers
BEAM automatically identifies time series data as outliers if the forecast model shows a significant deviation Contains. This is the case for data points whose error is greater than 3.5 times the standard deviation of the Average. These outliers are later not used by the model for calculation.
What is clustering?
With BEAM you can automatically group similar objects into clusters. For this purpose, the k-Means method applied. You simply select the number of groups and the characteristics (e.g. profitability or turnover) and Board creates the k-Means clusters. These are available to you directly as dimensions for your Business Intelligence and Corporate Performance management applications available in the Board CPM Platform. This seamless integration of the clustering function into the board environment not only enables the direct creation of of reports, graphics or self-service analyses based on clusters, but also the use of the Cluster for Forecasting and Planning. Thanks to the possibility of managing different scenarios, you can create a set of of clustering options and choose the most appropriate one for your current analysis or use the planned business process.
We at Bluebird help you to achieve your goals and are always at your side with our professional know-how as well as with our technical expertise.
We look forward to hearing from you.
Please fill out the following registration form and we will contact you shortly!