ACE

ACE

Decision-making capabilities should be widespread in organizations at any level and enterprise decision management strategy should apply rule-based systems in conjunction with advanced analytics models to automate, improve, and distribute decision-making capabilities across an organization.

Organizations need a new way to map processes and to provide insights at the decision point: that is exactly what our innovative “applications configuration environment” or ACE is all about. It allows the rapid implementation of easy-to-use advanced analytics capabilities to solve complex problems applied to “big data” in order to drive high impact processes and support effective actions throughout the enterprise.

ACE is big data ready offering native connection with appliances (IBM Pure Data Systems for Analytics, SAP Hana and Hadoop), it leverages on predictive modeling state of the art and it has a proprietary workflow engine and real-time scoring engine. Moreover it offers a unique configuration console based on Formula Editor and Wizard Editor to easily create new APPs and adapt exiting ones.

ACE advanced analytics functionalities are available through pillars: predictive, optimizations and forecasting pillars.

ACE

Pillars

pillar

Pillars streamline the analytical process, predictive optimization and forecasting, to support decision-making activities across the enterprise and analytical results industrialization, making easy to produce real-time outcomes. Pillars allow 2 different and complementary approaches to use algorithms:

  • Automated modeling approach. As a business user or domain expert you focus on managing business processes and you should rely on a self-extracting technology able to mine data, apply models and algorithms and return insights and executive summary reports so you can evaluate the drivers influencing your models with a business perspective
  • Advanced user approach. As a data scientist you need to deeply analyze data, work to improve high priority models and algorithms and support your team with very effective predictions and forecasts. Selecting the right modeling approach in an exhaustive list of techniques is a key factor for gaining competitive advantage

Pillars combine the 2 approaches letting business users as well as data scientist work together in the same technological framework; they can work simultaneously on the same data, applying different techniques, simulating different scenario to reach the best outcomes and drive the business forward.

For more info Contact us

Forecasting Pillar

pillar

Forecasting Pillar main functionalities are:

  • Data visualization, exploration and comparison
  • Time series Predictor/drivers selection/upload
  • Data cleaning and manipulation techniques
  • Automation of performance evaluation through business and statistical KPIs (MASE, MAPE, MAE, ME, etc.)
  • State-of-the-art algorithms for forecasting and simulations like linear, non-linear, casual and boosted models (e.g. ARIMAX, regressions, Generalized Additive Models, Exponential Smoothing, K-nearest neighbors, Neural Networks)
  • Business oriented summaries as well as statistical reports

Leading organizations are using Forecasting pillar to forecast demand, sales, managing spare parts, revenues and budgets, creating simulations and what-if scenarios.

For more info Contact us

Optimization Pillar

pillar

Optimization Pillar main functionalities are:

  • Gathering of constrains and parameters data from ACE
  • Target function and optimization model definition into Wizards and Formulas;
  • State-of-the-art techniques for optimization (symplex, genetic and neural algorithms for linear, non linear and mixed integer problems)
  • Optimization and simulation scenario generation and comparison

Leading organizations are using Optimization pillar to create and evaluate scenarios and to optimize decisions.

For more info Contact us

Predictive Pillar

Risk

Predictive Pillar main functionalities are:

  • Data preparation techniques
  • Key predictors calculations and evaluations
  • Drivers and variables selection/upload
  • State of the art algorithms for scoring, clustering, profiling, associations and text mining (decision trees, neural nets, machine learning techniques)
  • Business rules integration
  • Score Matrix to mix up different scoring techniques
  • Business oriented summaries as well as statistical reports

Leading organizations are using Predictive pillar to prevent fraud, reduce risk, predict asset downtime, increase redemption rates for marketing campaigns and anticipate customer’s attrition.

For more info Contact us

Deployer

Deployer

Advanced Analytics projects produce a number of complex scripts, each one plays a role. A typical sequence is: data acquisition, cleaning, dataset preparation, model estimation, prediction/forecast production, forecast and reliability indicators reporting.

Many valuable Advanced Analytics project lay unused because of this human driven deployment strategy, first step in making analytics Actionable is to allow them to be run on scheduled basis.

The application is made of three main functionalities:

  • Scripts and streams are uploaded, default parameters can be set for each one
  • Planning of the execution schema
  • Execution feedbacks for IT  managers and for Analysts

Deployer supports key Analytics tools, such as IBM SPSS Modeler, IBM SPSS Statistics and the Open Source R, very appreciated by the statisticians communities.

Deployer