Predictive Maintenance Modelling Guide

A major problem faced by businesses in asset-heavy industries such as manufacturing is the significant costs associated with delays in the production process due to mechanical problems. Most of these businesses are interested in predicting these problems in advance so that they can proactively fix these issues before they occur which will reduce the costly impact caused by downtime. This collection is a supplement to the [Playbook for Predictive Maintenance][1] which covers the common use cases in predictive maintenance and modelling approaches. The collection only focuses on the data science part of an end-to-end predictive maintenance solution to demonstrate the steps of implementing a predictive model by following the techniques presented in the playbook for a generic scenario that is based on a synthesis of multiple real-world business problems. This example scenario brings together common elements observed among many predictive maintenance use cases. The business problem for this example scenario is about predicting problems caused by component failures such that the question “What is the probability that a machine will fail in the near future due to a failure of a certain component” can be answered. The problem is formatted as a multi-class classification problem and a machine learning algorithm is used to create the predictive model that learns from historical data collected from machines. This collection provides an R notebook and two experiments. - [Predictive Maintenance Modelling Guide Data Sets][2]: The experiment that contains the data sets used in the collection. - [Predictive Maintenance Modelling Guide R Notebook][3]: The R notebook that explains the steps of implementing the solution. - [Predictive Maintenance Modelling Guide Experiment][4]: The experiment that demonstrates the feature engineering, training and evaluation of the predictive model using Azure Machine Learning Studio. It is recommended to follow the above order when examining the collection. As a first step, you will need to open the first experiment in studio that contains the modules that read the data sets used in the R notebook. it is essential that you save these data sets to your workspace before going through the R notebook, you can find the instructions in the experiment description. After you saved the datasets, you can continue with the R Notebook of the collection where feature engineering, labeling, training and evaluation are demonstrated using R language. As a final step, you can use the third experiment that follows the same steps of the R Notebook to feature engineer, label, train and evaluate your models in the Studio. John Ehrlinger ( a Microsoft employee) is a contributor of this collection. [1]: https://azure.microsoft.com/en-us/documentation/articles/cortana-analytics-playbook-predictive-maintenance/ [2]: https://gallery.cortanaintelligence.com/Experiment/Predictive-Maintenance-Implementation-Guide-Data-Sets-1 [3]: https://gallery.cortanaintelligence.com/Notebook/Predictive-Maintenance-Implementation-Guide-R-Notebook-2 [4]: https://gallery.cortanaintelligence.com/Experiment/Predictive-Maintenance-Implementation-Guide-Model-1