An integrated approach for weather forecasting over Internet of Things: A brief review

Weather forecasting has been major challenge due to uncertain nature of climatic conditions. Due to advancement in domain of Internet of Things (IoT), weather monitoring has been done with help of low powered sensor devices. Need has come to make central node or server efficient in terms of memory and power. Efforts are being put to on those algorithms which are less prone to prediction errors. Neural networks algorithms are not affordable to work on IoT devices. This paper focuses on Structural as well as Time Series based weather forecasting algorithms which are Linear Regression, Multiple Linear regression (MLR), Support Vector Regression (SVR) as well as Auto Regressive Integrated Moving Average (ARIMA). The predictive ability of each of these algorithms is computed using Root Mean Square Error which indicates ARIMA as the best prediction model.