Understanding how deep hierarchical models build their knowledge is a key issue in the usage of artificial intelligence to interpret the reality behind data. Depending on the discipline and models used, such knowledge may be represented in ways that are more or less intelligible for humans, limiting further improvements on the performance of the existing models. In order to delve into the characterization and modeling of volcano-seismic signals, this article emphasizes the idea of deciphering what and how recurrent neural networks (RNNs) model, and how this knowledge can be used to improve data interpretation. The key to accomplishing these objectives is both analyzing the hidden state dynamics associated with their hidden units as well as pruning/trimming based on the specialization of neurons. In this article, we process, analyze, and visualize the hidden states activation maps of two RNN architectures when managing different types of volcano-seismic events. As a result, the class-dependent discriminative behavior of most active neurons is analyzed, thereby increasing the comprehension of the detection and classification tasks. A representative dataset from the deception island volcano (Antarctica), containing volcano-tectonic earthquakes, long period events, volcanic tremors, and hybrid events, is used to train the models. Experimental analysis shows how neural activity and its associated specialization skills change depending on the architecture chosen and the type of event analyzed.

Purpose: Port throughput analysis is a challenging task, as it consists of intertwined interactions between a variety of cargos and numerous influencing factors. This study aims to propose a quantitative method to facilitate port throughput analysis by identification of important cargos and key macroeconomic variables. Design/methodology/approach: Mutual information is applied to measure the linear and nonlinear correlation among variables. The method gives a unique measure of dependence between two variables by quantifying the amount of information held in one variable through another variable. Findings: This study uses the mutual information to the Port of Isafjordur in Iceland to underpin the port throughput analysis. The results show that marine products are the main export cargo, whereas most imports are fuel oil, industrial materials and marine product. The aggregation of these cargos, handled in the port, meaningfully determines the non-containerized port throughput. The relation between non-containerized export and the national gross domestic product (GDP) is relatively high. However, non-containerized import is mostly related to the world GDP. The non-containerized throughput shows a strong relation to the national GDP. Furthermore, the results reveal that the volume of national export trade is the key influencing macroeconomic variable to the containerized throughput. Originality/value: Application of the mutual information in port throughput analysis effectively reduces epistemic uncertainty in the identification of important cargos and key influencing macroeconomic variables. Thus, it increases the reliability of the port throughput forecast.