In recent years many user-oriented and personalized computing technologies have been developed, in which users are immersed in a virtual world and surrounded by processing units. Such computing technologies require distributed signals to be collected, and perform intelligent analysis with data fusion depending on user preferences and the surrounding environment. In such human-oriented analysis, it is also necessary to consider different user preferences and even behavioral factors, that influence the final computing results. Development of user-oriented computing approaches are especially apparent in virtual reality and interactive technologies, multimedia, and decision-making systems, as well as user-oriented security protocols. Such human-oriented protocols allow the intelligent analysis of a great amount of information, perform analytics processes, extract meaning and manage systems in a secure manner. These subjects, as well as a number of others, such as personalized protocols for data analysis and security, computing approaches based on behavioral or perceptual factors, and bio-inspired technologies for knowledge extraction, will form the topics of this Special Issue on “Human oriented solutions for intelligent analysis, multimedia and communication systems” in the journal Concurrency and Computation: Practice and Experience.

For this Special Issue eleven articles of particular interest were selected, which present the most interesting research activities and results within the subject matter of this special issue.

The article “Towards human oriented solutions for deep semantic data analysis” by Ogiela and Snasel,¹ presents novel solutions for efficient semantic analysis of data on the basis of cognitive reasoning and an assessment of marketing preferences registered in the course of the human perception process. Obtaining information from data on the basis of its interpreted meaning, with a view to determine individual preferences, makes it possible to designate the set of features on whose occurrence (or absence) attention is focused and those features whose occurrence has an impact on ‘interest’ within a given piece of information, product, service, and so forth. The approach presented is based on application of cognitive resonance processes implemented in cognitive information systems.

The article entitled “A method to generate context information sets from analysis results with a unified abstraction model based on an extension of data enrichment scheme” by Park et al.² presents studies on a method for processing analysis modules that can enrich result datasets with context information based on a data abstraction model. Data abstraction provides not only capabilities for context-aware systems and users to inspect the context at four levels from raw datasets to situational relationships, but also supports unified context levels for each entity that can be deployed at any location where systems deal with context to provide dedicated services.

The article “Customer-oriented sales modeling strategy in a big data environment” by Chen et al.³ presents a new idea for a data mining technology application in business services. Different factors are analyzed, which may affect the profit of shopping malls in a big data environment and the most critical factors are found by data mining technology. This allows different sales promotion strategies to be provided for merchants to facilitate the expansion of sales. In management, small profits and quick returns is a popular sales strategy used by many shopping malls to increase turnover.

The article “An online cognitive authentication and trust evaluation application programming interface for cognitive security gateway based on distributed massive Internet of Things network” by Chen et al.⁴ presents a new online cognitive authentication and trust evaluation API for CSG based on distributed massive IoT network. An online identity generation API is proposed, together with the modified EPC Class 1 Gen2 tag translator which is used to create both provider's as well as client's online identities.

The article entitled “Dealing with Noise in Crowdsourced GPS Human Trajectory Logging Data” by Adhinugraha et al.⁵ presents new solutions for classifying the noise that might be found from public GPS traces. More than 5300 trajectories that started in the state of Victoria, Australia, were considered, and noise was classified into four types: spike noise, point noise, track noise, and logical noise. The authors tested the behavior of noise when processed with convex hull-based non-map-matching preprocessing methods to reduce spikes, followed by granularity reduction to reduce point density.

In the article “An Effective Architecture of Digital Twin System to Support Human Decision Making and AI-Driven Autonomy” by Mostafa et al.⁶ a data analytic maturity model is presented, which consists of four phases with ordered activities. It shows that any data analytic project needs to be gradually developed from foundations to powerful AI algorithms. The effort and time spent on a routine will create an exponential increase in business value. The digital twin starts in phase two which immediately follows the event that the big data infrastructure is established. It is started by shallowly replicating the characteristics, features and states of its physical twin, and then dives deeper to copy its behaviors, which is achieved by AI technologies, typically machine learning models.

The article “On the Undetectability of Payloads generated through Automatic Tools: a Human-oriented Approach” by Carpentieri et al.⁷ describes tools for the automatic generation of custom executable payloads. Such payloads typically enable to improve the interaction between the security experts and the asset under evaluation. However, due to the actions they take (i.e., remote access, privilege escalation, and so forth), these payloads are most likely classified as malicious by AVs, thus preventing their execution on a system. This article aims to provide the research community with an awareness of the possible security threats that can arise from automated tools commonly used in security assessment processes.

The article entitled “QoS-aware Big Service Composition using Distributed Co-Evolutionary Algorithm” by Dutta et al.⁸ presents an efficient QoS-aware big service composition model using a distributed co-evolutionary algorithm in Spark. In the proposed model the authors designed a distributed NSGA-III for finding the optimal Pareto front and a distributed multi-objective algorithm to compare the solutions of NSGA-III. They also discuss the parallel implementation of distributed co-evolutionary algorithm that makes the algorithms faster and scalable to find the near-optimal solution.

The article entitled “Probability and Topic-Based Data Transmission Protocol” by Saito et al.⁹ presents the MPSFC model to efficiently implement the IoT, where mobile fog nodes such as vehicles communicate with other nodes over wireless networks. Here, each fog node calculates output data on input data received from other fog nodes and forwards the output data to other fog nodes in the epidemic routing way. The authors proposed the TTLBDT and PTBDT protocols and compared them with the TBDT protocol.

The article “Implementation and evaluation of the information flow control for the Internet of Things” by Nakamura et al.¹⁰ describes the OI protocol and evaluates the authorization process of the OI protocol in terms of the execution time. In the evaluation, the authors make clear the features of the execution time of authorization processes for GET, PUT, POST, and DELETE operations in the OI protocol. The OI protocol was implemented on a hybrid device realized in Raspberry Pi 3 Model B+.

The article entitled “Chatbots: Security, Privacy, Data Protection and Social Aspects” by Hasal et al.¹¹ presents all security aspects concerning communication with chatbots. It provides a review describing important steps in chatbot design techniques considering the chatbot security. It also defines possible security threats and vulnerabilities, and presents detailed methods which allow the development of a safe chatbot platform.

We believe that the articles included in this Special Issue will have a great impact for future scientific research, and also contribute to the studies conducted by other researchers and practitioners, who work in the area of advanced information processing systems and computer security.

We would like to express my sincere appreciation of the valuable contributions made by all authors. We'd like also to express our special thanks to Professor David W. Walker from the School of Computer Science and Informatics, Cardiff University, UK, Editor-in-Chief of Concurrency and Computation: Practice and Experience, for allowing the publication of this Special Issue, and for his great support throughout the entire publication process.

近年来，出现了许多面向用户的个性化计算技术，其中用户沉浸在虚拟世界中，被处理单元包围。此类计算技术需要收集分布式信号，并根据用户偏好和周围环境进行数据融合的智能分析。在这种以人为本的分析中，还需要考虑不同的用户偏好甚至行为因素，这些因素会影响最终的计算结果。面向用户的计算方法的发展在虚拟现实和交互技术、多媒体和决策系统以及面向用户的安全协议中尤为明显。这种以人为本的协议允许对大量信息进行智能分析，以安全的方式执行分析流程、提取意义和管理系统。这些主题以及其他一些主题，例如用于数据分析和安全的个性化协议、基于行为或感知因素的计算方法以及用于知识提取的仿生技术，将构成本期“人类”特刊的主题。面向智能分析、多媒体和通信系统的解决方案”，发表在《并发与计算：实践与经验》杂志上。

本期特刊选择了 11 篇特别感兴趣的文章，介绍了本期特刊主题范围内最有趣的研究活动和成果。

Ogiela 和 Snasel 的文章“面向面向人类的深度语义数据分析解决方案” ¹提出了基于认知推理和对人类感知过程中记录的营销偏好评估的数据有效语义分析的新解决方案。根据数据的解释意义从数据中获取信息，以确定个人偏好，可以指定出现（或不存在）注意力集中的特征集，以及出现对“兴趣”有影响的特征。 ' 在给定的信息、产品、服务等中。所提出的方法基于在认知信息系统中实施的认知共振过程的应用。

Park 等人发表的题为“一种基于数据丰富方案扩展的统一抽象模型从分析结果中生成上下文信息集的方法”的文章。^图2展示了一种用于处理分析模块的方法的研究，该方法可以基于数据抽象模型用上下文信息来丰富结果数据集。数据抽象不仅为上下文感知系统和用户提供了从原始数据集到情境关系四个级别检查上下文的能力，而且还支持每个实体的统一上下文级别，这些实体可以部署在系统处理上下文的任何位置以提供专门的服务。

Chen 等人的文章“大数据环境中以客户为导向的销售建模策略”。^{图 3}为数据挖掘技术在商业服务中的应用提供了一种新思路。分析了大数据环境下可能影响商场利润的不同因素，通过数据挖掘技术找出了最关键的因素。这允许为商家提供不同的促销策略以促进销售的扩展。在管理上，薄利多销是许多商场用来增加营业额的流行销售策略。

Chen等人的文章“基于分布式海量物联网网络的认知安全网关的在线认知认证和信任评估应用程序接口”。^{图 4}提出了一种新的基于分布式海量物联网网络的 CSG 在线认知认证和信任评估 API。提出了一种在线身份生成 API，以及用于创建提供商和客户在线身份的修改后的 EPC Class 1 Gen2 标签转换器。

Adhinugraha 等人的题为“处理众包 GPS 人体轨迹记录数据中的噪声”的文章。^{图 5}提出了对可能从公共 GPS 轨迹中发现的噪声进行分类的新解决方案。考虑了始于澳大利亚维多利亚州的 5300 多条轨迹，并将噪声分为四种类型：尖峰噪声、点噪声、轨道噪声和逻辑噪声。作者测试了使用基于凸包的非地图匹配预处理方法处理时噪声的行为以减少尖峰，然后降低粒度以减少点密度。

在 Mostafa 等人的文章“支持人类决策和人工智能驱动的自治的数字孪生系统的有效架构”中。^{图 6 提出}了一个数据分析成熟度模型，该模型由具有有序活动的四个阶段组成。它表明，任何数据分析项目都需要从基础逐步发展到强大的人工智能算法。花在例行公事上的努力和时间将创造商业价值的指数增长。数字孪生在第二阶段开始，紧接在大数据基础设施建立之后。它首先从浅层复制其物理双胞胎的特征、特征和状态开始，然后更深入地复制其行为，这是通过人工智能技术（通常是机器学习模型）实现的。

Carpentieri 等人的文章“关于通过自动工具生成的有效载荷的不可检测性：一种以人为本的方法”。^{图 7}描述了用于自动生成自定义可执行负载的工具。此类有效载荷通常能够改善安全专家与被评估资产之间的交互。但是，由于它们采取的操作（即远程访问、权限提升等），这些有效载荷很可能被 AV 归类为恶意的，从而阻止它们在系统上执行。本文旨在让研究社区了解安全评估过程中常用的自动化工具可能产生的安全威胁。

Dutta 等人的题为“QoS-aware Big Service Composition using Distributed Co-Evolutionary Algorithm”的文章。^{图 8}使用 Spark 中的分布式协同进化算法提出了一种高效的 QoS 感知大服务组合模型。在提出的模型中，作者设计了一个分布式 NSGA-III 来寻找最优帕累托前沿和一个分布式多目标算法来比较 NSGA-III 的解决方案。他们还讨论了分布式协同进化算法的并行实现，该算法使算法更快、可扩展以找到接近最优的解决方案。

Saito 等人的题为“Probability and Topic-Based Data Transmission Protocol”的文章。^{图 9}展示了有效实施物联网的 MPSFC 模型，其中移动雾节点（如车辆）通过无线网络与其他节点进行通信。在这里，每个雾节点根据从其他雾节点接收到的输入数据计算输出数据，并将输出数据以流行病路由方式转发给其他雾节点。作者提出了TTLBDT和PTBDT协议，并与TBDT协议进行了比较。

Nakamura 等人的文章“物联网信息流控制的实现和评估”。^{图 10}描述了 OI 协议，并根据执行时间评估了 OI 协议的授权过程。在评测中，作者明确了OI协议中GET、PUT、POST、DELETE操作的授权流程执行时间的特点。OI 协议是在 Raspberry Pi 3 Model B+ 中实现的混合设备上实现的。

Hasal 等人撰写的题为“聊天机器人：安全、隐私、数据保护和社会方面”的文章。^{图 11}展示了与聊天机器人通信相关的所有安全方面。它提供了一个评论，描述了考虑到聊天机器人安全性的聊天机器人设计技术中的重要步骤。它还定义了可能的安全威胁和漏洞，并提供了允许开发安全聊天机器人平台的详细方法。

我们相信，本期特刊所收录的文章将对未来的科学研究产生重大影响，也有助于其他从事高级信息处理系统和计算机安全领域的研究人员和从业人员进行的研究。

我们衷心感谢所有作者所做的宝贵贡献。我们还要特别感谢来自英国卡迪夫大学计算机科学与信息学学院的 David W. Walker 教授，并发与计算：实践与经验的主编，他允许本刊的出版特刊，并感谢他在整个出版过程中的大力支持。