Predictive Maintenance: Machine Learning vs Rule Based Algorithms

While basic predictive maintenance concepts are discussed in various articles, there is actually little to find when it comes to selecting the best approach on predicting an error. In this article we get you started with a short introduction on predictive maintenance and then focus on which way to go when it comes to choosing the best predictive algorithm for you: Is it better to go with a machine learning model or should you get started with a rule based algorithm first? Let's get started by understanding where we are coming from and what it is all about, we need some context: Predictive Maintenance is basically as old as it gets and in its foundation nothing new. If in the past a mechanics was servicing a machine and found unusual visual or acoustic behaviour in a certain part, the machine may be shut down before breaking and the part was exchanged. That is already predictive maintenance.