Quality of Experience (QoE) is key to Internet of Things (IoT) applications. Indeed, depending on the nature of the application, different application message loss and latency requirements lead to different QoE metrics. In this context, and considering as inspiration the Mean Option Scores (MOS) typically associated with Real Time Communication (RTC), this paper introduces QoE scores that can be used to assess the performance of IoT applications. Taking into account physical, link, network, transport, and application layer features like the transmission rate, network layer loss and latency, the topology and the session layer protocol under consideration, linear regression mechanisms are used to estimate these QoE scores. Some of these features are regular system parameters while others are extracted from standard mechanisms and protocols used for traffic analysis as Key Performance Indicators (KPIs). This model is then trained and evaluated against a large number of Wireless Personal Area Network (WPAN) and Low Power Wide Area Network (LPWAN) architectures.