Frequent Episode Mining to Support Pattern Analysis in Developmental Biology.

Learning predictive models that use pattern discovery %2D A bootstrap evaluative approach applied in organ functioning sequences.

Advances in Intelligent Data Analysis VIII

Compressing tags to find interesting media groups.
On photo sharing websites like Flickr and Zooomr, users are offered the possibility to assign tags to their uploaded pictures. Using these tags to find interesting groups of semantically related pictures in the result set of a given query is a problem with obvious applications. We analyse this problem from a Minimum Description Length (MDL) perspective and develop an algorithm that finds the most interesting groups. The method is based on Krimp, which finds small sets of patterns that characterise the data using compression. These patterns are sets of tags, often assigned together to photos. The better a database compresses, the more structure it contains and thus the more homogeneous it is. Following this observation we devise a compression-based measure. Our experiments on Flickr data show that the most interesting and homogeneous groups are found. We show extensive examples and compare to clusterings on the Flickr website.
Identifying the Components.
To understand a component and how it is used in a system, an engineer must understand the externally visible properties and relationships with other components. We applied a technique, Interface Identification, to a product&#039;s source code to identify all of the interfaces between a component and the system. This permitted us to quickly harvest a component from one product for reuse in a new one. The technique relies on source code and knowledge of what source files make up a component. A reverse engineering tool aids the process.
Mining Databases to Mine Queries Faster.

Characteristic relational patterns.

Low%2DEntropy Set Selection.

Identifying the components.
To understand a component and how it is used in a system, an engineer must understand the externally visible properties and relationships with other components. We applied a technique, Interface Identification, to a product&#039;s source code to identify all of the interfaces between a component and the system. This permitted us to quickly harvest a component from one product for reuse in a new one. The technique relies on source code and knowledge of what source files make up a component. A reverse engineering tool aids the process.
StreamKrimp Detecting Change in Data Streams.

Filling in the Blanks %2D Krimp Minimisation for Missing Data.

Parallel Universes and Local Patterns%2C 01.05. %2D 04.05.2007

Discovering Relational Items Sets Efficiently.
Frequent item set mining is a major data mining research area. Generalising from the standard single table case to a multirelational setting is simple in principle, but hard in practice. That is, it is simple to define frequent item sets in the multirelational setting, as well as extending the A-Priori algorithm. It is hard, because the well-known frequent pattern explosion at low min-sup settings is far worse than it is in the standard case. In this paper we introduce an effective algorithm for the discovery of frequent, multi-relational item sets. These relational patterns show which item sets occur together. Answering questions like: ‘What type of Books are bought together with what Record types?’. Hence, they provide a symmetric insight in the relation and reveal patterns that are relevant with respect to the relation. It extends our earlier work on using MDL to discover a small set of characteristic item sets. The algorithm, R-KRIMP, first discovers the small set of characteristic patterns in the single tables and then combines these to find a small set of characteristic multi-relational item sets. This reduces the original search space dramatically and, hence, brings down the computational complexity by orders of magnitude. In the experiments we show that this approach yields a very good approximation of the naive approach, joining all tables into one huge table, while being far more efficient. Keywords: Frequent Patterns, Relational Mining, MDL. 1.
Parallel Universes and Local Patterns%2C 01.05. %2D 04.05.2007

Understanding Discrete Classifiers with a Case Study in Gene Prediction.

Preserving Privacy through Data Generation.

Characterising the difference.
Characterising the differences between two databases is an often occurring problem in Data Mining. Detection of change over time is a prime example, comparing databases from two branches is another one. The key problem is to discover the patterns that describe the difference. Emerging patterns provide only a partial answer to this question. In previous work, we showed that the data distribution can be captured in a pattern-based model using compression [12]. Here, we extend this approach to define a generic dissimilarity measure on databases. Moreover, we show that this approach can identify those patterns that characterise the differences between two distributions. Experimental results show that our method provides a wellfounded way to independently measure database dissimilarity that allows for thorough inspection of the actual differences. This illustrates the use of our approach in real world data mining.
Finding Composite Episodes.

07181 Abstracts Collection %2D%2D Parallel Universes and Local Patterns .

07181 Introduction %2D%2D Parallel Universes and Local Patterns .

Reducing the Frequent Pattern Set.

Compression Picks Item Sets That Matter.

Item Sets that Compress.
ABSTRACT: Association rule mining has been the main technique to identify meaningful patterns in market basket data (also known as synthetic transactional data). However, this approach has several limitations. Especially the large number of rules that have been generated from association creates difficulties in extracting and analyzing useful information. So, clustering association rules is introduced to overcome the problem. In this paper, we discuss three combined approaches to cluster association rules. First, we create hierarchical concepts and relationships of domain ontologies to merge and simplify items into more general concepts. This significantly reduces the total number of rules that are created. Adding to this, subspace clustering further reduces the resulting association rules to a more relevant problem space according to the level of details. Finally, a new ranking function is introduced to give more emphasis to the meaningful clustered groups via frequency, recency, and monetary value, which are the most important factors in the marketing domain. Our experiment shows that the total number of rules has decreased while the value of lift has increased. Also, the result of hierarchical subspace clustering shows that our approach groups data effectively to more simple, understandable and relevant problem space that is more suitable for further analysis.
Item Sets that Compress.
ABSTRACT: Association rule mining has been the main technique to identify meaningful patterns in market basket data (also known as synthetic transactional data). However, this approach has several limitations. Especially the large number of rules that have been generated from association creates difficulties in extracting and analyzing useful information. So, clustering association rules is introduced to overcome the problem. In this paper, we discuss three combined approaches to cluster association rules. First, we create hierarchical concepts and relationships of domain ontologies to merge and simplify items into more general concepts. This significantly reduces the total number of rules that are created. Adding to this, subspace clustering further reduces the resulting association rules to a more relevant problem space according to the level of details. Finally, a new ranking function is introduced to give more emphasis to the meaningful clustered groups via frequency, recency, and monetary value, which are the most important factors in the marketing domain. Our experiment shows that the total number of rules has decreased while the value of lift has increased. Also, the result of hierarchical subspace clustering shows that our approach groups data effectively to more simple, understandable and relevant problem space that is more suitable for further analysis.
Combination of text%2Dmining algorithms increases the performance.

Combination of text%2Dmining algorithms increases the performance.

Introduction.
The use of monads to structure functional programs is described. Monads provide a convenient framework for simulating e ects found in other languages, such as global state, exception handling, output, or non-determinism. Three case studies are looked at in detail: how monads ease the modi cation of a simple evaluator; how monads act as the basis of a datatype of arrays subject to in-place update; and how monads can be used to build parsers.  
KDID 2004
Abstract. We are designing new data mining techniques on boolean contexts to identify a priori interesting bi-sets (i.e., sets of objects or transactions associated to sets of attributes or items). A typical important case concerns formal concept mining (i.e., maximal rectangles of true values or associated closed sets by means of the so-called Galois connection). It has been applied with some success to, e.g., gene expression data analysis where objects denote biological situations and attributes denote gene expression properties. However in such real-life application domains, it turns out that the Galois association is a too strong one when considering intrinsically noisy data. It is clear that strong associations that would however accept a bounded number of exceptions would be extremely useful. We study the new pattern domain of α/β concepts, i.e., consistent maximal bi-sets with less than α false values per row and less than β false values per column. We provide a complete algorithm that computes all the α/β concepts based on the generation of concept unions pruned thanks to anti-monotonic constraints. An experimental validation on synthetic data is given. It illustrates that more relevant associations can be discovered in noisy data. We also discuss a practical application in molecular biology that illustrates an incomplete but quite useful extraction when all the concepts that are needed beforehand can not be discovered. 1
Local Pattern Detection
 Separate-and-conquer or covering rule learning algorithms may be viewed as a technique for using local pattern discovery for generating a global theory. Local patterns are learned one at a time, and each pattern is evaluated in a local context, with respect to the number of positive and negative examples that it covers. Global context is provided by removing the examples that are covered by previous patterns before learning a new rule. In this paper, we discuss several research issues that arise in this context. We start with a brief discussion of covering algorithms, their problems, and review a few suggestions for resolving them. We then discuss the suitability of a well-known family of evaluation metrics, and analyze how they trade off coverage and precision of a rule. Our conclusion is that in many applications, coverage is only needed for establishing statistical significance, and precision is the metric that should be optimized for rules. The main problem with optimizing precision is its unreliability for low example sizes, which is mainly caused by overfitting. We then report some preliminary experiments that addresses this problem by meta-learning a predictor for the true accuracy of a rule based on its coverage on the training set.  
Advances in Intelligent Data Analysis VI
The growing number of publicly available information sources makes it impossible for individuals to keep track of all the various opinions on one topic. The goal of our artificial believer system presented in this paper is to extract and analyze statements of opinion from newspaper articles. Beliefs are modeled using a fuzzy-theoretic approach applied after nlp-based information extraction. A fuzzy believer models a human agent, deciding what statements to believe or reject based on different, configurable strategies.
CONAN An Integrative System for Biomedical Literature Mining.
Abstract. The amount of information about the genome, transcriptome and proteome, forms a problem for the scientific community: how to find the right information in a reasonable amount of time. Most research aiming to solve this problem, however, concentrate on a certain organism or a very limited dataset. Complementary to those algorithms, we developed CONAN, a system which provides a full-scale approach, tailored to experimentalists, designed to combine several information extraction methods and connect the outcome of these methods to gather novel information. Its methods include tagging of gene/protein names, finding interaction and mutation data, tagging of biological concepts, linking to MeSH and Gene Ontology terms, which can all be found back by querying the system. We present a full-scale approach that will ultimately cover all of PubMed/MEDLINE. We show that this universality has no effect on quality: our system performs as well as existing systems. 1
Instability of Classifiers on Categorical Data.

Data Mining in Inductive Databases.
The objective of this thesis is to study data mining query optimization in a Logicbased Knowledge Discovery Support Environment. i.e, a flexible knowledge discovery system with capabilities to obtain, maintain, represent, and utilize both induced and deduced knowledge. In particular, we focus on frequent pattern queries, since this kind of query is at the basis of many mining tasks, and it seems appropriate to be encapsulated in a DBMS as a primitive operation. We introduce an inductive language for frequent pattern queries, which is simple enough to be highly optimized and expressive enough to cover the most of interesting queries. Then we study optimization issues, defining novel algorithms for constrained frequent pattern mining i.e., finding all itemsets included in a transaction database that satisfy a given set of constraints. Finally, all the contributions in both the linguistic part of the thesis and the algorithmic part, are amalgamated together in order to define an optimized constraint-pushing operational semantics for the proposed language. 4
Discovery of Regulatory Connections in Microarray Data.

Constructing %28Almost%29 Phylogenetic Trees from Developmental Sequences Data.

Involving Aggregate Functions in Multi%2Drelational Search.
Abstract The fact that data is scattered over many tables causes many problems in the practice of data mining. To deal with this problem, one either constructs a single table by propositionalisation, or uses a Multi-Relational Data Mining algorithm. In either case, one has to deal with the non-determinacy of one-to-many relationships. In propositionalisation, aggregate functions have already proven to be powerful tools to handle this non-determinacy. In this paper we show how aggregate functions can be incorporated in the dynamic construction of patterns of Multi-Relational Data Mining. 1
Complex Data Mining Using Patterns.

Principles of Data Mining and Knowledge Discovery
 We consider the problem of finding statistically unusual subgroups  in a multi-relation database, and extend previous work on singlerelation  subgroup discovery. We give a precise definition of the multirelation  subgroup discovery task, propose a specific form of declarative  bias based on foreign links as a means of specifying the hypothesis space,  and show how propositional evaluation functions can be adapted to the  multi-relation setting. We then describe an algorithm for this problem  setting that uses optimistic estimate and minimal support pruning, an  optimal refinement operator and sampling to ensure efficiency and can  easily be parallelized.  
MAMBO Discovering Association Rules Based on Conditional Independencies.
We present the Mambo algorithm for the discovery of association rules. Mambo is driven by conditional independence relations between the variables instead of the minimum support restrictions of algorithms like Apriori. We argue that making use of conditional independencies is an intuitively appealing way to restrict the set of association rules considered. Since we only have a finite sample from the probability distribution of interest, we have to deal with uncertainty concerning the conditional independencies present. Bayesian methods are used to quantify this uncertainty, and the posterior probabilities of conditional independence relations are estimated with the Markov Chain Monte Carlo technique. We analyse an insurance data set with Mambo and illustrate the differences in results compared to Apriori.
Propositionalisation and Aggregates.
Abstract The fact that data is scattered over many tables causes many problems in the practice of data mining. To deal with this problem, one either constructs a single table by hand, or one uses a Multi-Relational Data Mining algorithm. In this paper, we propose a different approach in which the single table is constructed automatically using aggregate functions, which repeatedly summarise information from different tables over associations in the datamodel. Following the construction of the single table, we apply traditional data mining algorithms. Next to an in-depth discussion of our approach, the paper presents results of experiments on three well-known data sets. 1
Propositionalisation and Aggregates.
Abstract The fact that data is scattered over many tables causes many problems in the practice of data mining. To deal with this problem, one either constructs a single table by hand, or one uses a Multi-Relational Data Mining algorithm. In this paper, we propose a different approach in which the single table is constructed automatically using aggregate functions, which repeatedly summarise information from different tables over associations in the datamodel. Following the construction of the single table, we apply traditional data mining algorithms. Next to an in-depth discussion of our approach, the paper presents results of experiments on three well-known data sets. 1


Multi%2DRelational Data Mining%2C Using UML for ILP.
Held in conjunction with the 12th European Conference on Machine Learning (ECML&#039;01) and the 5th European
BioInformatics Databases   Data Mining %28abstract%29.

Data Mining Methods for Knowledge Discovery
Abstract: Mining knowledge from large amounts of spatial data is known as spatial data mining. It becomes a highly demanding field because huge amounts of spatial data have been collected in various applications ranging from geo-spatial data to bio-medical knowledge. The amount of spatial data being collected is increasing exponentially. So, it far exceeded human’s ability to analyze. Recently, clustering has been recognized as a primary data mining method for knowledge discovery in spatial database. The database can be clustered in many ways depending on the clustering algorithm employed, parameter settings used, and other factors. Multiple clustering can be combined so that the final partitioning of data provides better clustering. In this paper, a novel density based k-means clustering algorithm has been proposed to overcome the drawbacks of DBSCAN and kmeans clustering algorithms. The result will be an improved version of k-means clustering algorithm. This algorithm will perform better than DBSCAN while handling clusters of circularly distributed data points and slightly overlapped clusters. Keywords: Clustering, DBSCAN, k-means, DBkmeans. I.
Priors on network structures. Biasing the search for Bayesian networks.
In this paper we show how a user can influence recovery of Bayesian Networks from a database by specifying prior knowledge. The main novelty of our approach is that the user only has to provide partial prior knowledge, which is then completed to a full prior over all possible network structures. This partial prior knowledge is expressed among variables in an intuitive pairwise way, which embodies the uncertainty of the user about his/her own prior knowledge. Thus, the uncertainty of the model is updated in the normal Bayesian way. 1991 Mathematics Subject Classification: 62A15, 46N30, 47N30, 60Bxx, 90B40, 93A30, 62-07  1991 Computing Reviews Classification System: H.3.3, I.2.8, I.6.5, G.3  Keywords and Phrases: bayesian network, prior probabilities, user knowledge.  Note: This work has been carried out under the Esprit-IV project Keso. 1. Introduction  Bayesian nets provide much insight in the conditional (in)dependencies among the attributes in a database. As such, Bayesian network re...
Measuring Time Series%27 Similarity through Large Singular Features Revealed with Wavelet Transformation.

The Haar Wavelet Transform in the Time Series Similarity Paradigm.
Abstract. Similarity measures play an important role in many data mining algorithms. To allow the use of such algorithms on non-standard databases, such as databases of financial time series, their similarity measure has to be defined. We present a simple and powerful technique which allows for the rapid evaluation of similarity between time series in large data bases. It is based on the orthonormal decomposition of the time series into the Haar basis. We demonstrate that this approach is capable of providing estimates of the local slope of the time series in the sequence of multi-resolution steps. The Haar representation and a number of related represenations derived from it are suitable for direct comparison, e.g. evaluation of the correlation product. We demonstrate that the distance between such representations closely corresponds to the subjective feeling of similarity between the time series. In order to test the validity of subjective criteria, we test the records of currency exchanges, finding convincing levels of correlation. 1
Multi%2Drelational Decision Tree Induction.
Discovering decision trees is an important set of techniques in KDD, both because of their simple interpretation and the efficiency of their discovery. One of their disadvantages is that they do not take the structure of the mining object into account. By going from the standard single-relation approach to the multi-relational approach as in ILP this disadvantage is removed. However, the straightforward generalization loses the efficiency of the standard algorithms. In this paper we present a framework that allows the efficient discovery of multi-relational decision trees through the exploitation of the domain knowledge encoded in the data model of the database.  Introduction  The induction of decision trees has been getting a lot of attention in the field of Knowledge Discovery in Databases over the past few years. This popularity has been largely due to the efficiency with which decision trees can be induced from large datasets, as well as to the elegant and intuitive representation ...
Wavelet Transform in Similarity Paradigm.
and their applications. SMC is sponsored by the Netherlands Organization for Scientific Research (NWO). CWI is a member of
Research and Business Challenges in Data Mining Technology.

Enriching Active Databases with Agent Technology.

Query Optimization to Support Data Mining.

KESO Minimizing Database Interaction.

DEGAS A Database of Autonomous Objects.

DEGAS Capturing Dynamics in Objects.

Object Histories as a Foundation for an Active OODB.

Deductive Databases Challenges

Data Mining What it is and how it is done.

Data Mining and the KESO Project.

Data Surveying Foundations of an Inductive Query Language.

Guiding Schema Integration by Behavioural Information.

An Approach to Schema Integratioin Based on Transformations and Behaviour

Homogeneous Discoveries Contain No Surprises Inferring Risk Profiles from Large Databases.
Many models of reality are probabilistic. For example, not everyone orders crisps with their beer, but a certain percentage does. Inferring such probabilistic knowledge from databases is one of the major challenges for data mining. Recently Agrawal et al. [1] investigated a class of such problems. In this paper a new class of such problems is investigated, viz., inferring risk-profiles. The proto-typical example of this class is: &#034;what is the probability that a given policy-holder will file a claim with the insurance company in the next year&#034;. A risk-profile is then a description of a group of insurants that have the same probability for filing a claim. It is shown in this paper that homogeneous descriptions are the most plausible risk-profiles. Moreover, under modest assumptions it is shown that covers of such homogeneous descriptions are essentially unique. A direct consequence of this result is that it suffices to search for the homogeneous description with the highest ass...
Schema Integration in Object%2DOriented Databases
This report presents a formal approach to support schema integration in objectoriented  databases. The basis of the approach is a synthetic subclass order, which is  defined in terms of a weak subtype relation on underlying types of classes and a subfunction  relation on functional forms of methods. Classes are identified using an equivalence  relation induced by the subclass order and factorised using a join operator w.r.t. the  subclass order. Class hierarchies are integrated by repeatedly identifying and factorising  classes. The novelty of this report is that both attributes and methods are used to compare  classes and behaviour of methods is used to compare attributes.  1991 CR Categories: D.1.5: [Software] Object-oriented programming, D.2.2: [Software]  Tools and techniques, H.2.1: [Database management] Logical design.  Keywords and Phrases: Database design, Object-oriented databases, Schema integration.  Note: This research is partly funded by the Dutch Organisation for Scientific Research  through NFI-grant NF74.  Contents  1 
Termination and Confluence of Rule Execution.
Rules provide the functionality for constraint enforcement and view maintenance. A provably  correct implementation of both issues based on rules, requires confluent and terminating behaviour of  the rule set. However, little work has been done so far on the static detection of these properties. In  this paper, a design theory for rule sets in an OODBMS is developed.  This paper introduces two predicates, viz., Terminate(n) and Independent, which imply respectively  termination and confluency. Both predicates are characterised under two kinds of rule execution semantics:  set and instance based. The decidability of the predicates is proven and it is shown that set  and instance based semantics coincide whenever the rule set is independent and terminates. Moreover,  sufficient conditions of low algorithmic complexity for both Terminate(n) and Independent under both  kinds of semantics are given.  1991 CR Categories: H.2.1[Information Systems]: Logical Design- data models; H.2.3[Informat...
Enforcing Confluence of Rule Execution.

Towards a design theory for Database triggers.
Advances in the area of active database management systems require the development  of a trigger design theory, which guides the user in the definition off wellbehaving  trigger based applications. The development of such a theory requires a  formal definition of trigger semantics. This paper describes a framework for such a  formalisation of triggers. That is, the parameters of trigger execution and the options  for setting them are identified and discussed. Furthermore, the development of  a trigger design theory is initiated with the formulation of a sufficient condition for  trigger independence.  
Using Design Axioms and Topology to Model Database Semantics.

