Xiwu Gu

LIMTopic: A Framework of Incorporating Link Based Importance into Topic Modeling

Topic modeling has become a widely used tool for document management. However, there are few topic models distinguishing the importance of documents on different topics. In this paper, we propose a framework LIMTopic to incorporate link based importance into topic modeling. To instantiate the framework, RankTopic and HITSTopic are proposed by incorporating topical pagerank and topical HITS into topic modeling respectively. Specifically, ranking methods are first used to compute the topical importance of documents. Then, a generalized relation is built between link importance and topic modeling. We empirically show that LIMTopic converges after a small number of iterations in most experimental settings. The necessity of incorporating link importance into topic modeling is justified based on KL-Divergences between topic distributions converted from topical link importance and those computed by basic topic models. To investigate the document network summarization performance of topic models, we propose a novel measure called log-likelihood of ranking-integrated document-word matrix. Extensive experimental results show that LIMTopic performs better than baseline models in generalization performance, document clustering and classification, topic interpretability and document network summarization performance. Moreover, RankTopic has comparable performance with relational topic model (RTM) and HITSTopic performs much better than baseline models in document clustering and classification.

Detecting network communities using regularized spectral clustering algorithm

The progressively scale of online social network leads to the difficulty of traditional algorithms on detecting communities. We introduce an efficient and fast algorithm to detect community structure in social networks. Instead of using the eigenvectors in spectral clustering algorithms, we construct a target function for detecting communities. The whole social network communities will be partitioned by this target function. We also analyze and estimate the generalization error of the algorithm. The performance of the algorithm is compared with the standard spectral clustering algorithm, which is applied to different well-known instances of social networks with a community structure, both computer generated and from the real world. The experimental results demonstrate the effectiveness of the algorithm.

An Efficient SSD-based Hybrid Storage Architecture for Large-Scale Search Engines

Large-scale search engines use hard disk drives (HDD) to store the mass index data for their capacity, whose performances are limited by the relatively low I/O performance of HDD. Caching is an effective optimization, and many caching algorithms have been proposed to improve retrieval performance. Considering the high cost of memory and huge amounts of data, the limited capacity of cache in memory cannot resolve the above problem thoroughly. In this paper, we adopt a solid state disk (SSD) based storage architecture, which uses SSD as a secondary cache for memory. We analyze the I/O patterns of search engines and propose SSD-based data management policies based on the hybrid storage architecture, including data selection, data placement and data replacement. Our main goal is to improve the performance of search engines while reducing operation cost inside SSD. The experimental results demonstrate the proposed architecture improves the hit ratio by 13.31%, the performance by 41.05%, the average access time inside SSD by 43.83%, and reduces block erasure operations by 71.52%.

Optimizing Academic Conference Classification Using Social Tags

Automatically classifying academic conference into semantic topic promises improved academic search and browsing for users. Social tagging is an increasingly popular way of describing the topic of academic conference. However, no attention has been devoted to academic conference classification by making use of social tags. Motivated by this observation, this paper proposes a method which utilizes social tags as well as the content of academic conference in order to improve automatically identifying academic conference classification. The proposed method applies different automatic classification algorithms to improve classification quality by using social tags. Experimental results show that this method mentioned above performs better than the method which only utilizes the content to classify academic conference with 1% Precision measure score increase and 1.64% F1 measure score increase, which demonstrates the effectiveness of the proposed method.

Optimizing ranking method using social annotations based on language model

Recent research has shown that more and more web users utilize social annotations to manage and organize their interested resources. Therefore, with the growing popularity of social annotations, it is becoming more and more important to utilize such social annotations to achieve effective web search. However, using a statistical model, there are no previous studies that examine the relationships between queries and social annotations. Motivated by this observation, we use social annotations to re-rank search results. We intend to optimize retrieval ranking method by using the ranking strategy of integrating the query-annotation similarity into query-document similarity. Specifically, we calculate the query-annotation similarity by using a statistical language model, which in a shorter form we call simply a language model. Then the initial search results are re-ranked according to the computational weighted score of the query-document similarity score and the query-annotation similarity score. Experimental results show that the proposed method can improve the NDCG score by 8.13%. We further conduct an empirical evaluation of the method by using a query set including about 300 popular social annotations and constructed phrases. More generally, the optimized results with social annotations based on a language model can be of significant benefit to web search.

RMiner: a tool set for role mining

Recently, there are many approaches proposed for mining roles using automated technologies. However, it lacks a tool set that can be used to aid the application of role mining approaches and update role states. In this demonstration, we introduce a tool set, RMiner, which is based on the core of WEKA, an open source data mining tool. RMiner implements most of the classic and latest role mining algorithms and provides interactive tools for administrator to update role states. The running examples of RMiner are presented to demonstrate the effectiveness of the tool set.

Efficient Online Index Maintenance for SSD-based Information Retrieval Systems

Solid state disks (SSDs) can potentially eliminate the I/O bottleneck for many conventional applications. However, they have a very unique characteristic of erase-before-write, which probably makes existing index maintenance methods inapplicable to SSDs. In this paper, we propose Hybrid Merge, a new online index maintenance strategy for information retrieval systems, which applies SSDs instead of hard disk drives (HDDs) to store inverted indexes. We analyze the existing indexing methods through experiments, and design a new merge-based indexing method with no random writes. We try to take the full advantage of the SSDs fast random reads to overcome the defects of existing methods. Experimental results show that the proposed method improves indexing and query performance with extremely low write traffic compare to existing approaches.

Type-2 fuzzy description logic

Description logics (DLs) are widely employed in recent semantic web application systems. However, classical description logics are limited when dealing with imprecise concepts and roles, thus providing the motivation for this work. In this paper, we present a type-2 fuzzy attributive concept language with complements (ALC) and provide its knowledge representation and reasoning algorithms. We also propose type-2 fuzzy web ontology language (OWL) to build a fuzzy ontology based on type-2 fuzzy ALC and analyze the soundness, completeness, and complexity of the reasoning algorithms. Compared to type-1 fuzzy ALC, type-2 fuzzy ALC can describe imprecise knowledge more meticulously by using the membership degree interval. We implement a semantic search engine based on type-2 fuzzy ALC and carry out experiments on real data to test its performance. The results show that the type-2 fuzzy ALC can improve the precision and increase the number of relevant hits for imprecise information searches.

Topic-based ranking in Folksonomy via probabilistic model

Social tagging is an increasingly popular way to describe and classify documents on the web. However, the quality of the tags varies considerably since the tags are authored freely. How to rate the tags becomes an important issue. Most social tagging systems order tags just according to the input sequence with little information about the importance and relevance. This limits the applications of tags such as information search, tag recommendation, and so on. In this paper, we pay attention to finding the authority score of tags in the whole tag space conditional on topics and put forward a topic-sensitive tag ranking (TSTR) approach to rank tags automatically according to their topic relevance. We first extract topics from folksonomy using a probabilistic model, and then construct a transition probability graph. Finally, we perform random walk over the topic level on the graph to get topic rank scores of tags. Experimental results show that the proposed tag ranking method is both effective and efficient. We also apply tag ranking into tag recommendation, which demonstrates that the proposed tag ranking approach really boosts the performances of social-tagging related applications.

Role mining based on cardinality constraints

Role mining was recently proposed to automatically find roles among user‐permission assignments using data mining technologies. However, the current studies about role mining mainly focus on how to find roles, without considering the constraints that are essentially required in role‐based access control systems. In this paper, we present a role mining algorithm with constraints, especially for the cardinality constraints. We illustrate it is essential for role mining to take cardinality constraints into account, and introduce the concepts of the cardinality constraints of roles and permissions. We further propose a role mining algorithm to generate roles based on these two kinds of cardinality constraints. The algorithm uses graph theory to model the role mining problem and maps the relation of two roles to the relation of graph elements. We set an optimization goal for role mining and employ graph optimization theory to find roles that satisfy the aforementioned cardinality constraints. We carry out the experiments to evaluate our approach. The experimental results demonstrate the rationality and effectiveness of the proposed algorithm. Copyright