Federated learning (FL) represents a transformative shift in machine learning, moving from conventional centralized approaches to a distributed framework that emphasizes data privacy and security. The FL server transmits an initial model to clients, which they train locally on their private data. After training, the server aggregates model updates from each client to update the global model. Selecting the best clients in FL is critical to improving the convergence speed and accuracy of the final model, which requires careful client selection approaches. The client selection phase of FL faces numerous challenges that impact overall training performance, including statistical heterogeneity and system heterogeneity resulting from the diversity of client data and resources. Communication costs present another challenge, especially in networks with limited client communication resources. Additionally, selecting trustworthy clients represents another challenge, as selecting malicious clients creates a significant risk within the FL training process. Moreover, the fairness challenge entails providing fair opportunities for all clients to participate in training. To address these challenges, we offer solutions that utilize effective techniques, approaches, and client selection methods. This survey presents a taxonomy of modern client selection methods in FL, highlighting the improvements in FL performance and effectiveness achieved through these methods, including greedy selection, reinforcement learning-based selection, multi-armed bandit-based selection, clustering-based selection, and reputation & security-based selection. Subsequently, it offers a general comparison between these methods in terms of their core ideas, advantages, limitations, and use cases. Finally, future and potential trends in client selection and improving performance in FL are identified.

Mobile ad hoc network is a creation of portable devices; they create a network with infrastructure on the fly that is ever changing. As in the previous network structure, each node performs both the action of a router and a host. Also, nodes cannot be fixed in the network; they can join or leave the network, and this increases the flexibility of the connectivity. There are routing protocols which are used for identification of efficient paths between the nodes in the network so as to seek the determination of the best routes between two nodes. This research show that routing is complex in MANETs and hence it demands the fine tuning of numerous routing protocols. We evaluate the effectiveness of these protocols by analyzing two primary metrics: are the average figures of throughput and average end to end delay. Simulation of this protocol was done using NS2 (Network Simulator) 2. 35, we investigate how well routing protocols fare in terms of different aspects including Size of the packets and number of nodes present.

The main access to modern healthcare using artificial technologies is related to the medical topic and prediction of risks to human health. Enhance patient medical care using intelligent prediction models such as machine learning, like gradient boosting trees, supervised machine learning and logistic regression which have a great importance in detecting diseases by analyzing medical images and diagnosing chronic diseases, in addition to use deep learning models like deep neural networks, recurrent neural networks, and long short term memory to predict many disease like depression risk, lung cancer, heart diabetic and kidney diseases. Enhance healthcare provider insights using intelligent prediction models to predict future health conditions, treatment outcomes, and disease progression. Moreover, the contribution of the intelligent prediction model helps the healthcare professional identify potential risks and intervene proactively by analyzing patient historical data for early diseases detection like using. Ultimately, the combination of medical history, intelligent prediction, and healthcare data analysis will empower healthcare providers with valuable tools to improve patient outcomes in efficient healthcare organizations.

contributing to the rate of water waste is overflow from overhead tanks. A crucial component that lessens human intervention in preventing such circumstances is the use of an automatic water level control mechanism. In this research, we suggest an intelligent system to regulate the water level and the process of turning on and off. The water level is detected by this method using an ultrasonic sensor and an Arduino. The water level is calculated in percentage based on the sound the water makes, and the result is displayed on the LCD screen. The water level is computed up to 100%. The relay switch attached to this system will automatically turn on and off when the water level exceeds the preset algorithmic value, which is then displayed on the LCD screen. When the water level falls to 0%, it will activate automatically, and once it reaches 100%, it will switch off. Additionally, we will observe that the pump will not operate in the absence of a water source to fill the tank since air will enter the pump and cause damage if it runs without water. Under these circumstances, the LCD panel will display the words "dry protection," which refers to the condition. Given that it would make it simpler for the user to fill the tank without assistance, the results showed that the automatic water level control device employing ultrasonic technology was effective based on readings that were practically monitored during the project's operation.

Diagnosing diseases in humans is the first step in treating diseases, and knowing it is important to determine treatment and deal with the disease in the correct way. Diagnosis is made in medical institutions using available tools and specialists in each medical field to determine the problem presented by the patient. Modeling and analysis of medical data is important in healthcare and social applications in areas related to disease prediction and diagnosis. The model selection strategy is an important determinant of the performance and acceptance of a medical diagnostic decision support system. This paper proposes a stacked learning model derived from multiple ensembles learning algorithms, including Random Forest, Catboost and XGBoost. To determine the effectiveness of the model, it was tested using eight data sets covering different diseases to help make disease diagnosis decisions. The results show that the proposed model generally outperforms individual machine learning models in terms of accuracy