SESSION Practical Application of Machine Learning for APM Track: Reactive to Proactive Decision Making John Slovensky, Asset Answers Product Manager Product Management, Meridium With machine learning technology, companies can automatically organize, qualify, and analyze asset data to ensure accuracy and dynamically give context to asset information and performance. But many organizations are leery about the quality of their own data and the information they’ve collected over the years, which is reinforced by the various ways this information is collected, via spreadsheets, sensor data, paper inputs or any others. Is this data ever really adequately and accurately put to good use? Machine learning, supported by sophisticated algorithms and innovative long text data mining and analysis, can utilize this dark data and provide a full view of asset health and performance. Machine learning algorithms can be trained to study industry-benchmarked data as well as look at new, unstructured text from EAM systems, interrogating and classifying it to provide predictions of whether events qualify as failures or not, and the physical failure mode indicating what failed and how it failed along with confidence factors on its own performance as well. Join us in this session to learn how your organization can easily put a sustainable, automated machine learning framework in place to make asset data credible in a standard model for comparative analysis, reliability analysis and performance improvement recommendations. |