The Uncertainty-aware graph-based deep learning to rank (UGD-LTR) framework tackles two critical gaps in learning to rank. It ignores neither query–document dependencies nor uncertainty quantification. It expands feature space by modeling ranking data as a weighted bipartite graph, from which 22 22 structural features are extracted using centrality, neighborhood, and projection-based metrics. Specialized deep networks estimate these features for unseen query–document pairs. An ensemble of rankers processes this expanded set, while the Choquet fuzzy integral reconciles their outputs by capturing interaction-based confidence. Tests on LETOR4.0 and WCL2R show strong results, including a 5.6

Introduction: Microblogging websites have massed rich data sources for sentiment analysis and opinion mining. In this regard, sentiment classification has frequently proven inefficient because microblog posts typically lack syntactically consistent terms and representatives since users on these social networks do not like to write lengthy statements. Also, there are some limitations to low-resource languages. The Persian language has exceptional characteristics and demands unique annotated data and models for the sentiment analysis task, which are distinctive from text features within the English dialect. Method: This paper first constructs a user opinion dataset called ITRC-Opinion in a collaborative environment and insource way. Our dataset contains 60,000 informal and colloquial Persian texts from social microblogs such as Twitter and Instagram. Second, this study proposes a new architecture based on the convolutional neural network (CNN) model for more effective sentiment analysis of colloquial text in social microblog posts. The constructed datasets are used to evaluate the presented architecture. Furthermore, some models, such as LSTM, CNN-RNN, BiLSTM, and BiGRU with different word embeddings, including Fasttext, Glove, and Word2vec, investigated our dataset and evaluated the results. Results: The results demonstrate the benefit of our dataset and the proposed model (72% accuracy), displaying meaningful improvement in sentiment classification performance.

Social networks play a crucial role in the transmission and dissemination of information within society. Analyzing information dissemination can help identify primary sources of information and understand how it spreads. Recognizing key and influential users in social networks can assist marketing managers in delivering their messages more effectively. This paper uses graph models and machine learning methods to identify, examine, and analyze key points of information dissemination on the Instagram social network. The results indicate that by conducting a thorough analysis of social networks, effective patterns of information dissemination can be identified, which are highly beneficial for optimizing marketing and communication strategies. Additionally, this paper presents various methods for analyzing and predicting information dissemination.

Online social networks (OSNs) such as Facebook, Twitter, Instagram, etc. have attracted many users all around the world. Based on the centrality concept, many methods are proposed in order to find influential users in an online social network. However, the performance of these methods is not always acceptable. In this paper, we proposed a new improvement on centrality measures called P-centrality measure in which the effects of node predecessors are considered. In an extended measure called EP-centrality, the effect of the preceding predecessors of node predecessors are also considered. We also defined a combination of two centrality measures called NodePower (NP) to improve the effectiveness of the proposed metrics. The performance of utilizing our proposed centrality metrics in comparison with the conventional centrality measures is evaluated by Susceptible-Infected-Recovered (SIR) model. The results show that the proposed metrics display better performance finding influential users than normal ones due to Kendall’s τ coefficient metric.

The spread of internet and smartphones in recent years has led to the popularity and easy accessibility of social networks among users. Despite the benefits of these networks, such as ease of interpersonal communication and providing a space for free expression of opinions, they also provide the opportunity for destructive activities such as spreading false information or using fake accounts for fraud intentions. Fake accounts are mainly managed by bots. So, identifying bots and suspending them could very much help to increase the popularity and favorability of social networks. In this paper, we try to identify Persian bots on Twitter. This seems to be a challenging task in view of the problems pertinent to processing colloquial Persian. To this end, a set of features based on user account information and activity of users added to content features of tweets to classify users by several machine learning algorithms like Random Forest, Logistic Regression and SVM. The results of experiments on a dataset of Persian-language users show the proper performance of the proposed methods. It turns out that, achieving a balanced-accuracy of 93.86%, Random Forest is the most accurate classifier among those mentioned above.

Label propagation algorithm is widely used for community detection in a network due to its linear time complexity. It also does not need any predefined information such as the number of communities. However, the results of this algorithm are not stable because of the randomness strategy used in its propagation process. In this paper, a modification on label propagation strategy is proposed in which labels are propagated based on nodes importance defined by their positions and popularity among neighbors. The proposed strategy is an updating process which reduces the instability of the label propagation algorithm. Experiments on real-world and synthetic networks show that the proposed method improves accuracy in terms of modularity, NMI and ARI. Also, the method has an acceptable execution time.

In recent years, the amount of data in the world is growing rapidly. Data growth also occurs in the government sector. All ministries and institutions at every level are data producers. These government-owned data have a high potential if they can be used properly. Open government data can stimulate innovation and economic growth and enhance business models. In order to increase the willingness of citizens to use open government data and enjoy the benefits mentioned, the quality of open government data needs to be improved. The quality of open government data encompasses a variety of dimensions and criteria. Also, the importance of each dimension and criterion in increasing the quality of open government data is different. Therefore, we are faced with a complex system that requires proper decision-making and management. In fact, we are dealing with decision-making in the complex management system. Given the importance of this issue, the purpose of this study is to provide a new and comprehensive method to improve the quality of open government data and increase the willingness of citizens to use the data by considering the complex network of citizens and organizations. For this purpose, library studies have been used to extract comprehensive and effective dimensions and criteria. The statistical population includes all articles related to the criteria of improving the quality of open government data and increasing the willingness of citizens to use the data. The probabilistic sampling method of simple random samples has been used, and 10 articles in this field have been reviewed. After extracting the criteria as well as the data of 112 governmental organizations and institutions related to each criterion from the open data portal, the complex network of citizens and governmental organizations and institutions has been analyzed in order to identify high-degree centrality organizations. Then, the data characteristics of the organizations that were most desired by the citizens were extracted using data mining techniques including the regression model. Also, field method and multicriteria decision-making technique including the DEMATEL technique have been used to express the solutions and identify the cause-and-effect relationships between the solutions. The criteria extracted in improving the quality of open government data and increasing the willingness of citizens to use the data are included: "data originality," "license openness," "up-to-datedness," "data access," "metadata completeness," "number of data sets," "format openness," "nondiscrimination," "understandable," "number of categories of data sets," "free," "lack of missing data," "data request ability," "visualization," "feedback," and "data subject matter." Based on the results obtained from the analysis of the complex network and the regression model, the criterion of "society subject" with a coefficient of 72.564 and a positive sign has the greatest impact on increasing the number of citizens' visits to open government data. After that, the criterion of "format openness" with a coefficient of 52.682 and a positive sign has the second rank in increasing the number of visits. Extracting comprehensive and effective criteria in improving the quality of open government data and increasing citizens' willingness to use data, calculating the weight and importance of each criterion by analyzing the complex network of citizens and organizations, as well as providing solutions, can help managers in decision-making and proper management in the complex system of citizens and government organizations.

Representing a method to identify and contrast with the fraud which is created by robots for developing websites' traffic ranking

Nowadays, social networks have gained a lot of popularity among people. With the growth of these networks and a large number of people using these networks, social network analysis has received special attention, so the need for highly accurate and fast algorithms on various issues is strongly felt. One of the important issues in these networks is community detection problem that many algorithms have been proposed for this purpose. In social networks, communities usually are formed around popular or influential nodes. Most algorithms in this field, that are usually density-based, are unable to detect this structure. In this paper, we propose a new community detection algorithm based on the local popularity structure. In this algorithm, the most popular person in neighborhood of each user is selected as a leader and the user falls into that group. Experimental results on six real networks show that the proposed method not only has comparable results in terms of NMI and ARI, but also has shorter execution time compared to existing algorithms.

Disseminating information through the World Wide Web as the most popular medium has resulted in creating a huge number of web pages and so growing the dimension of the web. In this era of big data, an efficient website ranking to satisfy the web user requirements in different areas such as marketing and E-commerce is a major challenge in the current Internet. In this context, the role of ranking algorithms as a tool to provide services such as measuring the website visibility and comparing the website position to the competitors is crucial. In this paper, we propose an architecture for web domain ranking which includes processing capability required for handling Big Data available on the web. The proposed architecture presents a new method for web domain ranking that is independent of the link structure of the web graph. The proposed method provides web domain ranking based on the number of unique visitors, the number of user sessions, and session duration.

The popularity of social networks has rapidly increased over the past few years. Social networks provide many kinds of services and benefits to their users like helping them to communicate, click, view and share contents that reflect their opinions or interests. Detecting important contents defined as the most visited posts and users whom disseminate them can provide some interesting insights from cyberspace user’s activities. In this paper, a framework for discovering important posts (most popular posts by views count) and influential users is introduced. The proposed framework employed on Telegram instant messaging service in this study but it is also applicable to other social networks such as Instagram and Twitter. This framework continuously works in a real social network analysis system named Zekavat to find daily important posts and influential users. The effectiveness of this framework was shown in experiments. The accuracy achieved in the advertisement detection model is 89%. Text-based clustering part of the framework was tested based on the human factor verification and clustering time is less than linear. Graph creation based on publishing relationships is more effective than mention relationship and in this process influential users can be identified in a precise manner. Keywords-social networks; clustering; LSH; machine learning; important posts;influential users

We consider a group of mobile users, in closed proximity, who are interested in downloading a common content (e.g., a video file). We address a cooperative solution where each mobile device is equipped with both cellular and Wi-Fi interfaces. The users exploit the cellular link to download different shares of the content from the based-station and leverage on Wi-Fi link to exchange the received data. In order to expedite content delivery, the base-station transmits random linear network-coded data to users. This paper presents an analytical study of the average completion time, i.e., the time necessary for all devices to successfully retrieve the data. We propose an analytical model to address the effect of random access MAC as well as the correlation among the received coded packets on the performance of content delivery. In our model, a p-persistent carrier sense multiple access approximation for IEEE 802.11 MAC is considered. We also derive the probability of a newly received packet to be innovative, where the coding coefficients are selected randomly and uniformly from GF(q). Our simulations confirms the theoretical analysis.

Mobile ecosystem contributes to gross domestic product (GDP). GDP contribution to the economy can be evaluated by the generated value added. Since value added and subscribers growth of mobile operators is decreasing in the world year-on-year from 2012, they have sensed that ought to immigrate to a new 5G generation network. 5G technology is appearing in 2020 and owns better properties than the previous generations e.g., LTE-A. Sole properties in 5G make bit rate 1000 fold and connections to 10 fold rather than 2010 and reduce the delay below 1 ms. This technology, in addition, is multiband/multimode/multilayer system, macrocell/microcell/picocell/femtocell hierarchical system and include GSM/UMTS/LTE/WiFi technologies. Besides, 5G uses distributed antenna systems (DAS) and massive MIMO. To access 5G technology, some fundamental and applied projects should be performed.

Manuscript Received 7-July-2012 and Revised 13-April-2013 ISSN: 2322-3936 Accepted on 15May-2013 Abstract— Spectrum auctions have been considered a promising approach to improve the efficiency of spectrum use. Spectrum reusability is also one of the important properties in spectrum auctions. To handle spectrum reusability, a buyer grouping procedure has been applied in many existing spectrum auction schemes. It is important to note that almost none of the proposed buyer grouping algorithms in the existing works has been specifically designed for spectrum allocation problem. However, buyer grouping in a practical spectrum auction mechanism has its own challenges such as heterogeneity and truthfulness. In this paper, first we illustrate the challenges of buyer grouping in a practical spectrum auction mechanism. Then we propose the novel algorithms for spectrum buyer grouping to solve these challenges. By extensive simulations, we show that our proposed algorithms can not only solve the challenges caused by radio spectrum properties but also provide good performance on various auction metrics.

Spectrum auctions are one of the best-known solutions to improve the efficiency of spectrum use. However, there can be many challenges in the design of a practical spectrum auction. Heterogeneity is one of the most major challenges. Unfortunately, most of the existing auction designs either do not take into account the various aspects of heterogeneity or assume only the scenario where each seller supplies one distinct channel and each buyer wishes to buy merely one channel. The authors propose a spectrum auction mechanism which considers the various aspects of heterogeneity as well as multi-channel purchasing. They prove that the auction design preserves three important economic aspects including truthfulness, budget balance and individual-rationality. Moreover, most of the existing works only provide the bidders a simple demand format. Their auction mechanism enables bidders to use diverse demand formats. Furthermore, they propose some novel adaptive grouping algorithms to improve the auction's performance. The simulation results demonstrate good performance of the proposed algorithms on various auction metrics.

The core aim of this work is the maximization of the achievable data rate of the secondary user pairs (SU pairs), while ensuring the QoS of primary users (PUs). All users are assumed to be equipped with multiple antennas. It is assumed that when PUs are present, the direct communications between SU pairs introduces intolerable interference to PUs and thereby SUs transmit signal using the cooperation of other SUs and avoid transmitting in the direct channel. In brief, an adaptive cooperative strategy for multiple-input/multiple-output (MIMO) cognitive radio networks is proposed. At the presence of PUs, the issue of joint relay selection and power allocation in Underlay MIMO Cooperative Cognitive Radio Networks (U-MIMO-CCRN) is addressed. The optimal approach for determining the power allocation and the cooperating SU is proposed. Besides, the outage probability of the proposed communication protocol is further derived. Due to high complexity of the optimal approach, a low-complexity approach is further proposed and its performance is evaluated using simulations. The simulation results reveal that the performance loss due to the low-complexity approach is only about 14%, while the complexity is greatly reduced.

Todays, there is an industry demand in telecommunication networks to provide a full-featured, commercially available, scalable and non-proprietary network management solution, where multi-vendor, multi-technology management systems interoperate in an open architecture environment. This paper presents a novel architecture to enable system interaction in management layer, NML-EML, based on MTNM solution package. In this architecture, communication is based on CORBA and JACORB is used as the CORBA interface. Also the MTNM interface software package has been improved so that it could be integrated as south interface for NML softwares and as north interface for EML softwares. As the implementation of this interface is platform independent it could be used in different platforms. Also by adding the communication medium proposed in this paper, the network management software could manage other standard element management systems. Furthermore, the proposed system can be used as a proof of validity and integrity of other interfaces for their considered functions.

Auctions have been widely studied as an efficient approach of allocating spectrum among secondary users in recent years. On the other side, a wide range of frequency bands could be available in a spectrum auction considering the current trend of deregulating wireless resources, therefore, channels provided by the primary users may reside in widely separated frequency bands, and due to the difference in propagation profile, would show significant heterogeneity in transmission range, channel error rate, path-loss, etc. Also, we can consider the channels with similar propagation and quality characteristics, for example, channels located in the same frequency band, are homogeneous and can be located in one spectrum type. Therefore, in this paper, we propose a novel double auction mechanism for both homogeneous and heterogeneous spectrums, called hybrid spectrums. The hybrid auction design has its own challenges, especially it also inherits the challenges related to heterogeneity. We prove that our auction design can not only solve the challenges caused by hybrid spectrums but also preserve three important economic aspects including truthfulness, budget balance and individual rationality.

Auctions have been widely studied as an efficient approach of allocating spectrum among secondary users in recent years. On the other side, a wide range of frequency bands could be available in a spectrum auction considering the current trend of deregulating wireless resources, therefore, channels provided by the primary users may reside in widely separated frequency bands, and due to the difference in propagation profile, would show significant heterogeneity in transmission range, channel error rate, path-loss, etc. Also, we can consider the channels with similar propagation and quality characteristics, for example, channels located in the same frequency band, are homogeneous and can be located in one spectrum type. Therefore, in this paper, we propose a double auction mechanism for both homogeneous and heterogeneous spectrums, called hybrid spectrums. The hybrid auction design has its own challenges, especially it also inherits the challenges related to heterogeneity. We prove that our auction design can not only solve the challenges caused by hybrid spectrums but also preserve three important economic aspects including truthfulness, budget balance and individual rationality. Also, we show that the proposed scheme increases spectrum utilization through spectrum reuse. Also, we offer a novel comprehensive grouping procedure to increase both the channel utilization and the seller satisfaction. Results from extensive simulation studies demonstrate good performance of the proposed algorithms on various auction metrics.

The topological characteristics of an IEEE 802.16 mesh network including the tree’s depth and degree of its nodes affect the delay and throughput of the network. To reach the desired trade-off between delay and throughput, all potential trees should be explored to obtain a tree with the proper topology. Since the number of extractable tree topologies from a given network graph is enormous, we use a genetic algorithm (GA) to explore the search space and find a good enough trade-off between per-node, as well as network-wide delay and throughput. In the proposed GA approach, we use the Pruefer code tree representation followed by novel genetic operators. First, for each individual tree topology, we obtain expressions analytically for per-node delay and throughput. Based on the required quality of service, the obtained expressions are invoked in the computation of fitness functions for the genetic approach. Using a proper fitness function, the proposed algorithm is able to find the intended trees while different constraints on delay and throughput of each node are imposed. Employing a GA approach leads to the exploration of this extremely wide search space in a reasonably short time, which results in overall scalability and accuracy of the proposed tree exploration algorithm.