Return to Blog


New Year



The past year generated a lot of buzz around Industry 4.0, prognostics and health management, artificial intelligence and the Industrial Internet of Things. In 2018, these areas will become even more relevant.

What developments should we expect? How will companies stay competitive? As a predictive analytics company, it is only natural for Predictronics to offer its predictions for the year ahead.

Here is what to look out for in 2018:

1. Lessons Learned

Projects can go wrong for a variety of reasons, including poor problem selection, data quality issues, unrealistic objectives and insufficient knowledge of machine learning and application domains. Those challenges are part of the learning curve – bumps in the road are to be expected when introducing new technologies. In 2018, companies and end-users will work together to learn from past mistakes. They will also use successful examples to improve their own solutions and stay competitive.

2. More Strategic Partnerships

Most companies don’t offer all the necessary capabilities for industrial predictive analytics solutions. Strategic partnerships and collaboration between companies, service providers, end-users and original equipment manufacturers are essential if the goal is to bring effective solutions to the market. Companies that work together in 2018 to complement each other will be the most successful.

3. Models that Learn and Adapt Over Time

Machine and process failures are relatively rare, so it is unlikely one will be left with a complete training set during the initial phase of a project. However, operating conditions change and failures still occur over time. Software platforms and models that quickly learn failure patterns are crucial to industrial applications. Companies that address those issues will emerge as leaders in 2018.

4. Predictive Quality / Process Optimization for Manufacturing Applications

In the past, interest surrounding manufacturing applications was mainly focused on reducing unplanned downtime and preventing unexpected failures. Though this topic is still relevant, interest is now shifting to process health and overall product quality. As this trend develops in 2018, it will incorporate process adjustment and optimization aspects as well.

5. Next Level Vertical Predictive Solutions

The fields of predictive analytics and industrial artificial intelligence are ever-changing and highly competitive. For applications that have out-of-the box vertical solutions, such as rotating machinery health monitoring and industrial robot predictive monitoring, substantial improvements will be made to machine learning, substantial methods, domain specific features and interface design. In 2018, expect companies to take a harder look at existing predictive analytics solutions to see how they can incorporate new features that make their products more accurate and valuable to customers.

David Siegel

About the Author

Dr. David Siegel is a Co-Founder and the Chief Technology Officer for Predictronics Corp. As the project lead for Predictronics, Dr. Siegel’s role includes creating targeted predictive solutions for numerous industrial applications, developing the technological road map for the Predictronics’ clients and their data projects, crafting new machine learning algorithms and methodologies, as well as guiding the Predictronics data science team in the configuration and deployment of various health monitoring and machine learning data solutions. Dr. Siegel has won numerous awards for his more than forty research publications and also for his success in international data challenge competitions. Dr. Siegel has a Doctorate in Mechanical Engineering from the University of Cincinnati, where he served as a research assistant for the NSF I/UCRC Center for Intelligent Maintenance Systems founded by Professor Jay Lee. Connect with him on LinkedIn.

For more information and weekly updates, follow Predictronics:

Know What Happens Next with PDX!