Building taxonomy of web search intents for name entity queries.

Automatic extraction of clickable structured web contents for name entity queries.

Exploring optimization of semantic relationship graph for multi%2Drelational Bayesian classification.

Multiple intents re%2Dranking.

Exploring optimization of semantic relationship graph for multi%2Drelational Bayesian classification.

Exploring the Power of Heuristics and Links in Multi%2Drelational Data Mining.

BibNetMiner mining bibliographic information networks.

Truth Discovery with Multiple Conflicting Information Providers on the Web.
The world-wide web has become the most important information source for most of us. Unfortunately, there is no guarantee for the correctness of information on the web. Moreover, different web sites often provide conflicting information on a subject, such as different specifications for the same product. In this paper we propose a new problem called Veracity, i.e., conformity to truth, which studies how to find true facts from a large amount of conflicting information on many subjects that is provided by various web sites. We design a general framework for the Veracity problem, and invent an algorithm called TruthFinder, which utilizes the relationships between web sites and their information, i.e., a web site is trustworthy if it provides many pieces of true information, and a piece of information is likely to be true if it is provided by many trustworthy web sites. Our experiments show that TruthFinder successfully finds true facts among conflicting information, and identifies trustworthy web sites better than the popular search engines.
Scalable mining and link analysis across multiple database relations.
Relational databases are the most popular repository for structured data, and are thus one of the richest sources of knowledge in the world. In a relational database, multiple relations are linked together via entityrelationship links. Unfortunately, most existing data mining approaches can only handle data stored in single tables, and cannot be applied to relational databases. Therefore, it is an urgent task to design data mining approaches that can discover knowledge from multi-relational data. In this thesis we study three most important data mining tasks in multi-relational environments: classification, clustering, and duplicate detection. Since information is widely spread across multiple relations, the most crucial and common challenge in multi-relational data mining is how to utilize the relational information linked with each object. We rely on two types of information,  — neighbor tuples and linkages between objects, to analyze the properties of objects and relationships among them. Because of the complexity of multi-relational data, efficiency and scalability are two major concerns in multi-relational data mining. In this thesis we propose scalable and accurate approaches for each data mining task studied. In order to achieve high efficiency and scalability, the approaches utilize novel techniques
Object Distinction Distinguishing Objects with Identical Names.

Truth discovery with multiple conflicting information providers on the web.
The world-wide web has become the most important information source for most of us. Unfortunately, there is no guarantee for the correctness of information on the web. Moreover, different web sites often provide conflicting information on a subject, such as different specifications for the same product. In this paper we propose a new problem called Veracity, i.e., conformity to truth, which studies how to find true facts from a large amount of conflicting information on many subjects that is provided by various web sites. We design a general framework for the Veracity problem, and invent an algorithm called TruthFinder, which utilizes the relationships between web sites and their information, i.e., a web site is trustworthy if it provides many pieces of true information, and a piece of information is likely to be true if it is provided by many trustworthy web sites. Our experiments show that TruthFinder successfully finds true facts among conflicting information, and identifies trustworthy web sites better than the popular search engines.
CrossClus user%2Dguided multi%2Drelational clustering.

VisFlowCluster%2DIP Connectivity%2DBased Visual Clustering of Network Hosts.
With the increasing number of hostile network attacks, anomaly  detection for network security has become an urgent task. As there have  not been highly e#ective solutions for automatic intrusion detection, especially  for detecting newly emerging attacks, network tra#c visualization  has become a promising technique for assisting network administrators  to monitor network tra#c and detect abnormal behaviors.
LinkClus Efficient Clustering via Heterogeneous Semantic Links.
Data objects in a relational database are cross-linked with each other via multi-typed links. Links contain rich semantic information that may indicate important relationships among objects. Most current clustering methods rely only on the properties that belong to the objects per se. However, the similarities between objects are often indicated by the links, and desirable clusters cannot be generated using only the properties of objects. In this paper we explore linkage-based clustering, in which the similarity between two objects is measured based on the similarities between the objects linked with them. In comparison with a previous study (SimRank) that computes links recursively on all pairs of objects, we take advantage of the power law distribution of links, and develop a hierarchical structure called SimTree to represent similarities in multi-granularity manner. This method avoids the high cost of computing and storing pairwise similarities but still thoroughly explore relationships among objects. An efficient algorithm is proposed to compute similarities between objects by avoiding pairwise similarity computations through merging computations that go through the same branches in the SimTree. Experiments show the proposed approach achieves high efficiency, scalability, and accuracy in clustering multi-typed linked objects. 1.
Efficient Classification across Multiple Database Relations A CrossMine Approach.



The Design of VisFlowConnect%2DIP A Link Analysis System for IP Security Situational Awareness.
Visualization of IP-based traffic dynamics on networks is a challenging task due to large data volume and the complex, temporal relationships between hosts. We present the architecture of VisFlowConnect-IP, a powerful new tool to visualize IP network traffic flow dynamics for security situational awareness. VisFlowConnect-IP allows an operator to visually assess the connectivity of large and complex networks on a single screen. It provides an overall view of the entire network and filter/drill-down features that allow operators to request more detailed information. Preliminary reports from several organizations using this tool report increased responsiveness to security events as well as new insights into understanding the security dynamics of their networks. In this paper we focus specifically on the design decisions made during the VisFlowConnect development process so that others may learn from our experience. The current VisFlowConnect architecture---the result of these design decisions---is extensible to processing other highvolume multi-dimensional data streams where link connectivity /activity is a focus of study. We report experimental results quantifying the scalability of the underlying algorithms for representing link analysis given continuous highvolume traffic flows as input.
Monitormining creating domain knowledge for system automation using a gray%2Dbox approach.

Cross%2Drelational clustering with user%27s guidance.

Efficient Classification from Multiple Heterogeneous Databases.
With the fast expansion of computer networks, it is inevitable to study data mining on heterogeneous databases. In this paper we propose MDBM, an accurate and efficient approach for classification on multiple heterogeneous databases. We propose a regression-based method for predicting the usefulness of inter-database links that serve as bridges for information transfer, because such links are automatically detected and may or may not be useful or even valid. Because of the high cost of inter-database communication, MDBM employs a new strategy for cross-database classification, which finds and performs actions with high benefit-to-cost ratios. The experiments show that MDBM achieves high accuracy in cross-database classification, with much higher efficiency than previous approaches.
Searching for Related Objects in Relational Databases.
To discover knowledge or retrieve information from a relational database, a user often needs to find objects related to certain source objects. There are two main challenges in building an e#ective object search system: the huge amount of objects in the database and the large number of di#erent relationships between objects. In this paper we introduce Ross, an e#cient and accurate relational object search system. Ross accepts complex queries that enable users to specify the relationships among objects. To measure the relationships of join paths, Ross considers the di#erent semantics of di#erent joins, and combines both selectivity and lengths of join paths to measure their strengths. A novel approach is used to find the best join paths between relations, which converts the database schema into a graph, so that the shortest paths in the graph correspond to best join paths in the database. Ross uses a stream-based system architecture to handle complex queries containing logical operators, which can find the most related objects upon users&#039; requests. Comprehensive experiments are conducted to show the high scalability and e#ectiveness of Ross.
Outsourcing Internet Security Economic Analysis of Incentives for Managed Security Service Providers.
Firms hesitate to outsource their network security to outside  security providers (called Managed Security Service Providers or  MSSPs) because an MSSP may shirk secretly to increase profits. In  economics this secret shirking behavior is commonly referred to as the  Moral Hazard problem. There is a counter argument that this moral  hazard problem is not as significant for the Internet security outsourcing  market because MSSPs work hard to build and maintain their reputations  which are crucial to surviving competition. Both arguments make  sense and should be considered to write a successful contract. This paper  studies the characteristics of optimal contracts (payment to MSSPs)  for security outsourcing market by setting up an economic framework  that combines both e#ects. It is shown that an optimal contract should  be performance-based. The degree of performance dependence decreases  if the reputation e#ect becomes more significant. We also show that if  serving a large group of customers helps the provider to improve service  quality significantly (which is observed in the internet security outsourcing  market), an optimal contract should always be performance-based  even if a strong reputation e#ect exists.
CrossMine Efficient Classification Across Multiple Database Relations.
Most of today&#039;s structured data is stored in relational databases. Such a database consists of multiple relations which are linked together conceptually via entity-relationship links in the design of relational database schemas. Multi-relational classification can be widely used in many disciplines, such as financial decision making, medical research, and geographical applications. However, most classification approaches only work on single &#034;flat&#034; data relations. It is usually difficult to convert multiple relations into a single flat relation without either introducing huge, undesirable &#034;universal relation&#034; or losing essential information. Previous works using Inductive Logic Programming approaches (recently also known as Relational Mining) have proven effective with high accuracy in multi-relational classification. Unfortunately, they suffer from poor scalability w.r.t. the number of relations and the number of attributes in databases.
CrossMine Efficient Classification Across Multiple Database Relations.
Most of today&#039;s structured data is stored in relational databases. Such a database consists of multiple relations which are linked together conceptually via entity-relationship links in the design of relational database schemas. Multi-relational classification can be widely used in many disciplines, such as financial decision making, medical research, and geographical applications. However, most classification approaches only work on single &#034;flat&#034; data relations. It is usually difficult to convert multiple relations into a single flat relation without either introducing huge, undesirable &#034;universal relation&#034; or losing essential information. Previous works using Inductive Logic Programming approaches (recently also known as Relational Mining) have proven effective with high accuracy in multi-relational classification. Unfortunately, they suffer from poor scalability w.r.t. the number of relations and the number of attributes in databases.
VisFlowConnect netflow visualizations of link relationships for security situational awareness.

Semantic image clustering using relevance feedback.

CPAR Classification based on Predictive Association Rules.
Recent studies in data mining have proposed a new classification approach, called associative classification, which, according to several reports, such as [7, 6], achieves higher classification accuracy than traditional classification approaches such as C4.5. However, the approach also su#ers from two major deficiencies: (1) it generates a very large number of association rules, which leads to high processing overhead; and (2) its confidence-based rule evaluation measure may lead to overfitting.
A New Textual%2FNon%2DTextual Classifier for Document Skew Correction.

