NASHRIYYAH -I MUHANDISI -I BARQ VA MUHANDISI -I KAMPYUTAR -I IRAN, B- MUHANDISI -I KAMPYUTAR
Year:2024 | Volume:22 | Issue:4|Papers Count:6

Social networks are not only tools for communication but also represent one of the key potentials in business and commerce. One of the most significant issues in this field is clustering nodes and extracting effective and useful patterns from them, known as community detection. A major challenge in community detection within social networks is the vast number of nodes, which makes any kind of analysis difficult. Another challenge is the overlap of cluster members, referred to as overlapping communities. In such networks, each node may belong to multiple groups. Considering overlaps between communities—especially in large-scale networks—poses significant challenges in accurately detecting and identifying communities. Therefore, many studies tend to overlook this issue. In this paper, an approach is proposed to address these challenges. The most time-consuming step in the proposed algorithm, identifying influential nodes, is performed in parallel. Moreover, overlaps between communities are taken into account and analyzed. The results of evaluating the proposed method, in comparison with other existing methods, indicate its superiority in terms of the uniformity of the detected communities.

Communication networks and the Internet of Things respond to various needs with continuous development. The limitations of size, computing power and energy consumption in Internet of Things devices are the main challenges of this space. This paper emphasizes combining UAVs with edge computing so that this integration provides advanced coverage and efficient computing support, especially in uncertain situations such as incident response. The proposed solution, in terms of the mobility of IoT nodes and with the aim of improving the energy efficiency of the whole system, the problem of UAV path planning and computation offloading is jointly modeled. Then, a mobility- and energy-aware computation offloading and path planning algorithm based on the efficient coverage set of UAVs is presented, which helps to maximize the energy efficiency of the system by using communication and consensus agreements between IoT nodes. The evaluation results show that the proposed method improves energy efficiency by 137% and energy consumption by 28% compared to previous works.

The rapid growth of IoT technology has led to the emergence of various latency-sensitive IoT applications. These applications require significant computational resources for real-time processing, resulting in high energy consumption in IoT devices. To address this issue, task offloading using fog computing has emerged as a novel solution. Fog-based task offloading reduces latency and enhances the flexibility of IoT devices. This study proposes a mathematical model aimed at minimizing end-to-end delay and energy consumption for task offloading in IoT-fog networks based on software-defined networking (SDN) infrastructure. The simulation results of the proposed model are compared with two metaheuristic algorithms (Genetic Algorithm and Firefly Algorithm) and a baseline paper, focusing on delay and energy consumption metrics. After implementing the scenario and conducting analysis, the simulation results indicate that the proposed model, using metaheuristic algorithms, achieved approximate average reductions of 18% in delay and 19% in energy consumption.

Stock market prediction has always been a focus of researchers. Advances in artificial intelligence and machine learning algorithms have enabled the use of textual data alongside numerical data for better stock market forecasting and performance. In this research, to predict the trend of the NewYork Stock Exchange (NYSE) index, numerical data, textual data, and a machine learning model were employed. The model's input includes numerical data as well as the results of sentiment analysis from texts extracted from X (formerly Twitter). Sentiment analysis is performed using a specific machine learning algorithm, Fin-BERT. Additionally, to improve prediction results, prior knowledge of data distribution is incorporated into the cost function of the proposed classifier (SVM). This knowledge is obtained through the calculation of sentiment entropy. Experimental results show that incorporating sentiment entropy into the model's cost function improves prediction performance.

Today, search engines are the gateway to the web. With the increasing popularity of the web, the efforts to exploit it for commercial, social, and political purposes have also increased, making it difficult for search engines to distinguish good content from spam. The concept of web spam was first introduced in 1996 and quickly became recognized as one of the key challenges for the search engine industry. The phenomenon of spam occurs primarily because a significant portion of web page visits comes from search engines, and users tend to check the first search results. The goal of identifying spam pages is to ensure that these pages cannot achieve high rankings using deceptive strategies. Our effort is to provide an effective method for identifying spam pages, thereby reducing the presence of spam in the top search results. In this article, two methods for combating web spam are proposed. The first method, called XGspam, identifies spam pages based on the XGBoost learning algorithm with an accuracy of 94.27%. The second method, named XGSspam, offers a solution to the challenge of imbalanced web data by combining the SMOTE oversampling algorithm with the XGBoost classification model, achieving an accuracy of 95.44% in identifying spam pages.

Deep neural networks are used in the detection of diseases and medical tasks due to their power and capability in extracting complex features and non-linear relationships. Following the emergence of COVID-19, deep learning approaches have been introduced as a powerful approach in diagnosing this disease. In some cases, data mining-based methods cannot definitively diagnose COVID-19 due to their lack of appropriate generalizability on the data. The aim of this research is to propose a solution to improve the diagnostic results in suspicious COVID-19 images. In this study, after diagnosing the disease using two deep networks, GoogleNet and AlexNet, the probability layer of the two learned networks is extracted, and the suspicious cases of the disease are identified. Then, the top features extracted from the two deep learners are combined and sent to a perceptron neural network for the diagnosis of suspicious cases. The extraction of the best features was performed using principal component analysis. The study database includes 224 CT scan images of COVID-19-infected lungs and 522 lung images of healthy individuals, obtained from the GitHub repository. The study results indicate that the aggregation of deep learners in the probability layer can lead to a 98.1% improvement in the diagnosis of COVID-19 in suspicious cases.