Missing Value Estimation for Mixed%2DAttribute Data Sets.

Cost Sensitive Classification in Data Mining.

Cost Sensitive Classification in Data Mining.

Web User Browse Behavior Characteristic Analysis Based on a BC Tree.

A New Classification Algorithm Using Mutual Nearest Neighbors.

Acquiring knowledge from inconsistent data sources through weighting.

Estimating Semi%2DParametric Missing Values with Iterative Imputation.

Combining kNN Imputation and Bootstrap Calibrated Empirical Likelihood for Incomplete Data Analysis.

Summary queries for frequent itemsets mining.

Cost%2Dsensitive classification with respect to waiting cost.

Strategies for complex data cube queries.

Confidence Metrics for Association Rule Mining.


POP algorithm Kernel%2Dbased imputation to treat missing values in knowledge discovery from databases.


Estimating confidence intervals for structural differences between contrast groups with missing data.

Mining class%2Dbridge rules based on rough sets.

Mining globally interesting patterns from multiple databases using kernel estimation.

A decremental algorithm of frequent itemset maintenance for mining updated databases.

Missing Data Analysis A Kernel%2DBased Multi%2DImputation Approach.

Secure Transaction Protocol Analysis Models and Applications

NIIA Nonparametric Iterative Imputation Algorithm.

Detecting Differences Between Contrast Groups.


Mining follow%2Dup correlation patterns from time%2Drelated databases.


Empirical likelihood confidence intervals for differences between two datasets with missing data.

Missing Value Imputation Based on Data Clustering.

Measuring the Uncertainty of Differences for Contrasting Groups.

Identifying Dependency Between Secure Messages for Protocol Analysis.

Cost%2DSensitive Imputing Missing Values with Ordering.

Identifying Dependency Between Secure Messages for Protocol Analysis.

Cost%2DTime Sensitive Decision Tree with Missing Values.

GBKII An Imputation Method for Missing Values.

Semi%2Dparametric optimization for missing data imputation.

On Data Structures for Association Rule Discovery.
Abstract. Data mining has been recognised as an essential element of decision support, which has increasingly become a focus of the database industry. Like all computationally expensive data analysis applications, for example Online Analytical Processing (OLAP), performance is a key factor for usefulness and acceptance in business. In the course of the CRITIKAL 1 project (Client-Server Rule Induction Technology for Industrial Knowledge Acquisition from Large Databases), which is funded by the European Commission, several kinds of architectures for data mining were evaluated with a strong focus on high performance. Speci-cally, the data mining techniques association rule discovery and decision tree induction were implemented into a prototype. We present the architecture developed by the CRITIKAL consortium and compare it to alternative architectures. 1
EDUA An efficient algorithm for dynamic database mining.

Editorial Special issue on mining low%2Dquality data.

Editorial Special issue on mining low%2Dquality data.

Cost%2DSensitive Test Strategies.
In medical diagnosis doctors must often determine what medical tests (e.g., X-ray, blood tests) should be ordered for a patient to minimize the total cost of medical tests and misdiagnosis. In this paper, we design cost-sensitive machine learning algorithms to model this learning and diagnosis process. Medical tests are like attributes in machine learning whose values may be obtained at cost (attribute cost), and misdiagnoses are like misclassifications which may also incur a cost (misclassification cost). We first propose an improved decision tree learning algorithm that minimizes the sum of attribute costs and misclassification costs. Then we design several novel “test strategies ” that may request to obtain values of unknown attributes at cost (similar to doctors ’ ordering of medical tests at cost) in order to minimize the total cost for test examples (new patients). We empirically evaluate and compare these test strategies, and show that they are effective and that they outperform previous methods. A case study on heart disease is given.
Kernel%2DBased Multi%2DImputation for Missing Data.

Efficient Frequent Itemsets Mining by Sampling.

Missing or absent%3F A Question in Cost%2Dsensitive Decision Tree.

Detecting Collusion Attacks in Security Protocols.

Adapting K%2DMeans Algorithm for Discovering Clusters in Subspaces.

Clustering Web Documents Based on Knowledge Granularity.

Enhancing DWT for Recent%2DBiased Dimension Reduction of Time Series Data.

Enhancing DWT for Recent%2DBiased Dimension Reduction of Time Series Data.

GridSM a Norm%2Ddriven State Machine Model of Grid Workflow.

Difference Detection Between Two Contrast Sets.

Identifying Follow%2DCorrelation Itemset%2DPairs.

Identifying bridging rules between conceptual clusters.

Efficient Computation of Multi%2Dfeature Data Cubes.
) Jiawei Han, Sonny H.S. Chee and Jenny Y. Chiang  Intelligent Database Systems Research Laboratory School of Computing Science, Simon Fraser University, British Columbia, Canada V5A 1S6  f han, schee, ychiang g@cs.sfu.ca  URL: http://db.cs.sfu.ca/  Abstract  Data warehouses and OLAP engines are expected to be widely available in the near future. The data in data warehouses has been cleansed, integrated, and preprocessed, and infrastructures have been built surrounding data warehouses for efficient data analysis. Therefore, data warehouses or OLAP databases are expected to be a major platform for data mining in the future. We discuss the issues related to efficient and effective data mining in large data warehouses and/or data marts, including the desired architectures for an integrated on-line analytical processing (OLAP) and on-line analytical mining (OLAM) system, the expected features of OLAM, and how to implement such a system effectively.  1 Introduction  With recent developments...
Is Frequency Enough for Decision Makers to Make Decisions%3F.

Optimized Parameters for Missing Data Imputation.

Mapping between Relational Database Schema and OWL Ontology for Deep Annotation.

Mining Multiple Data Sources Local Pattern Analysis.
Many large organizations process data from multiple data sources, such as the different branches of an interstate or international company. Also the Web has emerged as a large, distributed data repository consisting of a variety of data sources and formats. Although the data collected from the Web or multiple local datasets brings us opportunities in
A logical framework for identifying quality knowledge from different data sources.

An norm%2Ddriven state machine model for CSCW systems.

Generalized Dimension%2DReduction Framework for Recent%2DBiased Time Series Analysis.
Abstract—Recent-biased approximations have received increased attention recently as a mechanism for learning trend patterns from time series or data streams. They have shown promise for clustering time series and incrementally pattern maintaining. In this paper, we design a generalized dimension-reduction framework for recent-biased approximations, aiming at making traditional dimensionreduction techniques actionable in recent-biased time series analysis. The framework is designed in two ways: equi-segmented scheme and vari-segmented scheme. In both schemes, time series data are first partitioned into segments and a dimension-reduction technique is applied to each segment. Then, more coefficients are kept for more recent data while fewer kept for older data. Thus, more details are preserved for recent data and fewer coefficients are kept for the whole time series, which improves the efficiency greatly. We experimentally evaluate the proposed approach, and demonstrate that traditional dimension-reduction techniques, such as SVD, DFT, DWT, PIP, PAA, and landmarks, can be embedded into our framework for recent-biased approximations over streaming time series. Index Terms—Time series analysis, feature extraction or construction, data mining. 1
AI 2005 Advances in Artificial Intelligence

In%2DDepth Data Mining and Its Application in Stock Market.

In%2DDepth Data Mining and Its Application in Stock Market.

OSM an organizational state machine model for CSCW systems.

A Decremental Algorithm for Maintaining Frequent Itemsets in Dynamic Databases.
Abstract. Data mining and machine learning must confront the problem of pattern maintenance because data updating is a fundamental operation in data management. Most existing data-mining algorithms assume that the database is static, and a database update requires rediscovering all the patterns by scanning the entire old and new data. While there are many efficient mining techniques for data additions to databases, in this paper, we propose a decremental algorithm for pattern discovery when data is being deleted from databases. We conduct extensive experiments for evaluating this approach, and illustrate that the proposed algorithm can well model and capture useful interactions within data when the data is decreasing. 1
Dynamic Test%2DSensitive Decision Trees with Multiple Cost Scales.

A Recent%2DBiased Dimension Reduction Technique for Time Series Data.
Abstract. There are many techniques developed for tackling time series and most of them consider every part of a sequence equally. In many applications, however, recent data can often be much more interesting and significant than old data. This paper defines new recent-biased measures for distance and energy, and proposes a recent-biased technique based on DWT for time series in which more recent data are considered more significant. With such a recent-biased technique, the dimension of time series can be reduced while effectively preserving the recent-biased energy. Our experiments have demonstrated the effectiveness of the proposed approach for handling time series. 1
Identifying Interesting Patterns in Multidatabases.

Armga Identifying Interesting Association Rules with Genetic Algorithms.

A Partial%2DRepeatability Approach to Data Mining.
Mining textual documents and time series concurrently, such as predicting the movements of stock prices based on the contents of the news stories, is an emerging topic in data mining community. Previous researches have shown that there is a strong relationship between the time when the news stories are released and the time when the stock prices fluctuate. In this paper, we propose a systematic framework for predicting the tertiary movements of stock prices by analyzing the impacts of the news stories on the stocks. To be more specific, we investigate the immediate impacts of news stories on the stocks based on the Efficient Markets Hypothesis. Several data mining and text mining techniques are used in a novel way. Extensive experiments using real-life data are conducted, and encouraging results are obtained.
Identifying Interesting Customers through Web Log Classification.

%22Missing Is Useful%27 Missing Values in Cost%2DSensitive Decision Trees.
Abstract—Many real-world data sets for machine learning and data mining contain missing values and much previous research regards it as a problem and attempts to impute missing values before training and testing. In this paper, we study this issue in costsensitive learning that considers both test costs and misclassification costs. If some attributes (tests) are too expensive in obtaining their values, it would be more cost-effective to miss out their values, similar to skipping expensive and risky tests (missing values) in patient diagnosis (classification). That is, “missing is useful ” as missing values actually reduces the total cost of tests and misclassifications and, therefore, it is not meaningful to impute their values. We discuss and compare several strategies that utilize only known values and that “missing is useful ” for cost reduction in cost-sensitive decision tree learning. Index Terms—Induction, knowledge acquisition, machine learning. 1
Database classification for multi%2Ddatabase mining.

ENDL A Logical Framework for Verifying Secure Transaction Protocols.

A Verification Model for Electronic Transaction Protocols.

Decision trees with minimal costs.
We propose a simple, novel and yet effective method for building and testing decision trees that minimizes the sum of the misclassification and test costs. More specifically, we first put forward an original and simple splitting criterion for attribute selection in tree building. Our treebuilding algorithm has many desirable properties for a cost-sensitive learning system that must account for both types of costs. Then, assuming that the test cases may have a large number of missing values, we design several intelligent test strategies that can suggest ways of obtaining the missing values at a cost in order to minimize the total cost. We experimentally compare these strategies and C4.5, and demonstrate that our new algorithms significantly outperform C4.5 and its variations. In addition, our algorithm’s complexity is similar to that of C4.5, and is much lower than that of previous work. Our work is useful for many diagnostic tasks which must factor in the misclassification and test costs for obtaining missing information. 1.
Discovering Interesting Association Rules by Clustering.
Abstract. There are a great many metrics available for measuring the interestingness of rules. In this paper, we design a distinct approach for identifying association rules that maximizes the interestingness in an applied context. More specifically, the interestingness of association rules is defined as the dissimilarity between corresponding clusters. In addition, the interestingness assists in filtering out those rules that may be uninteresting in applications. Experiments show the effectiveness of our algorithm.
Cost%2DSensitive Decision Trees with Multiple Cost Scales.

Decision trees with minimal costs.
We propose a simple, novel and yet effective method for building and testing decision trees that minimizes the sum of the misclassification and test costs. More specifically, we first put forward an original and simple splitting criterion for attribute selection in tree building. Our treebuilding algorithm has many desirable properties for a cost-sensitive learning system that must account for both types of costs. Then, assuming that the test cases may have a large number of missing values, we design several intelligent test strategies that can suggest ways of obtaining the missing values at a cost in order to minimize the total cost. We experimentally compare these strategies and C4.5, and demonstrate that our new algorithms significantly outperform C4.5 and its variations. In addition, our algorithm’s complexity is similar to that of C4.5, and is much lower than that of previous work. Our work is useful for many diagnostic tasks which must factor in the misclassification and test costs for obtaining missing information. 1.
Mining Term Association Rules for Heuristic Query Construction.

Dealing with Inconsistent Secure Messages.

Is Minimum%2DSupport Appropriate to Identifying Large Itemsets%3F

Efficient mining of both positive and negative association rules.

Mining Dependent Patterns In Probabilistic Databases.

Guest Editors%27 Introduction Information Enhancement for Data Mining.

Identifying Global Exceptional Patterns in Multi%2Ddatabase Mining.
In multi-database mining, there can be many local patterns (frequent itemsets or association rules) in each database. At the end of multi-database mining, it is necessary to analyze these local patterns to gain global patterns, when putting all the data from the databases into a single dataset can destroy important information that reflect the distribution of global patterns. This paper develops an algorithm for synthesizing local patterns in multi-database is proposed. This approach is particularly fit to find potentially useful exceptions. The proposed method has been evaluated experimentally. The experimental results have shown that this method is efficient and appropriate to identifying exceptional patterns.
Guest Editors%27 Introduction Information Enhancement for Data Mining.

A fuzzy logic based method to acquire user threshold of minimum%2Dsupport for mining association rules.

An efficient strategy for mining exceptions in multi%2Ddatabases .

Identifying Software Component Association With Genetic Algorithm.

Verifying the Purchase Request in SET Protocol.

A Database%2DIndependent Approach of Mining Association Rules with Genetic Algorithm.

Verifying the Payment Authorization in SET Protocol.

Identifying Frequent Terms in Text Databases by Association Semantics.

A Database%2DIndependent Strategy for Confidence Determination.

Discovering Associations in Very Large Databases by Approximating.

Mining Dynamic databases by Weighting.

Identifying Approximate Itemsets of Interest in Large Databases.

Temporal Constraint Satisfaction in Matrix Method.
Constraint satisfaction is a challenging problem in Interval Algebra (IA). So there are many efforts to attack this problem. After building a matrix method to deal with temporal reasoning problems, we develop basic techniques for applying the matrix method to constraint satisfaction in this paper. Thus, the propagating rules and the algorithms of 3- and path-consistency are studied. If our matrix method is used, then the temporal constraint satisfaction problem can be transformed into a problem that can be effectively solved.
Data Preparation for Data Mining.
Data preparation is a fundamental stage of data analysis. While a lot of low-quality information is available in various data sources and on the Web, many organizations or companies are interested in how to transform the data into cleaned forms which can be used for high-profit purposes. This goal generates an urgent need for data analysis aimed at cleaning the raw data. In this paper, we first show the importance of data preparation in data analysis, then introduce some research achievements in the area of data preparation. Finally, we suggest some future directions of research and development.
An Agent%2DBased Hybrid Framework for Database Mining.
While knowledge discovery in databases (KDD) is defined as an iterative sequence of the following steps: data pre-processing, data mining, and post data mining, a significant amount of research in data mining has been done, resulting in a variety of algorithms and techniques for each step. However, a single data-mining technique has not been proven appropriate for every domain and data set. Instead, several techniques may need to be integrated into hybrid systems and used cooperatively during a particular data-mining operation. That is, hybrid solutions are crucial for the success of data mining. This paper presents a hybrid framework for identifying patterns from databases or multi-databases. The framework integrates these techniques for mining tasks from an agent point of view. Based on the experiments conducted, putting different KDD techniques together into the agent-based architecture enables them to be used cooperatively when needed. The proposed framework provides a highly flexible and robust data-mining platform and the resulting systems demonstrate emergent behaviors although it does not improve the performance of individual KDD techniques.
Cooperative Strategy for Web Data Mining and Cleaning.

Toward Databases Mining Pre%2DProcessing Collected Data.

Multi%2DDatabase Mining.
Multi-database mining is an important research area  because (1) there is an urgent need for analyzing data in different sources, (2) there are essential differences between mono- and multi-database mining, and (3) there are limitations in existing multi-database mining efforts. This paper designs a new multidatabase mining process. Some research issues involving mining multi-databases, including database clustering and local pattern analysis, are discussed.
Synthesizing High%2DFrequency Rules from Different Data Sources.

Post%2Dmining maintenance of association rules by weighting.

A Probability Data Model and its Semantics.

Association Rule Mining%2C Models and Algorithms

Identifying Exceptional Patterns in Multi%2DDatabases.

Mind the Trends When You Mine Incremental Data Mining.

Mining Both Positive and Negative Association Rules.
In this paper, we describe a novel technique, called APACS2, for mining interesting quantitative association rules from very large databases. To effectively mine such rules, APACS2 employs adjusted difference analysis. The use of this technique has the advantage that it does not require any user-supplied thresholds which are often hard to determine. Furthermore, APACS2 also has the advantage that it allows users to discover both positive and negative association rules. A positive association rule tells us that a record having certain attribute value will also have another attribute value whereas a negative association rule tells us that a record having certain attribute value will not have another attribute value. The fact that APACS2 is able to mine both positive and negative association rules and that it uses an objective yet meaningful measure to determine the interestingness of a rule makes it very effective at different data mining tasks.
Association Rule Mining by Agents.
mining System. The EMADS vision is that of a community of data mining agents, contributed by many individuals and interacting under decentralized control, to address data mining requests. EMADS is seen both as an end user platform and a research tool. This chapter details the EMADS vision, the associated conceptual framework and the current implementation. Although EMADS may be applied to many data mining tasks; the study described here, for the sake of brevity, concentrates on agent based Association Rule Mining and agent based classification. A full description of EMADS is presented.
Mining Very Large Databases Using Software Agents.

Propagating Temporal Relations of Intervals by Matrix.
Traditional temporal relations propagating is based on Allen ’ s Interval Algebra. This paper proposes an alternative method to propagate temporal relations among intervals, in which 5 £ 5 matrices are used to represent temporal relations of intervals. Hence, the propagation of temporal relations is transformed into a numerical computation. For ef ® ciency, we use the special values of the thirteen matrices to determine the possible temporal relations between two given intervals by using only the ® nal resultant matrix so as to optimize the propagation. To evaluate the utility of the proposed technique, we have implemented the matrix representation in Java. The experimental results demonstrate that the approach is ef ® cient and promising. In the real world, changes cannot be avoided. To conquer and exploit the real world for our life, we must catch the properties of time. So, Allen proposed a time world model of time interval (simply called interval) calculusÐ Interval Algebra (IA) (Allen 1983). The model divides the relationships among intervals into 13 kinds of temporal forms, laying a foundation for dealing with temporal relationships in applications. The major challenge in IA is the propagation of temporal relations. To confront this problem, traditional models are designed to improve IA. For example, a number of researches have been focused on constraint satisfactory
Discovering Causality in Large Databases.

Anytime mining for multiuser applications.

Encoding probability propagation in belief networks.

Collecting Quality Data for Database Mining.

Pattern Discovery in Probabilistic Databases.

A Logical Framework for Knowledge Sharing in Multi%2Dagent Systems.
Abstract. The issue of knowledge sharing has been an important topic in multi-agent research. Knowledge sharing leads to that agents analyze, judge and synthesize the told information so as to make agents’ own knowledge. To match these applications, this paper builds a logical framework for knowledge sharing among agents. We developa multimodal logic for reasoning about both agents ’ knowledge and told information. For formalizing the relationshipbetween knowledge and told information, we present a framework of semantics, with respect to which a sound and complete proof theory is given.
Mining Small Databases by Collecting Knowledge.

Estimating Itemsets of Interst by Sampling.

Large Scale Data Mining Based on Data Partitioning.

A model for compressing probabilities in belief networks.

Anytime algorithm for agent%2Dmediated merchant information gathering.

A nearest neighborhood algebra for probabilistic databases.

A Temporal Logic for Supporting Historical Databases.

Aggregation and maintenance for database mining.

IMC A Method for Interval Calculus in Matrix.
. Time representation is important in many applications, such as temporal databases, planning, and multi-agents. Since Allen&#039;s work on binary interval relatons (called interval algebra), many researchers have further investigated temporal informaton processing based on interval calculus. However, there are still some limitations such as constraint satisfaction is a NP-hard problem in interval calculus. For this reason, we propose a new interpretation for interval relationships and their calculus in this paper. That is to establish a new method to transform interval calculus into matrix calculus. Our experiments show that it is faster to propagate temporal relations than interval algebra. Keywords: Planning, temporal logic, interval algebra, interval calculus, temporal reasoning. 1 Introduction In 1983, Allen [1] proposed a time world model of interval algebra (IA). This model has been successfully applied to multi-agents [7], temporal databases [6], simulations [8], and concurrent sy...
