Visualization literacy is increasingly critical for interpreting and making decisions from data-driven visual representations, yet existing assessments often emphasize low-level decoding and value lookup and undersample mid- to high-level visual-analytic reasoning. We introduce CoViLT, the Cognitively-Ordered Visualization Literacy Test. This task-grounded framework operationalizes Bloom’s revised taxonomy through observable visualization tasks aligned with Brehmer and Munzner’s typology (Retrieve/Identify, Compute/Derive, Compare/Detect, Judge/Decide). CoViLT includes 48 dichotomously scored multiple-choice items (possible total score range: 0–48), balanced across four cognitive levels and 12 visualization types. Content validity was confirmed by 20 experts (CVR threshold = 0.42; all items met the threshold). Of the 187 undergraduates who started the pilot, six did not complete the test and 14 were excluded because they combined item-response times under 10 s with an invariant response-position pattern (consistently selecting either the first or the fifth option), yielding a final analytic sample of 167. Item difficulty ranged from 0.16 to 1.00 (M = 0.76), and discrimination ranged from 0.000 to 0.449 (M = 0.224), with α =.735 (SEM = 2.47). These pilot findings provide preliminary support for using CoViLT’s total score in group-level research on visual-analytic reasoning; further validation is required before individual-level or high-stakes use.
Time series with rich payloads (TSRP) are sequences of irregularly sampled, high-dimensional observations in which temporal alignment is informative but sample sizes are often limited. Such data arise in numerous domains, yet existing classification methods are ill-suited for these challenges. This setting falls outside the assumptions of most existing time-series classifiers, which rely on dense numerical observations or large training samples. We present a framework for classifying TSRP that integrates representation learning with sparse dynamic time warping (DTW) and support vector machines (SVMs). High-dimensional payloads are first embedded into lower-dimensional representations, after which sparse DTW aligns sequences for SVM-based classification. Our approach is evaluated on six datasets derived from social media, capturing medically relevant transitions such as shifts in mental health status. Results demonstrate that the proposed framework substantially outperforms a baseline based of average pooling across five of six datasets. We further analyze performance relative to dataset characteristics, provide insights into model behavior, and introduce an interpretable mechanism linking temporal alignments to input features. Overall, this work establishes a general, interpretable, and effective approach for classifying sparse time series with rich payloads, providing a foundation for applications in healthcare and other high-stakes domains.
This study proposes an integrated intuitionistic fuzzy group decision-making framework for the early quality assessment of newly discovered meteorites. To address the challenges of expert weighting and data uncertainty, the framework introduces a novel normalized projection measure and a dual-mechanism quality evaluation system combining inversion-based consistency and entropy-based uncertainty quantification. An adaptive expert weighting scheme is developed using a dynamic compromise coefficient. Empirical analysis based on an illustrative case study demonstrates the potential advantages of the proposed framework: the median-based data center shows improved expert differentiation compared with the arithmetic mean; the comprehensive measure exhibits enhanced fairness and discriminative ability relative to inversion-only and entropy-only approaches, respectively; and the model demonstrates notable robustness, with stable expert rankings across a substantial range of parameter variations and invariant final meteorite rankings. These findings, derived from the case study presented, suggest that the proposed approach provides a reliable and theoretically grounded methodology for preliminary meteorite evaluation, with extensible applications in management and quality-assured group decision-making. Further validation on additional datasets is recommended to establish the generalizability of the quantitative findings.
In recent years, multilayer networks have become a hot research topic in the field of complex networks. Extracting shared community structures from multilayer networks is crucial for analyzing complex multilayer network systems. However, existing methods either merge multilayer networks into single-layer weighted networks or extend single-layer network algorithms using consensus clustering approaches; these methods suffer from limited accuracy due to their neglect of inter-layer structural correlations. To address this issue, this paper proposes a multilayer network community detection model based on cross-layer coupled random walks. The proposed method defines a multilayer stochastic transition process through three structure-aware mechanisms: (1) a node-adaptive starting-layer selection strategy based on node activity and inter-layer structural correlation; (2) a multiplicatively coupled intra-layer transition mechanism that jointly considers triangle-motif support and cross-layer first-order connection persistence; and (3) a node-conditioned inter-layer transition mechanism guided by cross-layer neighborhood similarity. The resulting random-walk sequences are subsequently used by Skip-gram to learn node representations, followed by K-means++ for community detection. Experimental results on synthetic and real-world networks demonstrate that the proposed algorithm outperforms existing state-of-the-art methods in multilayer network community detection. Case studies further confirm that the communities identified in the human-HIV-1 interaction network exhibit significant biological functional consistency.
In intelligent information systems that rely on explicit-feedback recommendation, observed ratings are noisy, subjective proxies for user preference. Most prior work on natural noise in collaborative filtering handles this problem through data cleansing. That strategy separates noise handling from model training and may discard informative observations. This paper instead studies robust learning directly in the training objective for explicit-rating prediction. We reformulate rating prediction as a multiclass problem under label noise. We adapt Forward Correction and Backward Correction to collaborative filtering and introduce an anchor-based heuristic for estimating the noise transition matrix when the corruption process is unknown. The approach is evaluated on MovieLens 100k, CiaoDVD, and Yahoo! Music under controlled uniform and pairwise corruption. It compares corrected losses with cleansing methods, robust-learning baselines, and classical collaborative-filtering predictors in a shared explicit-rating setting. Across datasets and corruption levels, Forward Correction with the proposed estimator delivers the most consistent balance between predictive quality and robustness. The estimated matrix also appears to capture part of the noise the datasets already carry, in addition to the injected corruption, although that evidence is necessarily indirect: no public explicit-feedback dataset annotates which ratings are inconsistent. These findings suggest that, for explicit-rating prediction under controlled corruption, robust learning is a strong alternative to pre-processing-centered noise handling in collaborative filtering.
