Clinically Inspired Endocrine Signaling Simulation Dataset for ML-Driven Source Localization in Molecular Communication

Description This record contains the simulation data used in the work "Toward Clinically-Inspired Validation of ML-Driven Source Localization in Molecular Communication." The dataset supports a 2-page abstract submitted to the 10th Workshop on Molecular Communications (MolCom 2026), Istanbul, Turkey, 15-17 April 2026. This work extends our prior Zenodo record, DOI: 10.5281/zenodo.15408120. The archive includes simulation runs generated with the MEHLISSA simulator and associated files used for ML-based source localization experiments. Short Description of the Paper Building on our prior work on source localization in the human circulatory system (HCS), this work-in-progress paper extends the framework to a clinically relevant endocrine-signaling setting for tumor-source localization. Specifically, it models endocrine signaling and translates adrenal venous sampling (AVS)-based clinical measurement into MELISSA simulations. The adrenal glands are modeled as hormone (aldosterone and cortisol) secreting transmitters, while the HCS serves as the communication channel and the adrenal veins and inferior vena cava act as passive receivers analogous to clinical AVS setting. Continuous aldosterone and cortisol secretion, together with advection- and diffusion-driven transport, generates receiver-side time-series concentration signals that are used as inputs to the machine leraning (ML) model. The resulting dataset enables validation of our earlier stacked ensemble ML model using LightGBM and BernoulliNB in this clinically grounded setting. Data and Code We publish our simulation data and the Python code to process it as described in the paper here on Zenodo under the CC BY and MIT licenses, respectively. Contact If you have any questions or suggestions for improvements, feel free to contact me. Saswati Pal Email: [e-mail]数据集说明 本存档包含发表于论文《面向分子通信中机器学习驱动源定位的临床导向验证》的仿真数据。本数据集用于支撑提交至2026年4月15日至17日于土耳其伊斯坦布尔举办的第10届分子通信研讨会（MolCom 2026）的2页摘要稿件。本研究基于我们此前的Zenodo存档（DOI: 10.5281/zenodo.15408120）进行拓展。存档内包含使用MEHLISSA仿真器生成的仿真运行数据，以及用于机器学习（Machine Learning, ML）驱动源定位实验的配套文件。 论文简要概述 本阶段性研究成果基于我们此前在人体循环系统（Human Circulatory System, HCS）源定位领域的研究工作，将原框架拓展至与临床相关的内分泌信号场景以实现肿瘤源定位。具体而言，本研究对内分泌信号进行建模，并将基于肾上腺静脉采样（Adrenal Venous Sampling, AVS）的临床测量数据转化为MELISSA仿真场景。研究中将肾上腺建模为分泌激素（醛固酮与皮质醇）的信号发射端，人体循环系统作为通信信道，肾上腺静脉与下腔静脉则作为被动接收端，类比临床AVS检测场景。醛固酮与皮质醇的持续分泌结合对流与扩散驱动的物质输运过程，生成接收端的时序浓度信号，作为机器学习模型的输入数据。本数据集可用于在该临床贴合场景中，验证我们此前采用LightGBM与BernoulliNB构建的堆叠集成机器学习模型。 数据与代码 我们将本论文中提及的仿真数据与处理Python代码发布于Zenodo平台，二者分别采用CC BY与MIT开源许可协议。 联系方式 若您有任何疑问或改进建议，欢迎随时联系。 Saswati Pal 邮箱：[e-mail]