A method for missing data recovery of waste gas monitoring in animal building based on GA-SVM

Abstract

In order to solve the data missing problem caused by sensor faults during the waste gas monitoring in animal building, a method for missing data recovery was presented based on support vector machine (SVM) combined with genetic algorithm (GA). Multiple factors that influence monitoring values of the waste gas in animal building such as temporal, spatial and environmental, were considered to established a SVM regression prediction model to estimate the missing data of the waste gas monitoring. Meanwhile, to obtain better prediction accuracy, model parameters were optimized by the GA. The data processing of the ammonia (NH3) concentration was taken as an example; monitoring data of 3 days were randomly selected in a farm to test the presented model in this paper. It is shown that there was a very little error between the estimated data and the monitoring data, the maximal relative error was 6.99 % (percent), and the average relative error was 2.15 % (percent). It is an effective method for missing data recovery and a practical way of data processing for waste gas monitoring in animal building.