The optimization is a discipline which is part of mathematics and which aims to model, analyse and solve analytically or numerically problems of minimization or maximization of a function on a specific dataset. Several optimization algorithms are used in systems based on deep learning (DL) such as gradient descent (GD) algorithm. Considering the importance and the efficiency of the GD algorithm, several research works made it possible to improve it and to produce several other variants which also knew great success in DL. This paper presents a comparative study of stochastic, momentum, Nesterov, AdaGrad, RMSProp, AdaDelta, Adam, AdaMax and Nadam gradient descent algorithms based on the speed of convergence of these different algorithms, as well as the mean absolute error of each algorithm in the generation of an optimization solution. The obtained results show that AdaGrad algorithm represents the best performances than the other algorithms with a mean absolute error (MAE) of 0.3858 in 53 iterations and AdaDelta one represents the lowest performances with a MAE of 0.6035 in 6000 iterations. The case study treated in this work is based on an extract of data from the keratoconus dataset of Harvard Dataverse and the results are obtained using Python.

Keratoconus is a noninflammatory disease characterized by thinning and bulging of the cornea, generally appearing during adolescence and slowly progressing, causing vision impairment. However, the detection of keratoconus remains difficult in the early stages of the disease because the patient does not feel any pain. Therefore, the development of a method for detecting this disease based on machine and deep learning methods is necessary for early detection in order to provide the appropriate treatment as early as possible to patients. Thus, the objective of this work is to determine the most relevant parameters with respect to the different classifiers used for keratoconus classification based on the keratoconus dataset of Harvard Dataverse. A total of 446 parameters are analyzed out of 3162 observations by 11 different feature selection algorithms. Obtained results showed that sequential forward selection (SFS) method provided a subset of 10 most relevant variables, thus, generating the highest classification performance by the application of random forest (RF) classifier, with an accuracy of 98% and 95% considering 2 and 4 keratoconus classes, respectively. Found classification accuracy applying RF classifier on the selected variables using SFS method achieves the accuracy obtained using all features of the original dataset.

Efficient traffic management has become a major concern within the framework of smart city projects. However, the increasing complexity of data exchanges and the growing importance of big data makes this task more challenging. Vehicular ad hoc networks (VANETs) face various challenges, including the management of massive data generated by different entities in their environment. In this context, a proposal is put forth for a real-time anomaly detection system with parallel data processing, thereby speeding up data processing. This approach accurately computes vehicle density for each section at any given time, enabling precise traffic management and the provision of information to vehicles regarding traffic density and the safest route to their destination. Furthermore, a machine learning-based prediction system has been developed to mitigate congestion problems and reduce accident risks. Simulations demonstrate that the proposed solution effectively addresses transportation issues while maintaining low latency and high precision.

Rapid urbanization to meet the needs of the growing population has led to several challenges such as pollution, increased and congested traffic, poor sustainability, and impact on the ecological environment. The conception of smart cities comprising intelligent convergence systems has been regarded as a potential solution to overcome these problems. Based on the information, communications, and technology (ICT), the idea of a smart city has emerged to decrease the impact of rapid urbanization. In this context, important efforts have been made for making cities smarter and more sustainable. However, the challenges associated with the implementation and evaluation of smart cities in developing countries are not examined appropriately, particularly in the Moroccan context. To analyze the efficacy and success of such efforts, the evaluation and comparisons using common frameworks are significantly important. For this purpose, the present research aims to investigate and evaluate the most influential dimensions and criteria for smart city development (SCD) in the Moroccan context. To reach this goal, this study proposes a new integrated Multi-Criteria Decision-Making (MCDM) model based on Intuitionistic Fuzzy Analytical Hierarchy Process (IF-AHP) and Intuitionistic Fuzzy Decision-Making Trial and Evaluation Laboratory (IF-DEMATEL). In the given context, the IF-AHP is employed to analyze the structure of the problem and calculate the weights of the qualitative and quantitative dimensions/criteria by incorporating the uncertainty values provided by the experts. Later, IF-DEMATEL is used to construct the structural correlation of dimensions/criteria in MCDM. The use of intuitionistic fuzzy set theory helps in dealing with the linguistic imprecision and the ambiguity of experts’ judgment. Results reveal that ‘Smart Living and Governance’ and ‘Smart Economy’ are major dimensions impacting the SCD in the Moroccan context. The proposed model focuses on enhancing the understanding of different dimensions/criteria and situations in smart cities compared to traditional cities and elevates their decision-making capability. Moreover, the results are discussed, as are the managerial implications, conclusions, limitations, and potential opportunities.