Designing a Machine Learning Decision Tree for Information Systems: A study into the implementation of supervised and unsupervised machine learning methods

A decision tree for predicting and detecting software breakdowns using machine learning techniques has been created during this research. It provides guidance for implementing suitable machine learning algorithms in information systems. The created decision tree is focused on both the supervised classification and unsupervised anomaly detection algorithms. By going through the decision tree some data shortcomings might be found. Several options to tackle these shortcomings are provided during an elaboration of the decision tree. When the decision tree is followed the practitioner has gained an overview of machine learning requirements that are not yet fulfilled as well as several methods to fulfil the unmet requirements. The application of the decision tree is demonstrated based on the situation of the company’s support department.