A Probabilistic Approach to Handle Missing Data for Multi-Sensory Activity Recognition

Context and activity recognition in complex scenarios is prone to data loss due to disconnections, sensor failure, transmission problems, etc. This generally implies significant changes in the recognition performance. In the case of classifier fusion faulty sensors can be removed from the recognition chain to overcome this issue. Alternatively, we can try to compensate or impute data to replace the missing signals. In this paper we proposed a probabilistic method for imputation of missing data. The proposed method is based on conditional Gaussian distribution and has been previously applied in other fields, such as speech recognition and bioinformatics, but not in for activity recognition. Our method exploits the correlation among classifier outputs to infer missing values of decision profile from available values in a probabilistic manner. We assess the method performance using two datasets in a car manufacturing and in a daily activities scenario with three different configuration of sensors. Results show the advantages of the probabilistic estimation over other common methods such as removing and clustering. The method is also applicable in other classification problems which uses fusion methods to combine decisions of classifiers.