Results 1  10
of
36
Naive Bayesian Classification of Structured Data
, 2003
"... In this paper we present 1BC and 1BC2, two systems that perform naive Bayesian classification of structured individuals. The approach of 1BC is to project the individuals along firstorder features. These features are built from the individual using structural predicates referring to related objects ..."
Abstract

Cited by 32 (0 self)
 Add to MetaCart
In this paper we present 1BC and 1BC2, two systems that perform naive Bayesian classification of structured individuals. The approach of 1BC is to project the individuals along firstorder features. These features are built from the individual using structural predicates referring to related objects (e.g. atoms within molecules), and properties applying to the individual or one or several of its related objects (e.g. a bond between two atoms). We describe an individual in terms of elementary features consisting of zero or more structural predicates and one property; these features are treated as conditionally independent in the spirit of the naive Bayes assumption. 1BC2 represents an alternative firstorder upgrade to the naive Bayesian classifier by considering probability distributions over structured objects (e.g., a molecule as a set of atoms), and estimating those distributions from the probabilities of its elements (which are assumed to be independent). We present a unifying view on both systems in which 1BC works in language space, and 1BC2 works in individual space. We also present a new, efficient recursive algorithm improving upon the original propositionalisation approach of 1BC. Both systems have been implemented in the context of the firstorder descriptive learner Tertius, and we investigate the differences between the two systems both in computational terms and on artificially generated data. Finally, we describe a range of experiments on ILP benchmark data sets demonstrating the viability of our approach.
MrSBC: a MultiRelational Naive Bayes Classifier
 Todorovski & H. Blockeel (Eds.), Knowledge Discovery in Databases PKDD 2003, Lecture Notes in Artificial Intelligence
, 2003
"... Abstract. In this paper we propose an extension of the naïve Bayes classification method to the multirelational setting. In this setting, training data are stored in several tables related by foreign key constraints and each example is represented by a set of related tuples rather than a single row ..."
Abstract

Cited by 15 (7 self)
 Add to MetaCart
(Show Context)
Abstract. In this paper we propose an extension of the naïve Bayes classification method to the multirelational setting. In this setting, training data are stored in several tables related by foreign key constraints and each example is represented by a set of related tuples rather than a single row as in the classical data mining setting. This work is characterized by three aspects. First, an integrated approach in the computation of the posterior probabilities for each class that make use of first order classification rules. Second, the applicability to both discrete and continuous attributes by means a supervised discretization. Third, the consideration of knowledge on the data model embedded in the database schema during the generation of classification rules. The proposed method has been implemented in the new system MrSBC, which is tightly integrated with a relational DBMS. Testing has been performed on two datasets and four benchmark tasks. Results on predictive accuracy and efficiency are in favour of MrSBC for the most complex tasks. 1
The Role of Feature Construction in Inductive Rule Learning
"... This paper proposes a unifying framework for inductive rule learning algorithms. We suggest that the problem of constructing an appropriate inductive hypothesis (set of rules) can be broken down in the following subtasks: rule construction, body construction, and feature construction. Each of the ..."
Abstract

Cited by 13 (0 self)
 Add to MetaCart
This paper proposes a unifying framework for inductive rule learning algorithms. We suggest that the problem of constructing an appropriate inductive hypothesis (set of rules) can be broken down in the following subtasks: rule construction, body construction, and feature construction. Each of these subtasks may have its own declarative bias, search strategies, and heuristics. In particular, we argue that feature construction is a crucial notion in explaining the relations between attributevalue rule learning and inductive logic programming (ILP). We demonstrate this by a general method for transforming ILP problems to attributevalue form, which overcomes some of the traditional limitations of propositionalisation approaches.
Schemas and Models
 IN PROCEEDINGS OF THE SIGKDD2002 WORKSHOP ON MULTIRELATIONAL LEARNING
, 2002
"... We propose the SchemaModel Framework, which characterizes algorithms that learn probabilistic models from relational data as having two parts: a schema that identifies sets of related data items and groups them into relevant categories; and a model that allows probabilistic inference about those ..."
Abstract

Cited by 11 (1 self)
 Add to MetaCart
We propose the SchemaModel Framework, which characterizes algorithms that learn probabilistic models from relational data as having two parts: a schema that identifies sets of related data items and groups them into relevant categories; and a model that allows probabilistic inference about those data items. The framework
On adaptability in grid systems
 In Future Generation Grids, CoreGRID series
, 2005
"... With the increasing size and complexity, adaptability is among the most badly needed properties in today’s Grid systems. Adaptability refers to the degree to which adjustments in practices, processes, or structures of systems are possible to projected or actual changes of their environment. In this ..."
Abstract

Cited by 11 (1 self)
 Add to MetaCart
(Show Context)
With the increasing size and complexity, adaptability is among the most badly needed properties in today’s Grid systems. Adaptability refers to the degree to which adjustments in practices, processes, or structures of systems are possible to projected or actual changes of their environment. In this paper, we review concepts, methods, algorithms, and implementations that are deemed useful for designing adaptable Grid systems, illustrating them with examples. Contrary to the existing literature, the portfolio of the proposed approaches includes unorthodox tools such as game theory. We also discusses methods which have not been fully exploited for purposes of adaptability, such as automated planning or time series analysis. Our inventory is done along the stages of the feedback loopknown fromcontrol theory. These stages includemonitoring, analyzing, predicting, planning, decision taking, and finally executing the plan. Our discussion reveals that several of the problems paving the way to fully adaptable system are of fundamental nature, which makes a ’quantum leap’ progress in this area unlikely.
Bridging the gap between distance and generalisation: Symbolic learning in metric spaces
, 2008
"... Distancebased and generalisationbased methods are two families of artificial intelligence techniques that have been successfully used over a wide range of realworld problems. In the first case, general algorithms can be applied to any data representation by just changing the distance. The metric ..."
Abstract

Cited by 10 (4 self)
 Add to MetaCart
Distancebased and generalisationbased methods are two families of artificial intelligence techniques that have been successfully used over a wide range of realworld problems. In the first case, general algorithms can be applied to any data representation by just changing the distance. The metric space sets the search and learning space, which is generally instanceoriented. In the second case, models can be obtained for a given pattern language, which can be comprehensible. The generalityordered space sets the search and learning space, which is generally modeloriented. However, the concepts of distance and generalisation clash in many different ways, especially when knowledge representation is complex (e.g. structured data). This work establishes a framework where these two fields can be integrated in a consistent way. We introduce the concept of distancebased generalisation, which connects all the generalised examples in such a way that all of them are reachable inside the generalisation by using straight paths in the metric space. This makes the metric space and the generalityordered space coherent (or even dual). Additionally, we also introduce a definition of minimal distancebased generalisation that can be seen as the first formulation of the Minimum Description Length (MDL)/Minimum Message Length (MML) principle in terms of a distance function. We instantiate and develop the framework for the most common data representations and distances, where we show that consistent instances can be found for numerical data, nominal data, sets, lists, tuples, graphs, firstorder atoms and clauses. As a result, general learning methods that integrate the best from distancebased and generalisationbased methods can be defined and adapted to any specific problem by appropriately choosing the distance, the pattern language and the generalisation operator.
Logical Characterisations Of Inductive Learning
, 2000
"... This chapter presents a logical analysis of induction. Contrary to common approaches to inductive logic that treat inductive validity as a realvalued generalisation of deductive validity, we argue that the only logical step in induction lies in hypothesis generation rather than evaluation. ..."
Abstract

Cited by 8 (2 self)
 Add to MetaCart
This chapter presents a logical analysis of induction. Contrary to common approaches to inductive logic that treat inductive validity as a realvalued generalisation of deductive validity, we argue that the only logical step in induction lies in hypothesis generation rather than evaluation.
Clustergrouping: From subgroup discovery to clustering
 In Proceedings of the 15th European Conference on Machine Learning (ECML
, 2004
"... Abstract. We introduce the problem of clustergrouping and show that it can be considered a subtask in several important data mining tasks, such as subgroup discovery, mining correlated patterns, clustering and classification. The algorithm CG for solving clustergrouping problems is then introduce ..."
Abstract

Cited by 8 (0 self)
 Add to MetaCart
Abstract. We introduce the problem of clustergrouping and show that it can be considered a subtask in several important data mining tasks, such as subgroup discovery, mining correlated patterns, clustering and classification. The algorithm CG for solving clustergrouping problems is then introduced, and it is incorporated as a component in several existing and novel algorithms for tackling subgroup discovery, clustering and classification. The resulting systems are empirically compared to stateoftheart systems such as CN2, CBA, Ripper, Autoclass and CobWeb. The results indicate that the CG algorithm can be useful as a generic local pattern mining component in a wide variety of data mining and machine learning algorithms.
Discovering and explaining abnormal nodes in semantic graphs
 IEEE Transactions on Knowledge and Data Engineering
"... Abstract—An important problem in the area of homeland security is to identify abnormal or suspicious entities in large data sets. Although there are methods from data mining and social network analysis focusing on finding patterns or central nodes from networks or numerical data sets, there has been ..."
Abstract

Cited by 6 (1 self)
 Add to MetaCart
(Show Context)
Abstract—An important problem in the area of homeland security is to identify abnormal or suspicious entities in large data sets. Although there are methods from data mining and social network analysis focusing on finding patterns or central nodes from networks or numerical data sets, there has been little work aimed at discovering abnormal instances in large complex semantic graphs, whose nodes are richly connected with many different types of links. In this paper, we describe a novel unsupervised framework to identify such instances. Besides discovering abnormal instances, we believe that to complete the process, a system has to also provide users with understandable explanations for its findings. Therefore, in the second part of the paper, we describe an explanation mechanism to automatically generate humanunderstandable explanations for the discovered results. To evaluate our discovery and explanation systems, we perform experiments on several different semantic graphs. The results show that our discovery system outperforms stateoftheart unsupervised network algorithms used to analyze the 9/11 terrorist network and other graphbased outlier detection algorithms by a significant margin. Additionally, the human study we conducted demonstrates that our explanation system, which provides natural language explanations for the system’s findings, allowed human subjects to perform complex data analysis in a much more efficient and accurate manner. Index Terms—Anomaly detection, data mining, knowledge and data engineering tools and techniques, semantic graphs. Ç 1
Learning in Clausal Logic: A Perspective on Inductive Logic Programming
 Computational Logic: Logic Programming and Beyond, volume 2407 of Lecture Notes in Computer Science
, 2002
"... Abstract. Inductive logic programming is a form of machine learning from examples which employs the representation formalism of clausal logic. One of the earliest inductive logic programming systems was Ehud Shapiro’s Model Inference System [90], which could synthesise simple recursive programs like ..."
Abstract

Cited by 5 (0 self)
 Add to MetaCart
(Show Context)
Abstract. Inductive logic programming is a form of machine learning from examples which employs the representation formalism of clausal logic. One of the earliest inductive logic programming systems was Ehud Shapiro’s Model Inference System [90], which could synthesise simple recursive programs like append/3. Many of the techniques devised by Shapiro, such as topdown search of program clauses by refinement operators, the use of intensional background knowledge, and the capability of inducing recursive clauses, are still in use today. On the other hand, significant advances have been made regarding dealing with noisy data, efficient heuristic and stochastic search methods, the use of logical representations going beyond definite clauses, and restricting the search space by means of declarative bias. The latter is a general term denoting any form of restrictions on the syntactic form of possible hypotheses. These include the use of types, input/output mode declarations, and clause schemata. Recently, some researchers have started using alternatives to Prolog featuring strong typing and real functions, which alleviate the need for some of the above adhoc mechanisms. Others have gone beyond Prolog by investigating learning tasks in which the hypotheses are not definite clause programs, but for instance sets of indefinite clauses or denials, constraint logic programs, or clauses representing association rules. The chapter gives an accessible introduction to the above topics. In addition, it outlines the main current research directions which have been strongly influenced by recent developments in data mining and challenging reallife applications. 1