This article presents a long-term inter-disciplinary research project situated at the intersection of the scientific disciplines of Musicology and Artificial Intelligence. The goal is to develop AI, and in particular machine learning and data mining, methods to study the complex phenomenon of expressive music performance. Formulating formal, quantitative models of expressive performance is one of the big open research problems in contemporary (empirical and cognitive) musicology. Our project develops a new direction in this field: we use inductive learning techniques to discover general and valid expression principles from (large amounts of) real performance data. The project is currently starting its third year and is planned to continue for at least four more years. In the

In this paper, we introduce hyperlink ensembles, a novel type of ensemble classifier for classifying hypertext documents. Instead of using the text on a page for deriving features that can be used for training a classifier, we suggest to use portions of texts from all pages that point to the target page. A hyperlink ensemble is formed by obtaining one prediction for each hyperlink that points to a page.

The paper presents a series of noise detection experiments in a medical problem of coronary artery disease diagnosis. The following algorithms for noise detection and elimination are tested: a saturation filter, a classification filter, a combined classification-saturation filter, and a consensus saturation filter. The distinguishing feature of the novel consensus saturation filter is its high reliability which is due to the multiple detection of potentially noisy examples. Reliable detection of noisy examples is important for the analysis of patient records in medical databases, as well as for the induction of rules from filtered data, representing genuine characteristics of the diagnostic domain. Medical evaluation in the problem of coronary artery disease diagnosis shows that the detected noisy examples are indeed noisy or non-typical class representatives.

This article presents a new rule discovery algorithm named PLCG that can find simple, robust partial rule models (sets of classification rules) in complex data where it is difficult or impossible to find models that completely account for all the phenomena of interest. Technically speaking,

This paper describes an extension of a GAbased, separate-and-conquer propositional rule induction algorithm called SIA [24]. While the original algorithm is computationally attractive and is also able to handle both nominal and continuous attributes efficiently, our algorithm further improves it by taking into account of the recent advances in the rule induction and evolutionary computation communities. The refined system has been compared to other GA-based and non GA-based rule learning algorithms on a number of benchmark datasets from the UCI machine learning repository. Results show that the proposed system can achieve higher performance while still produces a smaller number of rules. 1 Introduction The increasingly widespread use of information system technologies and the internet has resulted in an explosive growth of many business, government and scientific databases. As these terabyte-size databases become prevalent, the traditional approach of using human experts to sift thro...

Text Categorization (TC) and Information Extraction (IE) are two important goals of Natural Language Processing. While handcrafting rules for both tasks has a long tradition, learning approaches gained much interest in the past. In the present paper we try to provide a solid basis for the application of ILP methods to these learning problems. We propose to introduce three basic types (namely a type for text, one for words and one for text positions) and three simple predicate definitions over these types which enable to write text categorization and information extraction rules as logic programs. Based on the proposed representation, we present the key concepts of our approach to the problem of learning rules for TC and IE in terms of ILP. We conclude the paper by comparing our approach of representing texts and rules as logic programs to others.

. Divide-and-Conquer (DAC) and Separate-and-Conquer (SAC) are two strategies for rule induction that have been used extensively. When searching for rules DAC is maximally conservative w.r.t. decisions made during search for previous rules. This results in a very efficient strategy, which however suffers from difficulties in effectively inducing disjunctive concepts due to the replication problem. SAC on the other hand is maximally liberal in the same respect. This allows for a larger hypothesis space to be searched, which in many cases avoids the replication problem but at the cost of lower efficiency. We present a hybrid strategy called Reconsider-and-Conquer (RAC), which handles the replication problem more effectively than DAC by reconsidering some of the earlier decisions and allows for more efficient induction than SAC by holding on to some of the decisions. We present experimental results from propositional, numerical and relational domains demonstrating that RAC si...

The primary goal of the research reported in this paper is to identify what criteria are responsible for the good performance of a heuristic rule evaluation function in a greedy topdown covering algorithm. We first argue that search heuristics for inductive rule learning algorithms typically trade off consistency and coverage, and we investigate this trade-off by determining optimal parameter settings for five different parametrized heuristics. In order to avoid biasing our study by known functional families, we also investigate the potential of using meta-learning for obtaining alternative rule learning heuristics. The key results of this experimental study are not only practical default values for commonly used heuristics and a broad comparative evaluation of known and novel rule learning heuristics, but we also gain theoretical insights into factors that are responsible for a good performance. For example, we observe that consistency should be weighed more heavily than coverage, presumably because a lack of coverage can later be corrected by learning additional rules.

Abstract not found

A novel native stochastic local search algorithm for solving k-term DNF problems is presented. It is evaluated on hard k-term DNF problems that lie on the phase transition and compared to the performance of GSAT and WalkSAT type algorithms on SAT encodings of k-term DNF problems. We also evaluate state-of-the-art separate and conquer algorithms on these problems. Finally, we demonstrate the practical relevance of our algorithm on a chess endgame database.