An enhanced ride sharing model based on human characteristics, machine learning recommender system, and user threshold time

Exploring non-users' intention to adopt ride-sharing services: Taking into account increased risks due to the COVID-19 pandemic among other factors

“… Guo et al (2021) established an effective real-time ride-sharing framework. Narman et al (2021) designed an enhanced ride-sharing model to increase the usage of ride-sharing services.…”

Cross‐network service recommendation in smart cities

“…Several recommendation approaches have been proposed in various smart city contexts, including smart health care, 7,25 smart markets, 8 smart transportation, 26,27 smart parking, 9,10 ride sharing, 11 e‐governance, 12 smart mobility, 13 smart tourism services, and attractions 14–16 (see Table 2). To achieve the smart service recommendation task, various methods have been adopted, mainly heuristic based and model based.…”

“…To achieve the smart service recommendation task, various methods have been adopted, mainly heuristic based and model based. We distinguish machine learning and multilabel deep learning classification, 11,16 fuzzy logic to handle the decision‐making uncertainty in smart parking, 9 support vector machines (SVM) and Bayesian networks to detect/forecast high concentrations of allergens for route recommendation, 13 SVM to predict riders characteristics, 11 unsupervised natural language processing (NLP) to analyze user‐provided data, 28 in addition to software‐defined networks (SDN), dimensionality reduction, decision tree‐based classification, and convolutional neural networks 7 . Most of these approaches are collaborative filtering‐ and content based, and few ones have adopted a hybrid filtering recommendation (e.g., Reference 14).…”

“…The above methods have exploited different recommendation criteria and data, which are mainly location‐based. We give, as examples, tourists' implicit preferences which are captured from social media photos, 15 sensing context data, 8,14 content‐based information (e.g., visited attractions), 16 social trust (explicit and implicit), 29 ontological concepts (e.g., risk factors, diseases and symptoms, contaminants, weather), 13 mobility and parking data (e.g., distance between parking location and visited place, parking price, number of free spots, electric vehicle charging), 9,30 personalized waiting time for riders and drivers 11 . In addition to sensors and IoT data, other user‐generated data could be useful in the recommendation, such as social networks' data (e.g., comments, social tags, interaction logs) and digital platforms' data (e‐administration, e‐participation) 3 .…”

“…For example, functional requirements include resource management, runtime environment, device mobility, service mobility; while nonfunctional requirements include latency, devices heterogeneity, scalability, reliability, context‐awareness, security, interoperability. The above requirements have not been tackled enough in current smart service recommendation solutions, like in References 7–16. In addition, users (e.g., patients, medical staff, tourists, drivers) are often involved in multiple provider information networks simultaneously, to satisfy their diverse and complex needs, which are translated to a combination of smart services.…”