Confidence-Convergence-Based Sequential Recognition of Underwater Moving Target Behaviors under Limited Observations

This dataset provides the figure source data, generated figures, and plotting/audit scripts associated with the manuscript “Confidence-Convergence-Based Sequential Recognition of Underwater Moving Target Behaviors under Limited Observations”. The data support the simulation-based evaluation of a reliability-aware sequential recognition method for underwater moving target behavior recognition under limited observations. The package includes CSV files for overall method comparison, coverage-risk analysis, observation-window effects, confidence-evolution examples, ablation studies, robustness tests under noise, missing observations and delay, and additional sensitivity tests. It also includes the corresponding publication figures and scripts for figure generation and data-quality auditing. The dataset is intended to support reproducibility of the numerical results and visualizations reported in the manuscript.本数据集提供了与学术论文《基于置信度收敛的有限观测下水下移动目标行为序列识别》（Confidence-Convergence-Based Sequential Recognition of Underwater Moving Target Behaviors under Limited Observations）相关的图表源数据、生成图表以及绘图与审核脚本。该数据集可支撑有限观测条件下水下移动目标行为识别任务中，可靠性感知序列识别方法的仿真评估工作。数据包内含多类CSV文件，分别用于整体方法对比、覆盖-风险分析、观测窗口效应分析、置信度演化示例实验、消融实验、噪声、缺失观测与延迟场景下的鲁棒性测试，以及额外敏感性测试。此外，该数据包还收录了对应已发表的论文图表，以及用于图表生成与数据质量审核的配套脚本。本数据集旨在为该论文中报告的数值结果与可视化内容的可复现性提供支撑。