Search

Recommendation systems are popular both commercially and in the research community. For example, Online in Social Networks (OSNs) like Twitter, they are gaining an increasing attention since a lot of connection are established between users without any previous knowledge. This highlights one of the key features of a lot of OSNs: the creation of relationships between users. Therefore, it is important to find new ways to provide interesting friendships suggestions. However, mining and analyzing data from large scale Social Networks can become critical in terms of computational resources. This is particularly true in the context of ubiquitous access, where resource-constrained mobile devices are used to access the social network services. To this end, designing architectures/solutions offering the possibility of operating in a Mobile Cloud scenario is of key importance. Accordingly, we present a new recommendation system scheme that tries to find the right trade-offs between the exploitation of the already existing links/relationships and the interest affinities between users. In particular, such scheme is based on an inherently parallel Hubs And Authorities algorithm together with similarity measures that, for scalability purposes, can be easily transposed in a cloud scenario. The first one let us leverage triadic closures while the second one takes into account homophily. The proposal is supported by an extensive performance analysis on publicly available Twitter data. In particular, we proved the effectiveness of the proposed recommendation system by using several performance metrics available in the literature which include precision, recall, F-measure and G-measure. The results show encouraging perspectives in terms of both effectiveness and scalability, that are driving our future research efforts.

World Wide Web is developing in a chaotic and unfocused process, and this process has resulted in production of documents which are linked with each other, and which are not logically organized. Therefore, the aim of recommender systems is guiding users to find their favorite resources and meet their needs, by using the information obtained from the previous users’ interactions. In this paper, to predict the users’ navigation pattern with high precision, a hybrid algorithm of FCM fuzzy clustering techniques, weighted association rules, and fuzzy systems are presented. This algorithm is implemented in two phases, namely offline and online phases. In offline phase, using the recorded data in log file of the web server, the users’ navigation patterns are extracted. In online phase, the recommender system suggests, as the initial proposed set, a list of the current user’s favorite webpages which he/she has not visited yet. Then it expands this set using HITS algorithm so that the new webpages which have recently been added to the website have the chance to be present in the list of the proposed webpages. The results of the simulation in real-world data indicate the higher efficiency of the proposed algorithm in terms of precision and coverage comparing to other algorithms.

the explosive growth of information in the Web, users face difficulties while finding their desired information. Search engine helps the user by retrieving useful information from this huge collection based on his/her search query and presents a list of relevant web pages as a search result. However, without proper ranking of pages in the result through the relevancy of pages to the search query, the user may need to explore the whole list for discovering the appropriate page(s), thereby involving huge search time. Although a number of ranking algorithms such as HITS, PageRank, Weighted PageRank and etc., are developed to assist the search engine, but none of them provides page ranking with high accuracy. In this paper, we propose a score- based web page ranking algorithm involving web content mining and usage information of the pages. Our algorithm considers both syntactical and semantic matches of the search query to the pages. For a web page, syntactical score is calculated based on the total number of exact matches of the search words in the page. Besides, semantic score is measured using synonym matches of the search words. Moreover, we incorporate the usage information of the pages as page popularity in order to comprise the user interest in the ranking order. The total relevant score of each page is calculated using the summation of the syntactical and semantic scores of the page and its page popularity. Finally, the pages are ranked according to their total relevant score. Based on several performance evaluation measures, experimental results show considerable improvement in the page ranking using our algorithm as compared to other known approaches.

A naive approach has been made by applying the concept of reachability for web page ranking and implementing multi resolution analysis using Haar wavelet to order the web pages. In this article, page ranking has been done by developing a structured signal using in links, out links and reachability values of the web pages of network graphs. Using Haar wavelet, the page ranking is proposed and developed. The average and detailed coefficients of the input signal and the down sampling process provides the necessary page ranking of web pages. This approach does not involve any iterative technique, damping factor or initialization of the page ranks. In this paper, comparison between the original page rank, category-based page rank and the proposed approach have been made. The result reflects the role of paths between the pages in page rankings.

With a large amount of geotagged resources from smart devices, it is important to provide users with intelligent location-based services. Particularly, in this work, we focus on spatial ranking service, which can retrieve a set of relevant resources with a certain tag. This paper designs ranking algorithm in order to find out a list of locations which are collected from geotagged resources on SNSs. As extending HITS algorithm [13], we propose a novel method (called GeoHITS) that can analyze an undirected 2-mode graph composed with a set of tags and a set of locations. Thereby, meaningful relationships between the locations and a set of tags are discovered by integrating several weighting schemes and HITS algorithm. To evaluate the proposed spatial ranking approach, we have shows the experimental results from the recommendation applications. This article has been corrected. Link to the correction 10.2298/CSIS151203064E

Multi-document summarization techniques aim to reduce documents into a small set of words or paragraphs that convey the main meaning of the original document. Many approaches to multi-document summarization have used probability-based methods and machine learning techniques to simultaneously summarize multiple documents sharing a common topic. However, these techniques fail to semantically analyze proper nouns and newly-coined words because most depend on an out-of-date dictionary or thesaurus. To overcome these drawbacks, we propose a novel multi-document summarization system called FoDoSu, or Folksonomy-based Multi-Document Summarization, that employs the tag clusters used by Flickr, a Folksonomy system, for detecting key sentences from multiple documents. We first create a word frequency table for analyzing the semantics and contributions of words using the HITS algorithm. Then, by exploiting tag clusters, we analyze the semantic relationships between words in the word frequency table. Finally, we create a summary of multiple documents by analyzing the importance of each word and its semantic relatedness to others. Experimental results from the TAC 2008 and 2009 data sets demonstrate the improvement of our proposed framework over existing summarization systems.

Herein, we propose an algorithm to approximate web communities from the topic related web pages. The approximation is achieved by subspace factorization of the topic related web pages. The factorization process reveals existing association between web pages such that the closely related web pages are extracted. We vary the approximation values to identify varied degrees of relationship between web pages. Experiments on real data sets show that the proposed algorithm reduces the impact of unrelated links and therefore can be used to control spam links in web pages.

Stackoverflow is one of the most popular question and answer (Q&A) websites. One of the reasons for the popularity is the use of badges and reputation score. The reputation score is cumulative of answers accepted by the community users on their posted questions on various topics. This paper tries to find out the most influential top users topic (tag) specific over a group of topics by forming social network based on the question owner and accepted user owner for a particular topic (tag) using SNAP (Stanford Network Analytics Platform). After formation of topic (tag)-specific social network graph, top 10 experts and learners are obtained using HITS algorithm. A simple scoring methodology is used to find top-n most influential experts and learners in the group of topics (tags). Three case studies are undertaken for three groups of topics (tags). We find that there are some common users appearing in the top 10 experts list in a group of topics (tags). It is also observed that there are no common users in the top 10 learners list for first two case studies. However, some users appear common in top 10 learners for more than 1 topic (tag) in the third case study which is more domain-specific. The latest Stackoverflow dataset has been used.

Link analysis method in general and Hyperlink-Induced Topic Search (HITS) algorithm in particular is used for analyzing graph in many important contexts, including ranking search results based on the hyperlink structure of the World Wide Web and privacy analysis of social networks with the purpose of identifying the role of each object (node) of graph by computing node weight. Similar to almost link analysis algorithms, HITS method consider all edges identically in each computation. However, we find that value of edges in real cases are not identical in analyzing and depend on the relationships/associations in content of two endings. Therefore, in HITS, edge weight should be also computed so that be suitable with value of corresponding relationship/association. Benefiting from the edge weighting, we can treat the problems of "invaluable edges" and their relevant nodes (normally as "rubbishes") in analysis. With this motivation, an improvement HITS (E-HITS) is proposed in this paper. In proposed method, we analyze the meaning of edge, corresponding relationship/association and its characteristics (especially as "group-based") as basis for edge-weight calculation, in which, the calculation is only based on the available hub and authority parameters on graph. Experiments on simulation graphs as well as on real web-graph data show evidently that along with edge weighting, our proposed method can remove the impacts of invaluable edges as well as rubbish nodes in analysis, give the result more positively and objectively than the original HITS method.

Ubiquitous use of social media such as microblogging platforms brings about ample opportunities for the false information to diffuse online. It is very important not just to determine the veracity of information but also the authenticity of the users who spread the information, especially in time-critical situations like real-world emergencies, where urgent measures have to be taken for stopping the spread of fake information. In this work, we propose a novel machine learning based approach for automatic identification of the users spreading rumorous information by leveraging the concept of believability, i.e., the extent to which the propagated information is likely to be perceived as truthful, based on the trust measures of users in Twitter's retweet network. We hypothesize that the believability between two users is proportional to the trustingness of the retweeter and the trustworthiness of the tweeter, which are two complementary measures of user trust and can be inferred from retweeting behaviors using a variant of HITS algorithm. With the retweet network edge-weighted by believability scores, we use network representation learning to generate user embeddings, which are then leveraged to classify users into as rumor spreaders or not. Based on experiments on a very large real-world rumor dataset collected from Twitter, we demonstrate that our method can effectively identify rumor spreaders and outperform four strong baselines with large margin.

This paper presents techniques of retrieving useful information from a mixture of Web pages collected from either question-answer sites (Q&A sites) or Web search engines. The proposed techniques are designed to discover the maximum possible amount of know-how knowledge from such collections of Web pages, where know-how knowledge is defined as text contents qualified as information source regarding specific domain of questions. The major intent is to build a framework that selects helpful information to provide answers to various problems of interest, such as useful tips to a question. Techniques in this paper primarily attempt to complement knowledge available on Q&A sites with pages collected from search engines via topic models. In order to argue that pages collected from search engine are truly supplements to know- how knowledge on Q&A sites we verify how much extra useful information the Web search engine is able to provide by manually inspecting Web pages aggregated by the topic model.

This paper addresses the problem of identifying irrelevant items from a small set of similar documents using Web search engine suggests. Specifically, we collected volumes of Web pages through Web search engines and inspected the page contents using topic models. Among each cluster of pages sharing the same topic indicated by the topic model, our technique discovers potential content organization in the current page cluster and identifies pages that are out of focus from that topic. The metrics in our approach mainly consist of search engine suggest frequency and inter-document similarity measures. The intuition is that Web pages collected via the same search queries are more likely to share similar contents. We verify this intuition by implementing a subtopic based document selection framework and making quantitative evaluation against human made labeled data sets. Our evaluation result reveals that suggest frequency analysis along with inter-document similarity measure is effective at filtering off-topic documents in small data sets with satisfactory performance.

Construction of Ontology is indispensable with rapid increase in textual information. Much research in learning Ontology are supervised and require manually annotated resources. Also, quality of Ontology is dependent on quality of corpus which may not be readily available. To tackle these problems, we present an iterative focused web crawler for building corpus and an unsupervised framework for construction of Domain Ontology. The proposed framework consists of five phases, Corpus Collection using Iterative Focused crawling with novel weighting measure, Term Extraction using HITS algorithm, Taxonomic Relation Extraction using Hearst and Morpho-Syntactic Patterns, Non Taxonomic relation extraction using association rule mining and Domain Ontology Building. Evaluation results show that proposed crawler outweighs traditional crawling techniques, domain terms showed higher precision when compared to statistical techniques and learnt ontology has rich knowledge representation.

The use of network analysis in legal studies has increased rapidly in recent years. Many legal scholars have been attracted to network analysis, among other things by the possibility of conducting large-scale studies of case law. However, when in time we study a citation network impacts the results, sometimes significantly. To reduce this problem, this paper proposes that scholars use PageRank when studying inward citations in legal citation network and a similar measurement, HubRank, when studying outward citations, rather than the HITS algorithm. However, the paper also advocates further methodological discussion and development.

Understanding the criteria (or “aspects”) that reviewers use to evaluate games is important to game developers and publishers, since this will give them important input on how to improve their products. Techniques for the extraction of such aspects have been studied by others, albeit not specific to the gaming industry. In this paper we demonstrate an aspect extraction and analysis system specific to computer games. The system extracts game review texts from a list of known websites and automatically extracts candidate aspects from the review text using techniques from natural language processing and sentiment analysis. It then ranks the candidate aspects using the HITS algorithm. To evaluate the correctness of the extracted aspects, we used the system to calculate an overall score for each game by aggregating its highly-rated aspects, weighted by the importance of the respective aspects. The aggregated scores resulted in a ranking of games, which we compared to a known ranking from a popular website – th...

The immense number of documents published on the web requires the utilization of automatic classifiers that allow organizing and obtaining information from these large resources. Typically, automatic web pages classifiers handle millions of web pages, tens of thousands of features, and hundreds of categories. Most of the classifiers use the vector space model to represent the dataset of web pages. The components of each vector are computed using the term frequency inversed document frequency (TFIDF) scheme. Unfortunately, TFIDF-based classifiers face the problem of the large-scale size of input data that leads to a long processing time and an increase in resource requests. Therefore, there is an increasing demand to alleviate these problems by reducing the size of the input data without influencing the classification results. In this paper, we propose a novel approach that improves web page classifiers by reducing the size of the input data (i.e. web pages and feature reduction) by using the hypertext induced topic search (HITS) algorithm. We employ HITS results for weighting remaining features. We evaluate the performance of the proposed approach by comparing it with the TFIDF-based classifier. We demonstrate that our approach significantly reduces the time needed for classification.

We analyze the performance of international ODI cricket teams through a hierarchical ranking scheme. Players are represented as the nodes of a graph with batsmen as authorities, and bowlers as hubs. Low level player ratings determined using weighted HITS algorithm, are fed into a higher level Elo rating system to rank teams. Strike rate, economy rate, number of boundaries, etc. determine the edge weights for player graph. Match characteristics like margin of victory, winning style and player rankings determine the K-factor in Elo ratings. We show that player composition along with match characteristics is also an important aspect of team ranking, which has not explored previously. We report significant improvements in predicting match outcomes against other ranking schemes.

Numerous centrality measures have been developed to quantify the importances of nodes in time-independent networks, and many of them can be expressed as the leading eigenvector of some matrix. With the increasing availability of network data that changes in time, it is important to extend such eigenvector-based centrality measures to time-dependent networks. In this paper, we introduce a principled generalization of network centrality measures that is valid for any eigenvector-based centrality. We consider a temporal network with N nodes as a sequence of T layers that describe the network during different time windows, and we couple centrality matrices for the layers into a supra-centrality matrix of size NTxNT whose dominant eigenvector gives the centrality of each node i at each time t. We refer to this eigenvector and its components as a joint centrality, as it reflects the importances of both the node i and the time layer t. We also introduce the concepts of marginal and conditional centralities, which facilitate the study of centrality trajectories over time. We find that the strength of coupling between layers is important for determining multiscale properties of centrality, such as localization phenomena and the time scale of centrality changes. In the strong-coupling regime, we derive expressions for time-averaged centralities, which are given by the zeroth-order terms of a singular perturbation expansion. We also study first-order terms to obtain first-order-mover scores, which concisely describe the magnitude of nodes' centrality changes over time. As examples, we apply our method to three empirical temporal networks: the United States Ph.D. exchange in mathematics, costarring relationships among top-billed actors during the Golden Age of Hollywood, and citations of decisions from the United States Supreme Court., 38 pages, 7 figures, and 5 tables

As a convenient and popular social network platform, microblogging has become an important medium for people’s information exchange. The development of microblogging network can be regarded as a growth of complex network, and dissemination of information in microblogging network shows strong regularity. In the face of big scale of users, the appraisal of user influence become an important index to measures the value of a user. How to recognize user influence becomes a hot topic in microblogging research. In this paper, we proposed a method named InformationRank to evaluate user influence, which based on measure the information of node betweenness centrality in the network. Then, we analyze and compare the efficiency of this method with PageRank algorithm and Hits algorithm. The experimental results indicate that InformationRank algorithm is effective in a large scale network.

and is a massive, explosive, data resource in the world. In information retrieval approach web Search engines are predominant tools for finding and getting access to the contents of web. The primary goal of Search engine is to provide relevant information to the users according to their needs, Usually Search engines gives large result set for a user’s query. To limit the result list, it is necessary to assign ranking the web pages in an efficient and effective manner. Artificial Bee Colony (ABC) is one of the new approaches used to solve optimization problems. This paper proposes Artificial Bee Colony (ABC) approach as a new method for web mining particularly in ranking web pages. It considers users interest, total web site linkage and growth analysis rate are used to assign rank the web pages. Proposed ABC approach for ranking web pages is implemented and tested on real datasets. The experimental results shows efficiency of the proposed method compared with tradition page Rank Algorithm.

PageRank algorithms only consider hyperlink information, without other page information such as page hits frequency, page update time and web page category. Therefore, the algorithms rank a lot of advertising pages and old pages pretty high and can’t meet the users' needs. This paper further studies the page meta-information such as category, page hits frequency and page update time. The Web page with high hits frequency and with smaller age should get a high rank, while the above two factors are more or less dependent on page category. Experimental results show that the algorithm has good results.

Determining the titles of Web pages is an important element in characterizing and categorizing the vast number of Web pages. There are a few approaches to automatically determining the titles of Web pages. As an R&D project for Naver, the operator of Naver (Korea's largest portal site), we developed a new method that makes use of anchor texts and analysis of links among Web pages. In this paper, we describe our method and show experiment results of its performance.

In this paper, a conceptual graph-based framework for summarizing short texts is proposed. A semantic representation is implemented through conceptual graph structures that consist of concepts and conceptual relations that stand for texts. To summarize conceptual graphs, the most important nodes are selected using a set of operations: generalization, association, ranking, and pruning, which are described. The importance of nodes on weighted conceptual graphs is measured using a modified version of HITS algorithm. In addition, some heuristic rules are used to keep coherent structures based on information from WordNet (hierarchy of concepts) and VerbNet (semantic patterns of verbs). The experimental results show that this approach is effective in summarizing short texts.

HITS algorithm developed by Jon Kleinberg made use of the link structure of the web network in order to discover and rank pages relevant to a particular topic. But it only took account of the hyperlink structure and completely excluded contents of web pages. Moreover, it ignored the fact that degrees of the importance of many links may be different. Therefore, this algorithm will lead to topic drifts. In this paper, we propose an improved HITS algorithm based on the theory of triadic closure and VSM. This method firstly computes the relevance between arbitrary two pages based on page topic similarity and common reference degree. Then, by using the relevance, a new adjacency matrix is constructed to iteratively calculate authorities and hubs. Preliminary experiments show the new algorithm improves the efficiency and quality of query, reduce the theme drifts.

Scientific collaborations commonly take place in a global and competitive environment. Coalitions and project consortia are formed among universities, companies and research institutes to apply for research grants and to perform jointly collaborative projects. In such a competitive environment, individual institutes may be strategic partners or competitors. Measures to determine partner importance have practical applications such as comparison and rating of competitors, reputation evaluation or performance evaluation of companies and institutes. Many network-centric metrics exist to measure the importance of individuals or companies in social and collaborative networks. Here we present a novel approach for measuring and combing various criteria for partner importance evaluation. The presented approach is cost sensitive, aware of temporal and context-based partner authority, and takes structural information with regard to structural holes into account. Well-established graph models such as the notion of hubs and authorities provide the basis for the presented authority ranking approach and are systematically extended towards a novel unified HITS/PageRank model. The applicability of the proposed approach and the effects of parameter selection are extensively studied using real data from the European Union's research program.

The web consists of a huge number of documents that have been published without any quality control. To retrieve necessary information from World Wide Web, search engines carry out number of tasks based on their respective structural design. Various Search engine follow, different algorithm for ranking pages & produce different result. Search engine generally returns a large number of web pages in response to user queries using algorithms. This paper discusses the idea of page rank algorithm based on the number of visits by the client or user. In this paper, a detailed study is being performed on the Page Rank, the Google's algorithm and check the damping factor value 0.85

As one of the new service model of Web 2.0, the emergence of Community Question-Answering system brings a new way for users to obtain information. However, the explosive growth of users and information, it will be hard for users to obtain the information quickly and accurately. Therefore, it is important to find experts in Community Question-Answering system to improve the accuracy and efficiency of information obtaining. This paper firstly analyzed the relationship among users, questions, and answers in Community Question-Answering system, and built the user diagram, and then by means of the Web mining technology, that is the link analysis weighted HITS algorithm, to find experts out. Finally, three evaluation indices were used to measure the validity of the experts finding algorithm. Experimental results show the effectiveness of the weighted HITS algorithm.

Inaccessible Web pages and 404 "Page Not Found" responses are a common Web phenomenon and a detriment to the user's browsing experience. The rediscovery of missing Web pages is, therefore, a relevant research topic in the digital preservation as well as in the Information Retrieval realm. In this article, we bring these two areas together by analyzing four content- and link-based methods to rediscover missing Web pages. We investigate the retrieval performance of the methods individually as well as their combinations and give an insight into how effective these methods are over time. As the main result of this work, we are able to recommend not only the best performing methods but also the sequence in which they should be applied, based on their performance, complexity required to generate them, and evolution over time. Our least complex single method results in a rediscovery rate of almost $$70\,\%$$ 70 % of Web pages of our sample dataset based on URIs sampled from the Open Directory Project (DMOZ). By increasing the complexity level and combining three different methods, our results show an increase of the success rate of up to $$77\,\%$$ 77 % . The results, based on our sample dataset, indicate that Web pages are often not completely lost but have moved to a different location and "just" need to be rediscovered.

the vast growth of the Internet, many web pages are available online. Search engines use a component called as web crawlers for collecting these web pages from the web for storage and indexing. Many web pages are autonomous and are updated independent of the users. .As the web pages are updated autonomously; users do not come to know of how often the sources change. An incremental crawler visits the web repeatedly after a specific interval of time for the updation of its collection. Users are benefited by knowing the page importance based upon the page accessing frequency. This paper finds out the page importance based on page accessing frequency and also architecture for the same is also proposed.

PageRank is an algorithm introduced in 1998 and used by the Google Internet search engine. It assigns a numerical value to each element of a set of hyperlinked documents (that is, web pages) within the World Wide Web with the purpose of measuring the relative importance of the page. The key idea in the algorithm is to give a higher PageRank value to web pages which are visited often by web surfers. On its website, Google describes PageRank as follows: ``PageRank reflects our view of the importance of web pages by considering more than 500 million variables and 2 billion terms. Pages that are considered important receive a higher PageRank and are more likely to appear at the top of the search results." Today PageRank is a paradigmatic problem of great interest in various areas, such as information technology, bibliometrics, biology, and e-commerce, where objects are often ranked in order of importance. This article considers a distributed randomized approach based on techniques from the area of Markov chains using a graph representation consisting of nodes and links. We also outline connections with other problems of current interest to the systems and control community, which include ranking of control journals, consensus of multi-agent systems, and aggregation-based techniques.

The World Wide Web has emerged to become the biggest and most popular way of communication and information dissemination. Every day, the Web is expending and people generally rely on search engine to explore the web. Because of its rapid and chaotic growth, the resulting network of information lacks of organization and structure. It is a challenge for service provider to provide proper, relevant and quality information to the internet users by using the web page contents and hyperlinks between web pages. This paper deals with analysis and comparison of web pages ranking algorithms based on various parameters to find out their advantages and limitations for ranking web pages and to give the further scope of research in web pages ranking algorithms. Six important algorithms: the Page Rank, Query Dependent-PageRank, HITS, SALSA, Simultaneous Terms Query Dependent-PageRank (SQD-PageRank) and Onto-SQD-PageRank are presented and their performances are discussed.

Background/Objectives: PageRank given by Brin and Larry in 1998, emerged as a dominant link analysis method used by web search engines for ranking of its search results. Efficient and fast computation of PageRank values for prodigious web graphs is indeed an important issue for web search engines today. Recognizing and fighting with spam web pages is also considered to be another necessary issue in web searching. Methods/Statistical Analysis: In this paper, we have proposed an efficient and accelerated parallel computation of PageRank scores on Graphics Processing Units (GPUs) which uses non even distribution of PageRank values. This work is experimented on datasets taken from Stanford Large Network Dataset Collection, on a system equipped with NVIDIA Quadro 2000 Graphics card using CUDA programming language. Findings: The proposed work has a speed up of 3.22 to 7.5 and is also capable of dealing with spam web pages. Application: The proposed algorithm helps in detecting spam web pages.

The extant work on network analyses has thus far paid little attention to the heterogeneity in time lags and speed of information propagation along edges. In this paper, we study this novel problem, modeling the time dimension and lags on network edges, in the context of paper and patent citation networks where the variation in the speed of knowledge flows between connected nodes is apparent. We propose to model time lags in knowledge diffusions in citation networks in one of the two ways: deterministic lags and probabilistic lags. Then, we discuss two approaches of computationally working with time lags in edges of citation networks. Experimentally, we study two different applications to demonstrate the importance of the time dimension and lags in citations: (1) HITS algorithm and (2) patent citation recommendation. We conduct experiments on millions of U. S. patent data and Web of Science (WOS) paper data. Our experiments show that incorporating time dimension and lags in edges significantly improve network modeling and analyses.

In recent years, with the popularity of smart mobile devices, mobile Internet has rapidly developed. When the social network meets the localization technology, it gives birth to a Location-Based Social Network (LBSN). The situational awareness based on the location has become more research significance. However, how to combine the context awareness, mobile sensors and abundant users' historical location data to make the platform more efficient, how to ensure that the proceeds can improve the accuracy of recommendation and perceived task assignment, are still challenges in the location-based social network. In this paper, we demonstrate a model to use historical location data of the participants and analyze the point-of-interest (POI). Then we propose the user location and the empirical value algorithm PTHS based on the HITS algorithm. Through analyzing the interest points of the selected scene perception task, we find that those users have similar point-of-interest, and rank them by the location and experience PTHS algorithm. Finally, much more appropriate users are assigned to the tasks. Through theoretical analysis and extensive simulations, we validate that proposed method is effective and efficient.

In recent years, fraud is becoming more rampant internationally with the development of modern technology and global communication. Due to the rapid growth in the volume of call logs, the task of fraudulent phone call detection is confronted with Big Data issues in real-world implementations. While our previous work, FrauDetector, has addressed this problem and achieved some promising results, it can be further enhanced as it focuses on the fraud detection accuracy while the efficiency and scalability are not on the top priority. Meanwhile, other known approaches suffer from long training time and/or cannot accurately detect fraudulent phone calls in real time. In this paper, we propose a highly-efficient parallelized graph-mining-based fraudulent phone call detection framework, namely PFrauDetector, which is able to automatically label fraudulent phone numbers with a "fraud" tag, a crucial prerequisite for distinguishing fraudulent phone call numbers from the normal ones. PFrauDetector generates smaller, more manageable sub-networks from the original graph and performs a parallelized weighted HITS algorithm for significant speed acceleration in the graph learning module. It adopts a novel aggregation approach to generate the trust (or experience) value for each phone number (or user) based on their respective local values. We conduct a comprehensive experimental study based on a real dataset collected through an anti-fraud mobile application, Whoscall. The results demonstrate a significantly improved efficiency of our approach compared to FrauDetector and superior performance against other major classifier-based methods.

Ranking web pages are generally based on in links and out links of a particular web page which may be exploited just to gain attention. Once weightage is calculated based on structural mining, the same may be excluded from the web page. Thus this approach is not trust worthy while calculating page importance for ranking. Research has been deviated from this traditional approach and started considering number of hits a particular web page got and the dwelling time. In this paper, we propose an algorithm for ranking scientific journals primarily based on considering the above metrics and the Research Author's profile.

The current information age has witnessed massive increase in the number of web based documents and services. Due to this explosive growth of online content, automated search engine programs are used to search and categorize billions of webpages and show only the most relevant pages for the search query submitted by the user. Search engines employ combination of automated algorithms, manually edited directories and advertisements to generate results for users' queries. In this paper, we propose and implement an augmented version of the standard PageRank algorithm [1] by using ‘weight’ of in-linked web pages. Rather than evenly dividing the weight of an in-linked webpage, our technique distributes it to all the out linked pages on the basis of their popularity. We name this enhanced page ranking algorithm as WIL (Weightage In-Link) PageRank algorithm. WIL uses the weights of in-linked webpages to calculate a new score of every individual webpage called WIL-score. Later the webpages can be ranked according to this WIL-score.

With the growing scale of networks, network attack strategies with high attack efficiency and computational efficiency are becoming more and more important. Various attack strategies with various sorting methods have been proposed. However most of them neglected the computational efficiency. Inspired by the high computational efficiency of PageRank and HITS algorithms that used in web pages network, this paper introduces those two algorithms to the network attack separately and explores the feasibility of those two new strategies. The initial experiments choose degree and BC attack as the contrast group, and compare the attack efficiency in six virtual networks. The results indicate that considering the computational efficiency and attack efficiency simultaneously, PageRank strategy has a better attack performance than other compared strategies. Initial discussions about the results are also given.

In this paper, the authors propose a novel bias detection method based on social preference learning for targets on competing topics such as “GalaxyTab vs. iPad” in Twitter. People tend to evaluate a topic by expressing their opinions towards the associated targets such as price and quality. To exploit characteristics of social data, targets are extracted by a modified HITS algorithm on a tripartite graph. The main contribution is that social preferences are learned with explicit sentiment, latent sentiment as social semantics, and lexical sentiment as contextual semantics on targets of the topic, and that the individual preference is considered together with social preferences for the bias detection of a tweet. Experimental results on Twitter collection show significant improvements over all baseline methods. The results indicate that the method deals with not only the lack of a sentiment lexicon but also social and contextual semantics on targets of social users.

Scientific collaboration commonly takes place in a global and competitive environment. Coalitions and consortia are formed among universities, companies and research institutes to apply for research grants and to perform joint projects. In such a competitive environment, individual institutes may be strategic partners or competitors. Measures to determine partner importance have practical applications such as comparison and rating of competitors, reputation evaluation or performance evaluation of companies and institutes. Many network-centric metrics exist to measure the important of individuals or companies in social and collaborative networks. Here we present a novel context-based metric to measure the importance of partners in scientific collaboration networks. Well-established graph models such as the notion of hubs and authorities provide the basis for this work and are systematically extended to a flexible, context-aware network importance measure.

SEO stands for Search Engine Optimization. It is a technique that searches various web pages for specified keywords and ranks these Web pages according to some parameters. They are used to feed pages to search engines.  The main importance of SEO is that it helps to find the relevant data and increase the rank of a webpage in search engines’ results. In our paper, we develop a new algorithm M-HITS (Modified HITS) to provide the page rank. M-HITS Algorithm is a new version of HITS algorithm. It is developed by extending the properties of HITS algorithm.

Search engines save copies of crawled web pages to provide instant search results, saved pages may become old and un-updated as original pages change providing new information and new links, and most of websites don’t submit these new changes to search engines so search engines don’t depend mainly on websites techniques of submitting changes. Keeping pages fresh and updated in search engine is important for giving real page ranks and for providing real time information. Techniques were invented to improve the page update process by search engine. In this paper the author combines two of good known techniques and implements the new one via experiments that improve better results in different experiment cases.

The notions of subgraph centrality and communicability, based on the exponential of the adjacency matrix of the underlying graph, have been effectively used in the analysis of undirected networks. In this paper we propose an extension of these measures to directed networks, and we apply them to the problem of ranking hubs and authorities. The extension is achieved by bipartization, i.e., the directed network is mapped onto a bipartite undirected network with twice as many nodes in order to obtain a network with a symmetric adjacency matrix. We explicitly determine the exponential of this adjacency matrix in terms of the adjacency matrix of the original, directed network, and we give an interpretation of centrality and communicability in this new context, leading to a technique for ranking hubs and authorities. The matrix exponential method for computing hubs and authorities is compared to the well known HITS algorithm, both on small artificial examples and on more realistic real-world networks. A few other ranking algorithms are also discussed and compared with our technique. The use of Gaussian quadrature rules for calculating hub and authority scores is discussed., Comment: 28 pages, 6 figures

The web is a huge and highly dynamic environment which is growing exponentially in content and developing fast in structure. No search engine can cover the whole web, but it has to focus on the most valuable pages for crawling. Many methods have been developed based on link and text analysis for retrieving the pages. In this paper, an algorithm based on link, text, logarithmic distance and probabilistic measure is presented to find the relevancy of the web pages. Here, the most relevant pages are retrieved. It has been proved experimentally that this method provides more number of relevant pages.

is a system of interlinked hypertext documents accessed via Internet. Internet is a global system of interconnected computer networks that serve billions of users worldwide. The huge amount of documents on the web is challenging for web search engines. Web contains multiple copies of the same content or same web page. Many of these pages on the Web are duplicates and near duplicates of other pages. Web search engines face substantial problems due to such duplicate and near duplicate web pages. These pages enlarge the space required to store the index, increase the cost of serving results and so frustrates the users. To assist search engines to provide search results free of redundancy to users and to provide distinct useful results on the first page, duplicate and near duplicate detection is required. The proposed approach will detect near duplicate web pages to increase search effectiveness and storage efficiency of search engine.

There are billions of web pages available on the World Wide Web (WWW). So there are lots of search results corresponding to a user's query out of which only some are relevant. The relevancy of a web page is calculated by search engines using page ranking algorithms. Most of the page ranking algorithm use web structure mining and web content mining to calculate the relevancy of a web page. In this paper, the standard Weighted PageRank algorithm is being modified by incorporating Visits of Links(VOL).The proposed method takes into account the importance of both the number of visits of inlinks and outlinks of the pages and distributes rank scores based on the popularity of the pages. So, the resultant pages are displayed on the basis of user browsing behavior.

According to the sensitive information exposure problem without authority in the classified network caused by negligence or even malice during the web construction in secret units, this paper proposes a discovery method to extract the sensitive information in the web pages based on Web information. Through hypertext transfer protocol connection module this method firstly realizes the function of acquiring all content in the web pages, and then uses regular expression technology to mark all sensitive information in every single web page. With the marked pages the method then uses the analysis function to judge whether the page belongs to the illegal pages and via the database management module stores the results of the analysis. Finally the method through visual display technology reveals the analysis results. Experiments show that this method based on information extraction technology can quickly and effectively find the sensitive information network has been released to help managers promote the efficiency of website maintenance.

A follower can be divided into supporter, non-supporter, or neutral according to a follower’s intention to a target user. Even though a follower is identified as a supporter, an opinion may not be positive to the target user. In this paper, we propose a method to classify a follower as supporter, non-supporter or neutral. To expand information of a follower, influential transmitters who support a target user are detected by using a modified HITS algorithm. To detect a follower’s specific opinion, social issues are extracted based on tweets of influential transmitters. The thread tweets are clustered based on Latent Dirichlet Allocation for social issues. Then, sentiment analysis is conducted for the clusters of a follower. To see the effectiveness of our method, a Korean tweet collection is constructed. As a result, we found that lots of supporting followers show opposite opinions depending on particular issues.

Key Opinion Leaders commonly referred to as thought leaders which play very important role in life science industry. We introduce the concept of identifying key opinion leaders using weighted Social Network Analysis (SNA). We use European PubMed Central dataset for creating weighted social network of healthcare related people and apply different centrality measures on it. We compare the centrality results of weighted and un-weighted social networks as well as weighted social network with PageRank and HITS algorithm. Initial experiments indicate that centrality results of weighted social network works 11 out 12 times better than un-weighted social networks and covers all the important results out of HITS and PageRank algorithm.