Missing value imputation method using Bayesian network for decision-making on HCR

Recently, varied types of robots have been developed in the field of human centered robotics. Human centered robots need to autonomously make their own decisions to live together with human. Accordingly, data mining, especially classification, has been drawing much attention as a component technology to develop decision-making systems. In the real-world data mining, missing value problem is happened, for example, speech containing noise, facial occlusion, and so on. In previous studies, various imputation methods have been developed. Previous imputation methods were developed to solve the missing value problem with lots of explanatory variable, even if some explanatory variables are ineffective for imputation. It has been said that using lots of variable deteriorates in learning efficiency, thus we believe that imputation methods should be developed considering relations among explanatory variables. Therefore we propose the imputation method using Bayesian network. Through the experiments, we can confirmed that proposed method imputes missing values with approximate values, and a classification system successfully classify the test sample, in which missing values are imputed by proposed method, in comparison with some conventional methods.