This study presents a machine learning framework for predicting pedestrian accident severity using Amman, Jordan's first complete 10-year traffic dataset (2014–2023). Addressing the critical class imbalance where minor injuries predominate (85%), causing standard models to poorly detect severe cases (< 25% recall), we implement cost-sensitive algorithms and specialized undersampling techniques, such as XGBoost with Balancing the Loss Function (XGBLF) and Random Data Partitioning with Voting Rule (RDPVR), which enhanced learning from underrepresented Major/Fatal cases while maintaining data authenticity. Through mixed-type correlation analysis and statistical testing, vehicle speed, road illumination, vehicle type, driver age, and road conditions emerged as the most significant predictive factors. RDPVR achieved a 63% true positive rate for Major/Fatal injuries, a 2.78-fold improvement over standard classifiers, and XGBLF achieved 95%, but this achievement was on account of the accuracy of the minor cases. Comprehensive interpretability analysis (SHAP, LIME, and Permutation Importance) revealed that heavy vehicles, poor lighting, and high-speed driving strongly predict Major/Fatal outcomes. Notably, the analysis demonstrates Jordan's improved safety trajectory, with 2020–2023 showing reduced severe accidents compared to 2014–2016, indicating measurable policy impact. This study delivers the first interpretable, context-sensitive AI framework for Amman/Jordan pedestrian safety, translating technical insights into actionable recommendations for targeted interventions, urban planning, and data-driven enforcement strategies to reduce pedestrian injury severity in high-risk zones.

Air traffic has seen a steady increase in the number of operations in recent years. This increase was not accompanied by an expansion of the systems capacity to continue operations without significant impacts on the performance and quality of transport services. It is very difficult to expand the airport infrastructure in todays conditions, so it is necessary to optimize the existing operations to increase the capacity and efficiency of the system. The paper focused on determining the correlation of individual key indicators of airport performance and their impact on delays, which was observed as an output variable. Three machine learning regression methods were applied, Decision Tree, Random Forrest and Support Vector Regression to predict delays with respect to other characteristics of traffic and surrounding airspace. All three methods were able to predict a delay with an error of 60s in at least 91% of the test set, which shows a clear potential for the use of machine learning in this domain.

Machine learning (ML) is a crucial component of artificial intelligence that has recently attracted attention for its application in logistics. ML algorithms are used on large datasets. They create logic correlations among given data and provide predictions of specific values. This research paper aims to conduct a systematic literature review to showcase the potential applications of machine learning in urban logistics systems, specifically focusing on enhancing satisfaction for postal logistics operators and their customers. The authors used various research publication databases in this context (Web of Science, Scopus, Google Scholar etc). The analysis of different models provides insights into diverse aspects, such as predicting product prices and types of cargo, evaluating user satisfaction, forecasting user departures, assessing optimal geographical locations for implementing postal centres, predicting purchase times before online orders, estimating delivery times in the last phase of the logistics chain and more. The significance of this research is highlighted through the identification of shortcomings in existing literature, offering guidelines for future research in developing new machine learning model for optimal operator selection. This model aims to achieve improvements in both customer and operator satisfaction simultaneously.

In the last few years, we have been active participants in globalization and the rapid growth of online commerce. This trend directly impacts logistics, postal, and courier companies. Challenges are visible in the growth and development of these companies, as well as quality of task execution and the ability to adapt to market conditions. With the impact of the SARS-COV-19 virus on the market, and due to locks during the pandemic, shoppers were forced to buy products online. This effect accelerated the trend of purchases via e-commerce, which increased the number of parcels in the technological phases of pick-up and delivery. These companies are the bearers of the first and last technological phase of shipment transport. Due to the increase in the number of shipments, postal and logistics companies are under great pressure. This is because of certain restrictions introduced during the pandemic, but also because of costs, especially fixed costs in the first and last phase of transport. This paper proposes an outsourcing model for the pick-up and delivery to convert fixed costs into variable ones.

Background: The growth of parcel volumes in urban areas, largely driven by e-commerce, has increased the complexity of pickup and delivery operations. To meet demands for cost efficiency, flexibility, and sustainability, CEP (Courier, Express, and Parcel) operators increasingly outsource segments of their last-mile networks. Methods: This study proposes a novel multidimensional cost model for outsourcing, integrating five key variables: transport unit type (parcel/pallet), service phase (pickup/delivery), vehicle category, powertrain type, and delivery point type. The model applies correction coefficients based on internal operational costs, further adjusted for location and service quality using a bonus/malus mechanism. Results: Each cost component is calculated independently, enabling full transparency and route-level cost tracking. A real-world case study was conducted using operational data from a CEP operator in Bosnia and Herzegovina. The model demonstrated improved accuracy and fairness in cost allocation, with measurable savings of up to 7% compared to existing fixed-price models. Conclusions: The proposed model supports data-driven outsourcing decisions, allows tailored cost structuring based on operational realities, and aligns with sustainable last-mile delivery strategies. It offers a scalable and adaptable tool for CEP operators seeking to enhance cost control and service efficiency in complex urban environments.