Estimation of Missing Data of Showcase Using Artificial Neural Networks

This paper proposes an estimation method of missing data of showcase using Artificial Neural Networks (ANN). When various abnormal conditions such as frost formation in the refrigerators and refrigerant leakage happen, vendors of showcases have to treat the abnormal conditions as quickly as possible. Therefore, various measured data of showcases such as temperatures or pressures of some portions of showcases have to be gathered correctly and symptoms of the abnormal conditions should be estimated in advance. However, in rare cases, there is a possibility to miss to gather the data due to various on-site conditions. The proposed method is applied to missing data of actual showcases. Effectiveness of the proposed method is verified by comparison with a conventional ARMA method.