The Effect of Variations in Micro-components of Domestic Water Consumption Data on the Classification of Excessive Water Usage

AbstractSeveral studies have been conducted to explore the potential of using consumer data for the estimation of domestic water consumption. However, it is not an easy task to collect and record consumer data that precisely represent daily, weekly or monthly household water usage. This paper investigates the effect of variations in water consumption data on the classification of domestic water usage levels. Two datasets were used in this study. The first dataset consists of ten predictive variables related to household water usage. The second dataset was generated based on the first dataset where four generic features were created to represent water consumption based on four categories of activities related to water usage. Selected classification algorithms were used for classification task. The findings show that variations in consumer data have very little effect on classification outcomes suggesting that data collected from consumer suitable to be used for predicting excessive domestic water usage