The proliferation of connected vehicles and the advent of New Radio (NR) technologies have ushered in a new era of intelligent transportation systems. Ensuring reliable and lowlatency communication between vehicles and their surrounding environment is of utmost importance for the success of these systems. This paper presents a novel approach to predict Quality of Service (QoS) in Vehicle-to-Everything (V2X) communications through nested cross-validation. Our methodology employs several machine learning (ML) methods to predict some QoS metrics, such as packet delivery ratio (PDR), and throughput, in NR-based V2X scenarios. In ML employment, nested cross-validation approach, unlike conventional cross-validation approach, prevents information leakage from parameter selection into hyperparameter selection, and this results in getting more robust results in terms of overfitting. The study utilizes real-world NR-V2X datasets to train and validate the proposed ML methods. Through extensive experiments, we demonstrate the efficacy of our approach in accurately predicting QoS parameters, even in dynamic and challenging vehicular environments. In summary, our research contributes to the advancement of NR-based V2X communication systems by introducing employment of ML methods with a novel approach for QoS prediction. The combination of accurate predictions through nested cross-validation not only enhances the reliability of communication in connected vehicles' landscape but also has a supportive role for stakeholders to make informed decisions for the optimization and management of vehicular networks.

翻译：连接车辆的普及和新无线电（NR）技术的出现，开启了智能交通系统的新纪元。确保车辆与其周围环境之间可靠且低延迟的通信，对于这些系统的成功至关重要。本文提出了一种通过嵌套交叉验证预测车辆与万物（V2X）通信中服务质量（QoS）的新方法。我们的方法采用多种机器学习（ML）方法，在基于NR的V2X场景中预测某些QoS指标，例如数据包传递率（PDR）和吞吐量。在机器学习应用中，嵌套交叉验证方法不同于传统的交叉验证方法，它能够防止参数选择向超参数选择的信息泄露，从而在过拟合方面获得更稳健的结果。本研究利用真实的NR-V2X数据集来训练和验证所提出的机器学习方法。通过大量实验，我们证明了该方法即使在动态且具有挑战性的车载环境中也能准确预测QoS参数的有效性。总之，我们的研究通过引入机器学习方法并结合一种新颖的QoS预测方法，为基于NR的V2X通信系统的进步做出了贡献。通过嵌套交叉验证实现的准确预测，不仅增强了连接车辆通信的可靠性，而且为利益相关者做出优化和管理车载网络的明智决策提供了支持作用。