Incorporate supervised machine learning into your IIoT solutions. Extends data science practices with automated predictive and prescriptive modeling. No algorithm expertise is required by users. ThingWorx Analytics learns from historical data and uses AI technology to automatically build, validate, and score advanced analytics models capable of making predictions and delivering optimization recommendations as part of your applications.

Better understand your IIoT data. Provides a variety of advanced algorithms that let you discover useful patterns and correlations within your data. Explanatory analysis is not required to generate prediction models, but it can provide insights that will improve the models you generate.

Monitor continuous and cyclic data streams to identify unexpected changes in behavior. Uses machine learning technology to observe and learn the “normal” pattern on a signal (single data stream) and then monitors for anomalies (outliers) on each input stream individually. Developers can easily integrate the output of real-time anomaly detection into a solution that help users or systems to take action when abnormal behavior is detected.

ThingWorx Analytics functionality includes the ability to make predictions based on data in the ThingModel. Relevant outcome-based predictions (time to failure, errors per hour, etc.) are easily added to applications by subscribing to events in the ThingModel. Predictive scoring can be deployed closer to the “thing” via edge technology, rather than in the cloud data.

Improve future performance and results. Automatically execute simulations to generate recommendations to optimize the performance of your product or process. Works as part of ThingWorx Analytics to identify the key factors causing a given outcome, to recommend the factors (“levers”) to change and to prescribe the optimal variable value by how much to improve that outcome. Prescriptive scoring uses prediction models generated by ThingWorx Analytics or equivalent PMML-compliant prediction model generation tools.

Simulate behavior of physical products in the digital world using integrated simulation and other computational providers into your application. Simulation models bring deep product specific knowledge and subject matter expertise. Utilize the knowledge of these models while the product is operating in the real world. Rapidly generate intelligence and insights that can be applied throughout the life cycle, impacting how you service, operate, and improve products.

Pre-built calculations and other transformations, useful in both general IoT application development and as a preparation step in changing raw data into useful insights for predictive analytics training and scoring