Prediction of the future behavior of the stock market has always attracted researchers' attention as an important challenge in the field of machine learning. In recent years deep learning methods have been successfully applied in this domain to improve prediction performance. Previous studies have demonstrated that aggregating information from related stocks can improve the performance of prediction. However, the capacity of modeling the stocks relations as directed graphs and the power of sophisticated graph embedding techniques such as Graph Attention Networks have not been exploited so far for prediction in this domain. In this work, we introduce a framework called DeepNet that creates a directed graph representing how useful the data from each stock can be for improving the prediction accuracy of any other stocks. DeepNet then applies Graph Attention Network to extract a useful representation for each node by aggregating information from its neighbors, while the optimal amount of each neighbor's contribution is learned during the training phase. We have developed a novel Graph Attention Network model called DGAT that is able to define unequal contribution values for each pair of adjacent nodes in a directed graph. Our evaluation experiments on the Tehran Stock Exchange data show that the introduced prediction model outperforms the state-of-the-art baseline algorithms in terms of accuracy and MCC measures.

Detecting fake news on social media platforms remains a significant challenge due to the dynamic nature of these networks, evolving user-news relationships, the difficulty in distinguishing real from fake information, and the use of advanced generative models to create fake content. In this study, we propose a novel approach, the Dynamic Graph Attention Network (DynGAT), for effective fake news detection. The DynGAT model utilizes the dynamic graph structure of social networks to capture the evolving interactions between users and news sources. It includes a graph construction module that updates the graph based on temporal data and a graph attention module that assigns importance to nodes and edges within the graph. The model applies attention mechanisms to prioritize critical interactions and uses deep learning techniques to classify news articles as real or fake. Experimental results on the TweepFake dataset (20,712 samples) show that DynGAT achieves 95% accuracy, outperforming existing methods such as Static GNN (87%), Transformer-based models (91%), and Hybrid models (89%). The model also demonstrates improvements in precision, recall, and F1 score. This work contributes to the ongoing efforts to combat misinformation and promote reliable information on social media platforms.

One of the standard criteria for expressing the relationship between two random variables is the correlation coefficient. Correlation between variables shows that changing the value of one variable leads to changing another variable in a certain direction. It is also possible to use the value of one variable to predict the value of another. In statistics, the correlation coefficient measures the direction and strength of the tendency to change. In machine learning, the correlation coefficient is known as a measure of classification quality. In fact, as a starting step for classification, the correlation between different samples should be estimated using a specific method. There are various methods to estimate the correlation of different data types, which have disadvantages such as low accuracy or high computational time. One of the methods that can overcome these problems, due to its high capability in modeling correlation between samples is graphical modeling. In this research, a new covariance model based on graph theory and graph neural network for estimating the correlation between samples is presented. The results show the improvement of the proposed model in accuracy, sensitivity, precision, F-Micro, F-Macro and statistical tests compared to Pearson and cosine methods.

Urban bus transportation is one of the common public transportation modes, which has a significant position among all other modes. Any attempt in order to increase the effectiveness and desirability of this mode has a positive influence on the urban trips. Bus network design due to its considerable influence on the performance of this mode is of great importance. And various approaches have been proposed in this context, which have different advantages and disadvantages. Reaching the optimal network in the shortest possible time, considering the parameters which are in conflict with each other most of the time, is the main reason of the complexity of bus network design problem, and makes constraints for the designers. The present paper proposes a heuristic approach for solving bus network design, which is based on graph theory. The proposed methodology designs bus network in two steps of route generating and route selection. For the evaluation purpose, this method compared to other heuristic methods. The results showed that the proposed method outperforms the other existing heuristic network design methods.

Kidney stones are solid crystals made of minerals and salts that form within the kidney, often creating a sharp, hard mass. These stones can block urine flow as they move into the urinary tract, making early detection crucial. Although deep neural networks (DNNs) have been used to diagnose kidney stones with some success, they still face performance and standardization issues. A new approach combines graph convolutional networks (GCNs) with DNNs to address these challenges. This method extracts orb features from images, converts them into graphs, and embeds nodes using a graph convolutional network, which includes a message-passing layer and node feature aggregation. The GCN updates node properties, enhancing efficiency and performance when integrated into a deep network. This approach enables more comprehensive and precise feature extraction from images, improving kidney stone diagnosis. The study highlights GCNs' potential in analyzing medical images for diagnosing kidney stones. The proposed architecture was tested using publicly available CT scan images and demonstrated outstanding accuracy, correctly identifying kidney stones or healthy conditions in 98.6% of cases. It outperformed other advanced techniques, especially in detecting stones of various sizes, including very small ones, proving its effectiveness in medical image analysis.

A hypertree is a special type of connected hypergraph that removes‎ ‎any‎, ‎its hyperedge then results in a disconnected hypergraph‎. ‎Relation between hypertrees (hypergraphs) and trees (graphs) can be helpful to solve real problems in hypernetworks and networks and it is the main tool in this regard‎. ‎The purpose of this paper is to introduce a positive relation (as $\alpha$-relation) on hypertrees that makes a connection between hypertrees and trees‎. ‎This relation is dependent on some parameters such as path‎, ‎length of a path‎, ‎and the intersection of hyperedges‎. ‎For any $q\in \mathbb{N}‎, ‎$ we introduce the concepts of a derivable tree‎, ‎$(\alpha‎, ‎q)$-hypergraph‎, ‎and fundamental $(\alpha‎, ‎q)$-hypertree for the first time in this study and analyze the structures of derivable trees from hypertrees via given positive relation‎. ‎In the final‎, ‎we apply the notions of derivable trees‎, ‎$(\alpha‎, ‎q)$-trees in real optimization problems by modeling hypernetworks and networks based on hypertrees and trees‎, ‎respectively.‎‎‎

Objective: The increase in ecological vulnerability in cities due to rising temperatures has received much attention in recent years. Accurate identification of the urban heat island network is very important for effectively reducing its impact. In many studies, the impact of heat island connectivity on these networks has been largely ignored. This study was conducted to fill this research gap by creating an urban heat island network based on the connectivity perspective to understand the structural characteristics of the effect of this network and evaluate it in order to determine the priority level for implementing temperature reduction measures in the Tehran metropolis.Method: To achieve the above goal, after analyzing the ground surface temperature, areas with high temperatures were identified. Then, morphological spatial pattern analysis, morphological structure evaluation, and recognition of the importance of heat island sources were carried out. In the next stage, the resistance level against thermal diffusion was constructed, and, then, using the minimum cumulative resistance method, heat transfer corridors were identified and analyzed.Results: The findings of this study identified 29 strong heat island cores in Tehran with a relatively scattered distribution, 8 of which showed very high heating power. 31 corridors connected these islands, 10 of which had the potential to increase temperatures significantly. In addition, in terms of spatial distribution, the heat island network fragments in Tehran were more densely located in the western and southern areas. The high density of heat islands in the western part of Tehran made planning to combat them more difficult and increased their influence. Also, the very dense islands located in the southwest of Tehran led to the identification of short heat corridors in this part of the city, which justified the increase in temperature. On the other hand, the results showed that the cores had the largest share in the heat island network in Tehran, which indicated the size of the heat islands and their regional distribution in the study area.Conclusions: In this study, special attention has been paid to the structural characteristics of the urban heat islands of the study area and their degree of importance. This approach is simpler than previous methods of determining the size or density of blue-green spaces to achieve cooling effects. This framework can be used as a strategic measure to prevent the coalescence and expansion of urban heat islands and to avoid the unplanned increase of blue-green spaces aimed at reducing temperatures in urban areas.

In recent years, graph convolutional networks (GCNs) have achieved significant performance in the field of skeleton-based action recognition. Existing GCN-based methods usually apply fixed graph topologies and a temporal convolutional filter to extract the spatial and temporal features of an action. Since an action is coordinated through different parts of the body in the time domain and shows different characteristics in the time domain, this work causes the loss of a lot of information about an action.To address this issue, in this paper, we present an attention-based graph neural (AT-AR) network to discover distinct features from both spatial and temporal aspects. The proposed model uses an SPG Net convolution to learn Spatio-temporal features. In addition, the attention mechanism creates an attention score using temporal features, which can enhance the temporal correlations of an action. After establishing the two-stream structure, the AT-AR network has reached 96% and 97% accuracy under X-View and X-Sub on the NTU RGB+D dataset.