Abstract:
This work focuses on interoperability issues, where we propose a novel approach to deal with the interfacing of internet of things devices, which is supposed to alleviate the interoperability issue. Based on Natural Language Processing and Machine Learning techniques, the proposed solution aims to classify API methods in order to automatically identify the specific action related to each method without any prior knowledge. This contribution may be an important piece of a larger solution that ensures the interoperability of IoT devices without imposing anything to the different IoT stakeholders. The proposed technique can be also used to detect the capabilities and the available physical interfaces of any device, which can also can help in improving the service recommendation techniques.
