Researchers usually extract domain experts only through analyzing network structure or partitioning users into several communities according to their label information. Combining structure and content to discovery domain experts is a new attempt. Motivated by that, this paper proposes a domain expert discovery method based on network structure and content semantics, called DEDSC, which can extract authority nodes in overlapping communities. To analyze the overall authority for each user in the social network, two definitions, structure authority value and content authority value, are proposed to evaluate the authority of users in different perspectives. Partitioning users into communities can make the results more accurate. Experimental results show that our proposed method can discover domain experts effectively. In addition, when we need to extract domain experts in a new test dataset, we do not need to re-train the data in the training dataset.

Heterogeneous networks are ubiquitous. People like to discover rare but meaningful objects and patterns from such networks. Regardless of high structure similarity or high content similarity, the corresponding objects can be used in data analysis. However, the vast differences between structure and contents should be paid more attention. In this paper, we propose an outlier correlation detection method, called Structure2Content, which discovers outlier correlation incrementally in structure-level and content-level. Structure2Content addresses three important challenges: (1) how can we measure the target object’s structure and content similarity? (2) how can we find the representative features of target objects? (3) how can we insert new data or delete the obsoleted data incrementally. To tackle these challenges, Structure2Content applies four main techniques: (1) two matrices are used to store structure and content similarity, respectively, (2) 3-tuples are used to represent the closeness degree between objects, (3) a mirror step and an iterative process are combined to obtain the top-K outlier correlations, and (4) only updating 3-tuples can help insert or delete data incrementally instead of training all data from the beginning. Substantial experiments show that our proposed method is very effective for outlier correlations detection.

Summary Learning from positive and unlabeled examples (PU learning) is a partially supervised classification that is frequently used in Web and text retrieval system. The merit of PU learning is that it can get good performance with less manual work. Motivated by transfer learning, this paper presents a novel method that transfers the ‘outdated data’ into the process of PU learning. We first propose a way to measure the strength of the features and select the strong features and the weak features according to the strength of the features. Then, we extract the reliable negative examples and the candidate negative examples using the strong and the weak features (Transfer‐1DNF). Finally, we construct a classifier called weighted voting iterative support vector machine (SVM) that is made up of several subclassifiers by applying SVM iteratively, and each subclassifier is assigned a weight in each iteration. We conduct the experiments on two datasets: 20 Newsgroups and Reuters‐21578, and compare our method with three baseline algorithms: positive example‐based learning, weighted voting classifier and SVM. The results show that our proposed method Transfer‐1DNF can extract more reliable negative examples with lower error rates, and our classifier outperforms the baseline algorithms. Copyright © 2016 John Wiley & Sons, Ltd.

Today more and more information on the Web makes it difficult to get domain-specific information due to the huge amount of data sources and the keywords that have few features. Anchor texts, which contain a few features of a specific topic, play an important role in domain-specific information retrieval, especially in Web page classification. However, the features contained in anchor texts are not informative enough. This paper presents a novel incremental method for Web page classification enhanced by link-contexts and clustering. Directly applying the vector of anchor text to a classifier might not get a good result because of the limited amount of features. Link-context is used first to obtain the contextual information of the anchor text. Then, a hierarchical clustering method is introduced to cluster feature vectors and content unit, which increases the length of a feature vector belonging to one specific class. Finally, incremental SVM is proposed to get the final classifier and increase the accuracy and efficiency of a classifier. Experimental results show that the performance of our proposed method outperforms the conventional topical Web crawler in Harvest rate and Target recall.

SUMMARY Term frequency–inverse document frequency (TF–IDF), one of the most popular feature (also called term or word) weighting methods used to describe documents in the vector space model and the applications related to text mining and information retrieval, can effectively reflect the importance of the term in the collection of documents, in which all documents play the same roles. But, TF–IDF does not take into account the difference of term IDF weighting if the documents play different roles in the collection of documents, such as positive and negative training set in text classification. In view of the aforementioned text, this paper presents a novel TF–IDF‐improved feature weighting approach, which reflects the importance of the term in the positive and the negative training examples, respectively. We also build a weighted voting classifier by iteratively applying the support vector machine algorithm and implement one‐class support vector machine and Positive Example Based Learning methods used for comparison. During classifying, an improved 1‐DNF algorithm, called 1‐DNFC, is also adopted, aiming at identifying more reliable negative documents from the unlabeled examples set. The experimental results show that the performance of term frequency inverse positive–negative document frequency‐based classifier outperforms that of TF–IDF‐based one, and the performance of weighted voting classifier also exceeds that of one‐class support vector machine‐based classifier and Positive Example Based Learning‐based classifier. Copyright © 2013 John Wiley & Sons, Ltd.

Concurrency and Computation: Practice and Experience 20(1):61–74 (January 2008) (DOI: 10.1002/cpe.1211) A bug was identified after publication in the testing program of this article. Figures 6 and 7 were subsequently affected. The correct results are shown as follows: Dynamic plot of average harvest rate versus number of crawled pages. Performance is averaged across topics and standard errors are also shown. The error bars correspond to ± standard error. Dynamic plot of average target recall versus number of crawled pages. Performance is averaged across topics and standard errors are also shown. The error bars correspond to ± standard error.

Abstract The complexity of web information environments and multiple‐topic web pages are negative factors significantly affecting the performance of focused crawling. A highly relevant region in a web page may be obscured because of low overall relevance of that page. Segmenting the web pages into smaller units will significantly improve the performance. Conquering and traversing irrelevant page to reach a relevant one (tunneling) can improve the effectiveness of focused crawling by expanding its reach. This paper presents a heuristic‐based method to enhance focused crawling performance. The method uses a Document Object Model (DOM)‐based page partition algorithm to segment a web page into content blocks with a hierarchical structure and investigates how to take advantage of block‐level evidence to enhance focused crawling by tunneling. Page segmentation can transform an uninteresting multi‐topic web page into several single topic context blocks and some of which may be interesting. Accordingly, focused crawler can pursue the interesting content blocks to retrieve the relevant pages. Experimental results indicate that this approach outperforms Breadth‐First, Best‐First and Link‐context algorithm both in harvest rate, target recall and target length. Copyright © 2007 John Wiley & Sons, Ltd.