Comparison of Data-Driven Models for Cleaning eHealth Sensor Data: Use Case on ECG Signal

Electronic Health Records (EHRs) enabled to store and process data recorded by sensors would mean standard-based personalization of medical services and would be a step further to guaranteeing a seamless care access. However, sensor data is subject to several sources of faults and errors which may further lead to imprecise or even incorrect and misleading answers. Thus, it is pivotal to ensure the quality of data collected from e.g. wearable sensors in wireless sensor networks for it to be used in a formal EHR. This article gives comparison of different data-driven models in cleaning eHealth sensor data from wireless sensor networks in order to make sure the data collected is precise and relevant and as such, may be included into a formal EHR. Furthermore, it then suggests optimization of the selected models with the goal of improving their results.