Data Mining Techniques in IoT Knowledge Discovery: A Survey

IoT is a buzzword nowadays and of course, it should be. The widespread of electronic and electromechanical devices with connecting ability to the Internet makes IoT be dominant from the user, manufacturer and services/goods provider perspective. Via IoT, the status of almost anything can be tracked, configured and maintained by different computing techniques using user devices or remotely from server ends. Determination of status can be easily known with data mining techniques that follow a distinct ladder until the representation of knowledge. In this survey work, we examined articles published from 2010 to date in the area of IoT. We followed a systematic literature review approach and scrutinize the different data mining steps followed by various scholars, and further classify the data mining techniques used in IoT as a conventional and non-conventional approach. Data cleaning, regression, model visualization, and summarization techniques were considered as challenging tasks due to the nature of IoT settings. This in turn demanded a new direction of research so as to come up with enhanced service provision in the area of IoT. Overlooked data mining techniques and comparison of the different approaches were criticized and reported. Moreover, the interdependency of IoT technologies with data mining approaches is discussed. Ultimately, an attempt has been made to indicate the research trend of IoT.