Enhanced Context Aware Comfort Level Prediction Model Using Internet of Things

Abstract

Internet of Things is an innovative, promising paradigm that bonds considerable newlineexpertise to identify seamlessly, sense and communicate between things. The Internet newlineof Things (IoT) is the network of connected devices, also called smart devices , newlineenabling the connected devices to communicate and exchange data. IoT enabled newlinesystems are driven by the combinations of sensors, communications, people and newlineprocesses. It plays an imperative role in supporting, coordinating and acting as an newlineeffective tool in predicting the comfort of the environment in every field like Smart newlineHome, Smart City, Smart Hospital, Smart Building, etc. IoT can be used for monitoring, newlineanalyzing and predicting the comfort of the environment. Several environment features newlineaffect individual behavior across the world. Hence, there is a requirement for newlinepredicting the comfort of the environment. newlineMany researchers have contributed many works to environment monitoring newlineusing the Internet of things. However, still there are certain areas for improvement newlineand development of the future environment monitoring needs using IoT. Environment newlinemonitoring is developed using various factors to control and suggest suitable measures newlinefor environment comfort. The complexity arises from various perspectives such as newlinepredicted comfort indices, input features, data sample size, algorithms applied, prediction newlineaccuracy, training and test data. However, prediction model using machine learning newlinealgorithm for environment comfort is a fast developing area and many of the models newlinemay not consider the complex relations between environment comfort and other newlineenvironment factors. newlineMonitoring and prediction based on various environment factors are the most newlinecritical and crucial task of the IoT environment. IoT environment is dynamic and the newlineviii newlineenvironment context varies every moment. Therefore, it is challenging to monitor the newlineenvironment factors at affordable cost with reliability to achieve higher model prediction newlineaccuracy. This research work proposes a framework to