Hyperlink-Induced Topic Search Algorithm Research Articles

HITS-based Propagation Paradigm for Graph Neural Networks

In this article, we present a new propagation paradigm based on the principle of Hyperlink-Induced Topic Search (HITS) algorithm. The HITS algorithm utilizes the concept of a “self-reinforcing” relationship of authority-hub. Using HITS, the centrality of nodes is determined via repeated updates of authority-hub scores that converge to a stationary distribution. Unlike PageRank-based propagation methods, which rely solely on the idea of authorities (in-links), HITS considers the relevance of both authorities (in-links) and hubs (out-links), thereby allowing for a more informative graph learning process. To segregate node prediction and propagation, we use a Multilayer Perceptron in combination with a HITS-based propagation approach and propose two models: HITS-GNN and HITS-GNN+. We provided additional validation of our models’ efficacy by performing an ablation study to assess the performance of authority-hub in independent models. Moreover, the effect of the main hyper-parameters and normalization is also analyzed to uncover how these techniques influence the performance of our models. Extensive experimental results indicate that the proposed approach significantly improves baseline methods on the graph (citation network) benchmark datasets by a decent margin for semi-supervised node classification, which can aid in predicting the categories (labels) of scientific articles not exclusively based on their content but also based on the type of articles they cite.

Distributed Graph Algorithms for Multiple Vector Engines of NEC SX-Aurora TSUBASA Systems

This paper describes the world-first attempt to develop distributed graph algorithm implementations, aimed for modern NEC SX-Aurora TSUBASA vector systems. Such systems are equipped with up to eight powerful vector engines, which are capable to significantly accelerate graph processsing and simultaneously increase the scale of processed input graphs. This paper describes distributed implementations of three widely-used graph algorithms: Page Rank (PR), Bellman-Ford Single Source Shortest Paths (further referred as SSSP) and Hyperlink-Induced Topic Search (HITS), evaluating their performance and scalability on Aurora 8 system. In this paper we describe graph partitioning strategies, communication strategies, programming models and single-VE optimizations used in these implementations. The developed implementations achieve 40, 6.6 and 1.3 GTEPS performance on PR, SSSP and HITS algorithm on 8 vector engines, at the same time achieving up to 1.5x, 2x and 2.5x acceleration on 2, 4 and 8 vector engines of Aurora 8 systems. Finally, this paper describes an approach to incorporate distributed graph processing support into our previously developed Vector Graph Library (VGL) framework – a novel framework for graph analytics on NEC SX-Aurora TSUBASA architecture.

GenHITS: A network science approach to driver gene detection in human regulatory network using gene’s influence evaluation

Cancer is among the diseases causing death, in which, cells uncontrollably grow and reproduce beyond the cell regulatory mechanism. In this disease, some genes are initiators of abnormalities and then transmit them to other genes through protein interactions. Accordingly, these genes are known as cancer driver genes (CDGs). In this regard, several methods have been previously developed for identifying cancer driver genes. Most of these methods are computational-based, which use the concept of mutation to predict CDGs. In this research, a method has been proposed for identifying CDGs in the transcription regulatory network using the concept of influence diffusion and by modifying the Hyperlink-Induced Topic Search algorithm based on the diffusion concept. Due to the type of these networks and the processes of abnormality progression in cells and the formation of cancerous tumors, high-influence genes can be the most likely considered as the driver genes. Therefore, we can use the influence diffusion concept as an acceptable theory to identify these genes. Recently, a method has been proposed to detect CDGs with the concept of the influence maximization. One of the challenges in these types of networks is finding the power of regulatory interaction between genes. Moreover, we have proposed a novel method to calculate the weight of regulatory interactions, based on the concept of diffusion. The performance of the proposed method was compared with other seventeen computational and network tools. Correspondingly, three cancer types were used as benchmarks as follows: breast invasive carcinoma (BRCA), Colon adenocarcinoma (COAD), and lung squamous cell carcinoma (LUSC). In addition, to determine the accuracy of the detected drivers using each method, CGC (Cancer Gene Census) and Mut-driver gene lists were utilized as gold standard. The results show that GenHITS performs better compared to the most of the other computational and network methods. Besides, it is also able to identify genes that have been identified by none of the other methods yet.

A HITS-based model for facility location decision

This paper presents a model to the facility location problem which incorporates both location characteristics and rival effects within a business cluster. Location characteristics in the model are predefined selection criteria such as rent, distance, safety and size. Rival effects represent the competition and collaboration relationship among close facility locations in the business cluster where candidate locations are located. The model presented uses data envelopment analysis (DEA) and weighted hyperlink-induced topic search (HITS) algorithm. DEA determines the efficient and inefficient locations as well as benchmarking relationship. The weighted HITS algorithm with the distance parameter considers the rival effects among locations to identify hubs and authorities. The applicability of our proposed model is demonstrated with a case study that rank apartments in the University area.

Media Roles in the Online News Domain: Authorities and Emergent Audience Brokers

This article empirically tests the role of legacy and digital-born news media, mapping the patterns of audience navigation across news sources and the relationship between news providers. We borrow tools from network science to bring evidence that suggest legacy news media retain control of the most central positions in the online news domain. Great progress has been made in discussing theoretically the impact of the Internet on the news media ecology. Less research attention, however, has been given to empirically testing changes in the role of legacy media and the rising prominence of digital-born outlets. To fill this gap, in this study we use the hyperlink-induced topic search algorithm, which identifies authorities by means of a hyperlink network, to show that legacy media are still the most authoritative sources in the media ecology. To further substantiate their dominant role, we also examine the structural position of news providers in the audience network. We gather navigation data from a panel of 30,000 people and use it to reproduce the network of patterns of news consumption. While legacy news media retain control of the brokerage positions for the general population, our analysis—focused on patterns of young news consumers—reveals that new digital outlets also occupy relevant positions to control the audience flow. The results of this study have substantive implications for our understanding of news organizations’ roles and how they attain authority in the digital age.

A novel method to predict essential proteins based on tensor and HITS algorithm

BackgroundEssential proteins are an important part of the cell and closely related to the life activities of the cell. Hitherto, Protein-Protein Interaction (PPI) networks have been adopted by many computational methods to predict essential proteins. Most of the current approaches focus mainly on the topological structure of PPI networks. However, those methods relying solely on the PPI network have low detection accuracy for essential proteins. Therefore, it is necessary to integrate the PPI network with other biological information to identify essential proteins.ResultsIn this paper, we proposed a novel random walk method for identifying essential proteins, called HEPT. A three-dimensional tensor is constructed first by combining the PPI network of Saccharomyces cerevisiae with multiple biological data such as gene ontology annotations and protein domains. Then, based on the newly constructed tensor, we extended the Hyperlink-Induced Topic Search (HITS) algorithm from a two-dimensional to a three-dimensional tensor model that can be utilized to infer essential proteins. Different from existing state-of-the-art methods, the importance of proteins and the types of interactions will both contribute to the essential protein prediction. To evaluate the performance of our newly proposed HEPT method, proteins are ranked in the descending order based on their ranking scores computed by our method and other competitive methods. After that, a certain number of the ranked proteins are selected as candidates for essential proteins. According to the list of known essential proteins, the number of true essential proteins is used to judge the performance of each method. Experimental results show that our method can achieve better prediction performance in comparison with other nine state-of-the-art methods in identifying essential proteins.ConclusionsThrough analysis and experimental results, it is obvious that HEPT can be used to effectively improve the prediction accuracy of essential proteins by the use of HITS algorithm and the combination of network topology with gene ontology annotations and protein domains, which provides a new insight into multi-data source fusion.

Filtering Instagram Hashtags Through Crowdtagging and the HITS Algorithm

Instagram is a rich source for mining descriptive tags for images and multimedia in general. The tags–image pairs can be used to train automatic image annotation (AIA) systems in accordance with the learning by example paradigm. In previous studies, we had concluded that, on average, 20% of the Instagram hashtags are related to the actual visual content of the image they accompany, i.e., they are descriptive hashtags, while there are many irrelevant hashtags, i.e., stop-hashtags, that are used across totally different images just for gathering clicks and for searchability enhancement. In this paper, we present a novel methodology, based on the principles of collective intelligence that helps in locating those hashtags. In particular, we show that the application of a modified version of the well-known hyperlink-induced topic search (HITS) algorithm, in a crowdtagging context, provides an effective and consistent way for finding pairs of Instagram images and hashtags, which lead to representative and noise-free training sets for content-based image retrieval. As a proof of concept, we used the crowdsourcing platform Figure-eight to allow collective intelligence to be gathered in the form of tag selection (crowdtagging) for Instagram hashtags. The crowdtagging data of Figure-eight are used to form bipartite graphs in which the first type of nodes corresponds to the annotators and the second type to the hashtags they selected. The HITS algorithm is first used to rank the annotators in terms of their effectiveness in the crowdtagging task and then to identify the right hashtags per image.

Conferences Events Suggestion Using Ranked Based Hyperlink-Induced Topic Search Algorithm

Conferences Events Suggestion Using Ranked Based Hyperlink-Induced Topic Search Algorithm

Identification of Essential Proteins Based on Improved HITS Algorithm.

Essential proteins are critical to the development and survival of cells. Identifying and analyzing essential proteins is vital to understand the molecular mechanisms of living cells and design new drugs. With the development of high-throughput technologies, many protein–protein interaction (PPI) data are available, which facilitates the studies of essential proteins at the network level. Up to now, although various computational methods have been proposed, the prediction precision still needs to be improved. In this paper, we propose a novel method by applying Hyperlink-Induced Topic Search (HITS) on weighted PPI networks to detect essential proteins, named HSEP. First, an original undirected PPI network is transformed into a bidirectional PPI network. Then, both biological information and network topological characteristics are taken into account to weighted PPI networks. Pieces of biological information include gene expression data, Gene Ontology (GO) annotation and subcellular localization. The edge clustering coefficient is represented as network topological characteristics to measure the closeness of two connected nodes. We conducted experiments on two species, namely Saccharomyces cerevisiae and Drosophila melanogaster, and the experimental results show that HSEP outperformed some state-of-the-art essential proteins detection techniques.

A New Journal Ranking Method: The Reputation Analysis of Citation Behavior Model

The authority of journals is usually a key indicator in guiding people to choose important journals, and the journal ranking is a common way to distinguish the journals' authority. The common ranking methods or models are based on citation data, such as the impact factor (IF), the PageRank algorithm, and the hyperlink-induced topic search algorithm. In this paper, we present a new model, named the reputation analysis of citation behavior (RACB) model, which not only considers the number of citations but also considers the reputation of inter-citation behavior. First, the model fits the function relation of the citation desire index (CDI) based on the citation data of the most reputable journal (named the top journal in the remainder of this paper). Then, the CDI values of the target journals are calculated by using the function relation of the top journal. An improved gray correlation analysis is used to describe the deviation between the CDI and the actual citation rate, named the random citation rate, indicating the inter-citation reputation value of the target journals. Finally, a case study showed that the ranking result of the RACB model has a high similarity with the IF values of 2017, which indicates that the ranking results of the RACB model that considered journal reputation evaluations are more reasonable than those of the traditional PageRank algorithm. This paper proposes a new method of journal ranking from the perspective of reputation evaluation, which can rank the journals more reasonably. It encourages researchers to cite articles more fairly in order to avoid a situation in which the low reputation of the citation behavior affects the quality of the journal that published the article.

Recommendations for Mobile Apps Based on the HITS Algorithm Combined With Association Rules

With the increasing popularity of intelligent devices, the mobile apps market has exploded. Due to a large number of candidate app services, it has become very difficult for users to choose the mobile apps that he/she wants to install. Therefore, it is crucial to improve users' experience and make personalized recommendations. In some cases, the traditional recommendation methods can be convenient, but they still have some shortcomings, resulting in inaccurate recommendations in general. To address this issue, this paper proposes a method for mobile app recommendations that are based on the Hyperlink-Induced Topic Search (HITS) algorithm combined with association rules. This method integrates the authority and hub scores into the candidate applications through the download and rating information, and it not only considers the importance of mobile apps in association rules but also takes the reliability factor of users into account. Experiments with the Huawei application market datasets show that the proposed method significantly improves the recommendation accuracies compared with the traditional methods.

Measuring the systemic importance of interconnected industries in the world economic system

PurposeThe purpose of this paper is to measure the systemic importance of industry in the world economic system under the system-wide event – the crisis of 2008-2009, by viewing this system as a weighted directed network of interconnected industries.Design/methodology/approachFirst, the authors investigate this crisis at three different levels based on network-related indicators: the “macro” global level, the “meso” country level, and the “micro” industry level. This investigation not only provides evidence for the systemic influence, that is, systemic risk, of the crisis, but also reveals the contagion mechanism of the crisis, which supports the stress testing. Second, the authors use a network-related business intelligence algorithm, the combined hyperlink-induced topic search (HITS) algorithm, to measure the contribution of a given individual industry to the overall risk of the economic system or, in other words, the systemic importance of the individual industry.FindingsThe HITS algorithm considers both the market information and the interconnectedness of the industries. Based on the stress testing, the performance of the combined HITS is compared with the purely market-based systemic risk measurement. The results show that the combined HITS outperforms the baseline in finding the top N systemically important industries.Practical implicationsThe combined HITS algorithm provides a novel network-based perspective of systemic risk measurement.Originality/valueMeasuring the systemic importance based on the combined HITS algorithm can help managers and regulators design effective risk management policies. In this respect, the work initiates a research direction of studying the systemic risk in a business system based on a network-related business intelligence algorithm because the business system can be viewed as an interconnected network.

Combining hyperlink-induced topic search and Bayesian approach for personalised manufacturing service recommendation

The explosive growth of Web service technologies for integrating virtual enterprises has resulted in the serious information overload. Thus the users are facing increasing difficulty in selecting the correct manufacturing services from the vast amount provided or recommended by collaborative partners for service-oriented supply chain deployment. Therefore, in this paper, a novel approach is presented for recommending personalised manufacturing services by combining a Hyperlink-Induced Topic Search (HITS) algorithm and the Bayesian approach. The personalised service recommendation problem is modelled to determine the optimal manufacturing services that are most probably the best selections to user preferences for some known manufacturing services. Further, the Bayesian approach decomposes such a problem of posterior probability into two sub-problems: the prior probability of a manufacturing service hypothesis before considering any user preferences, and the conditional probability of considering the aggregated preferences for some known manufacturing services as evidences when the user actually wants an unknown most preferable manufacturing service. Next, the personalised HITS algorithm is adapted to the network of service-oriented supply chain to rank the authority scores of manufacturing services that determine the relative probabilities of service execution through personalised trust propagation. The experimental results show that the proposed method can produce more accurate manufacturing service recommendation results than the existing approaches.

A time-aware hyperlink-induced topic search-based reputation evaluation method for optimal manufacturing service recommendation in distributed peer-to-peer networks

With the explosive growth of manufacturing services on a global scale, it has become an enormous challenge to recommend optimal manufacturing service in distributed manufacturing environments. Reputation evaluation is a key step to make optimal manufacturing service recommendation. However, the reputation values calculated through conventional reputation evaluation methods are static and local, and have a severe impact on the effectiveness of manufacturing service recommendation. This article presents a novel and effective reputation evaluation method for optimal manufacturing service recommendation in distributed peer-to-peer networks. The main contribution of the proposed approach is that it makes it possible to calculate the global and dynamic reputation values of enterprises by employing a time-aware hyperlink-induced topic search algorithm. This algorithm not only disseminates the reputation value of transaction enterprises, but also takes account of the temporal dimension in the process of computing reputation values. A concrete example demonstrates the feasibility of calculating reputation values of the proposed method, and an evaluation experiment reveals that both of recall rate and precision rate are improved, employing our proposed method for recommending manufacturing service.

A novel page ranking algorithm based on triadic closure and hyperlink-induced topic search

The Hyperlink-Induced Topic Search (HITS) algorithm developed by Jon Kleinberg made use of the link structure of the web pages on the Web in order to discover and rank web pages being relevant to a particular topic. However it only took account of the hyperlink structure, while completely excluded contents of web pages, and it ignored the fact that degrees of the importance of many hyperlinks on the Web may be different. In this paper, to overcome the topic drifts, we proposed a novel page ranking algorithm combining the hyperlink with the triadic closure theory by considering fully the Vector Space Model (VSM) and the TrustRank algorithm. The method firstly computed the relevance between two randomly arbitrary web pages based on web page topic similarity and common reference degree. Then, by using that model as a point of reference, a new adjacency matrix was constructed to iteratively calculate the authority and hub values of web pages. Next, we calculated the trust-degree for each web page in the basic set by the trust-score algorithm. Finally, the score for each web page is computed by linearly merging the authority and the trust-degree. In our experiments, we used five classic HITS-based algorithms to compare with our proposed page ranking algorithm-PCTHITS (Web Page Topic Similarity, Common Reference Degree, Trust-degree) algorithm. The experimental results demonstrated that our proposed algorithm outperform the other four classic improved algorithms and HITS algorithm.

Research on Ranking Algorithms in Web Structure Mining

In the past few decades, the Web has emerged as a treasure of information and web mining is a technique to handle this treasure. During recent years web mining has been a well-researched area. Web mining is the application of the data mining which is useful to extract the knowledge from web. With the progress of web, more and more data are now available for users on web. Web structure mining deals with the contents and hyperlinks on web pages. In this review paper, we have focused on three basic algorithms for evaluating the importance of pages i.e. Page Rank, Weighted Page Rank, and Hyperlink-Induced Topic Search and comparison of those algorithms. Page Rank algorithm is based on back links of the page and it calculates the rank of web pages at indexing time. Weighted Page Rank algorithm scores pages according to their relevancies and rank of a page is calculated by its number of incoming and outgoing links. Hyperlink-induced topic search algorithm is an iterative algorithm developed to quantify each pages value as an authority and as a hub. This study was done basically to explore the link structure algorithms for ranking pages.

Hubs and authorities in the world trade network using a weighted HITS algorithm.

We investigate the economic hubs and authorities of the world trade network (WTN) from to , an era of rapid economic globalization. Using a well-defined weighted hyperlink-induced topic search (HITS) algorithm, we can calculate the values of the weighted HITS hub and authority for each country in a conjugate way. In the context of the WTN, authority values are large for countries with significant imports from large hub countries, and hub values are large for countries with significant exports to high-authority countries. The United States was the largest economic authority in the WTN from to . The authority value of the United States has declined since , and China has now become the largest hub in the WTN. At the same time, China's authority value has grown as China is transforming itself from the “factory of the world” to the “market of the world.” European countries show a tendency to trade mostly within the European Union, which has decreased Europe's hub and authority values. Japan's authority value has increased slowly, while its hub value has declined. These changes are consistent with Japan's transition from being an export-driven economy in its high economic growth era in the latter half of the twentieth century to being a more mature, economically balanced nation.

MCHITS: Monte Carlo based Method for Hyperlink Induced Topic Search on Networks

Hyperlink Induced Topic Search (HITS) is the most authoritative and most widely used personalized ranking algorithm on networks. The HITS algorithm ranks nodes on networks according to power iteration, and has high complexity of computation. This paper models the HITS algorithm with the Monte Carlo method, and proposes Monte Carlo based algorithms for the HITS computation. Theoretical analysis and experiments show that the Monte Carlo based approximate computing of the HITS ranking reduces computing resources a lot while keeping higher accuracy, and is significantly better than related works

Study on theme-drift of hyperlink-induced topic search algorithm

Hyperlink-Induced Topic Search (HITS) algorithm is a classic hyperlink-based algorithm. But the HITS algorithm is purely based on the hyperlink,and it ignores the text of the linked page and does not distinguish the importance between the different hyperlinks. Because of this,a theme-drift phenomenon often happens when using HITS algorithm. The improved algorithm based on the HITS algorithm makes use of the classic tf-idf algorithm to calculate the related weight between the linked page and the query. The improved algorithm can make the search engine ranking results more in line with the query,and the corresponding precision rate has also been greatly improved.