Abstract
With the fluctuations of energy pricing, the electricity bills for many households is becoming an increasing burden in the monthly bills to be paid. Many households in many countries would appreciate Smart Electricity Monitors and Smart Electricity controllers that can be used both in Smart Cities and regular households. Weather conditions play a vital role in electricity consumption, for example, in a hot desert climate, the electricity bill is tripled or quadrupled due to the use of air conditioning in summer. In this paper, we propose a machine learning based technique to anticipate the electricity consumption using weather data. Using a dataset consisting of reading over 350 days of household appliances with weather condition, the proposed method is able to predict the electricity consumption with a correlation coefficient of 75.7 % using Random Tree. The results obtained provide an excellent basis for the ultimate goal of setting up an accurate smart electricity consumption monitor to be used both in Smart Cities and even in regular households which could ultimately help households to reduce electricity by regulating the use of the most electricity consuming appliances according to weather conditions.