A Model for Using Machine Learning in Smart Environments

Interoperability Work Package participated in International Workshop on Self-managing Solutions for Smart Environments (S3E 2011), presenting the paper "A Model for Using Machine Learning in Smart Environments".

The work presents a model for using machine learning in the adaptive control of smart environments. The model is based on an investigation of the existing works regarding smart environments and an analysis of the machine learning uses within them. Four different categories of machine learning in smart environments were identified: prediction, recognition, detection and optimisation. These categories can be deployed to different phases of a self-adaptive application utilising the adaptation loop structure. The use of machine learning in one phase of the adaptation loop was demonstrated by carrying out an experiment utilising neural networks in the prediction of latencies.