IMPROVEMENT THE QUALITY OF IOT BIG DATA: AN OUTLIER DETECTION APPROACH

Abstract

With the advent of the Internet, many concepts have been introduced in our technological life. One of the common and promising concepts that have attracted research communities is the Internet of Things (IoT). This concept assumes that objects (e.g., devices, sensors, processors, appliances, etc.) around us are connected and communicated with each other as a single network. The quality of data exchanged and its uncertainty are considered the main challenges that face developers when designing IoT models. This is due to the large-scale data generated by network objects that leads to redundancy, noise, and inconsistency in the collected data, which, in turn, yield a variety of issues. Moreover, IoT network is considered heterogeneous since different types of devices and applications are gathered to generate complex-considered data that is difficult to analyze. This data may follow anomalous behavior that leads to having abnormal data points, which impact the quality of data.