TrendMiner, a Software AG company delivers advanced analytics software to optimize process performance in chemical, petrochemical, oil & gas, pharmaceutical, food & beverage, metals & mining, water & wastewater, and other process manufacturing industries, has introduced machine learning hub (MLHub) as part of its 2022.R2 release.

The new module democratizes machine learning by fostering the creation, training, and deployment of ML models while bridging the gap between operations and analytics teams. The release also includes new context analytics and additional dashboarding features that provide a better view of operational performance.

TrendMiner’s vision to democratize analytics goes beyond its self-service tools. The software provides closer collaboration among a variety of experts to solve operational performance issues. Some require the introduction of (citizen) data scientists who bring specialized techniques and expertise to the table. Keeping data science in the loop allows companies to squeeze the deepest insights out of available data by using advanced statistics and machine learning models.

“After a very successful trial program with notebooks, we are introducing a new industrial MLHub for time-series data,” said Kim Rutten, Machine Learning Product Manager at TrendMiner.

“MLHub extends the analytic and machine learning capabilities of TrendMiner. With MLHub, (citizen) data scientists can access unprocessed, processed, and contextualized data in TrendMiner views and validate hypotheses. They also can create, train, and easily deploy machine learning models using the new notebook environment. Such analysis and its results then can be leveraged by other TrendMiner users through machine learning model tags in TrendHub and advanced and interactive visualizations in DashHub. This allows analytics and data science to become a team sport more than ever before,” added Rutten.

Unlike (big) data modelling tools or AI/ML platforms in the context of production process improvements, MLHub makes all pre-processed, time-series, and contextual operational data available for advanced visualizations and machine learning modelling. MLHub supports quick deployment into operations. It bridges the gap between central analytics teams and operations, which allows for fast iterative improvements.

The models developed in MLHub and deployed in DashHub enable operational experts to address and solve even more complex use cases in areas such as safety improvements, quality control, sustainability, and overall profitability.

Contextualized event data can help identify new areas for performance improvement. TrendMiner provides this from events captured during process monitoring or from data residing in other business applications, such as asset availability data, batch records, lab samples, alerts, and so forth.

The next step in analyzing contextual data is using multivariate scatter plots. This allows you to plot context events and their attributes on a scatter chart. As an additional benefit, insights through correlations and distributions can be extracted.

https://www.softwareag.com/en_corporate/company/connected-enterprise.html