Large and high-dimensional real-world datasets are being gathered across a wide range of application disciplines to enable data-driven decision making. Interactive data visualization can play a critical role in allowing domain experts to select and analyze data from these large collections. However, there is a critical mismatch between the very large number of dimensions in complex real-world datasets and the much smaller number of dimensions that can be concurrently visualized using modern techniques. This gap in dimensionality can result in high levels of selection bias that go unnoticed by users. The bias can in turn threaten the very validity of any subsequent insights. This article describes Adaptive Contextualization (AC), a novel approach to interactive visual data selection that is specifically designed to combat the invisible introduction of selection bias. The AC approach (1) monitors and models a user’s visual data selection activity, (2) computes metrics over that model to quantify the amount of selection bias after each step, (3) visualizes the metric results, and (4) provides interactive tools that help users assess and avoid bias-related problems. This article expands on an earlier article presented at ACM IUI 2016 [16] by providing a more detailed review of the AC methodology and additional evaluation results.
A novel radar target recognition method based on the deep one-dimensional residual-inception network is proposed for a high-resolution range profile (HRRP). The traditional methods based on shallow models can hardly extract the complete information of the targets HRRP from different angles. The deep models, such as sparse autoencoder, have been adopted to solve this problem. However, these deep models with a huge amount of parameters require more training samples to guarantee the generalization performance. To solve the above-mentioned problem, a model based on the one-dimensional convolutional kernel and a pooling layer is proposed. It is worth noting that the one-dimensional convolutional kernel and pooling operation have the potential to overcome the translation sensitivity and target aspect sensitivity of the HRRP, and both of them can greatly reduce the parameters and improve the generalization performance of the model. In addition, a new loss function is proposed to further enhance the separability of features. The experimental results show that compared with other four deep models, the proposed model can achieve a good performance in recognition accuracy and robustness. INDEX TERMS Radar target recognition, high-resolution range profile, one-dimensional residual-inception network, cosine center loss.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.