Controlling in-patient environment by mining sensor data

This paper demonstrates the application of data mining on sensor data in order to develop a classification method for determination of inpatient´s light and thermal comfort preferences. Hierarchical clustering along with incremental learning is proposed to develop an intelligent approach to control the temperature and light usage for each patient according to their comfort. By using the incremental learning the clustering model is adoptable to new usage patterns. For example, as new usage patterns are observed they are incrementally learned by the model and grouped into appropriate cluster. Application of two clustering techniques named Hierarchical and SimpleK-Means are evaluated on real dataset collected from medical unit to investigate the appropriate method for the classification. The results shows that Hierarchical clustering outperforms SimpleK-Means in term of number of clusters, cluster size, accuracy, update time and change rate. The proposed approach is evaluated on 10 days data which shows that the energy consumption is decreased by 11.32% and patient comfort level is increased by 16.74%.