Research and Development in Intelligent Systems XVIII: by Derek Sleeman (auth.), Max Bramer BSc, PhD, CEng, Frans

By Derek Sleeman (auth.), Max Bramer BSc, PhD, CEng, Frans Coenen PhD, Alun Preece BSc, PhD (eds.)

M.A. BRAMER collage of Portsmouth, united kingdom This quantity includes the refereed technical papers offered at ES200 l, the Twenty-fIrst SGES overseas convention on wisdom dependent platforms and utilized synthetic Intelligence, held in Cambridge in December 2 hundred l, including an invited keynote paper through Professor Derek Sleeman. The convention was once organised via SGES, the British machine Society professional team on wisdom established structures and utilized man made Intelligence. The papers during this quantity current new and cutting edge advancements within the box, divided into sections on desktop studying, Constraint delight, brokers, wisdom illustration, wisdom Engineering, and clever structures. The refereed papers commence with a paper entitled 'Detecting Mismatches between specialists' Ontologies received via wisdom Elicitation', which describes a scientific method of the research of discrepancies inside and between specialists' ontologies. This paper was once judged to be the simplest refereed technical paper submitted to the convention. the rest papers are dedicated to subject matters in vital components equivalent to brokers, wisdom engineering, wisdom illustration, making plans and constraint delight, with laptop studying back the biggest subject coated by way of the variety of papers approved for e-book. this can be the eighteenth quantity within the study and improvement sequence. the appliance circulate papers are released as a spouse quantity lower than the identify functions and options in clever structures IX.

Show description

Read Online or Download Research and Development in Intelligent Systems XVIII: Proceedings of ES2001, the Twenty-first SGES International Conference on Knowledge Based Systems and Applied Artificial Intelligence, Cambridge, December 2001 PDF

Best research books

Advanced Genealogy Research Techniques

A qualified genealogist finds tips to get prior brick partitions in examine and unearth hard-to-find ancestors

Advanced family tree study concepts makes use of updated and hugely prepared equipment and methods to teach you the way to discover the elusive info to around out your family tree examine. You'll get prior the brick partitions that experience stumped you and notice the right way to circulation past easy varieties of family tree assets. The publication covers a number of software program courses and really good family tree instruments and exhibits you the way to create an internet genealogical examine log to maintain information stumbled on and arrange it in how one can assist you comprehend what you've exposed. approximately each kind of sleek social networking is addresses as is utilizing DNA documents. This useful, in-depth consultant offers the subsequent point of aspect for someone who desires to extend past the newbie strategies and techniques.

• makes use of confirmed study how to move deeper and discover elusive details
• permits you to comprehend the main points you discover and retain music of data
• Covers numerous software program courses and really expert family tree tools
• bargains a number of situations and examples to force domestic the study equipment defined

Research and Development in Intelligent Systems XVIII: Proceedings of ES2001, the Twenty-first SGES International Conference on Knowledge Based Systems and Applied Artificial Intelligence, Cambridge, December 2001

M. A. BRAMER collage of Portsmouth, united kingdom This quantity contains the refereed technical papers provided at ES200 l, the Twenty-fIrst SGES foreign convention on wisdom established platforms and utilized synthetic Intelligence, held in Cambridge in December 2 hundred l, including an invited keynote paper by means of Professor Derek Sleeman.

Additional resources for Research and Development in Intelligent Systems XVIII: Proceedings of ES2001, the Twenty-first SGES International Conference on Knowledge Based Systems and Applied Artificial Intelligence, Cambridge, December 2001

Example text

Since the value for (3) is lower than for (I), J-pruning takes place and branch (3) is truncated. Table 5 shows the results obtained using J-pruning with a variety of datasets and the comparative figures for unpruned rules. 4 Table 5. g. 0 for wake_vortex). This again confirms that the basic (unpruned) form ofTDIDT leads to substantial overfitting of rules to the instances in the training set. The predictive accuracy is higher with J-pruning for 10 of the datasets, lower for 6 and unchanged for one (the smallest dataset, lens24).

Research and Development in Intelligent Systems XVIII © Springer-Verlag London Limited 2002 26 A II but two of the datasets used (wake ~vortex and wake~vortex2) were either created by the author or downloaded from the on-line repository of machine learning datasets maintained at the University of California at Irvine [4]. 2. 1 Example and Basic Terminology The following example is taken from [2]. Table I records a golf player's decisions on whether or not to play each day based on that day's weather conditions.

Though usually associated with nearest-neighbour or instance-based algorithms [22], a lazy approach to learning is increasingly being used in algorithms for decision-tree learning [10,11,23,24,25]. As in the eager version of PURSUE, attribute selection in the lazy version is based on the single strategy of increasing the probability of the target outcome class, with evidential power as the measure of attribute usefulness. However, instead of partitioning the data set into subsets corresponding to the values of a selected attribute, PURSUE now asks the user for the value of the selected attribute in the target problem and creates a single subset of the data set consisting of all instances for which the attribute's value is the same as the reported value.

Download PDF sample

Rated 4.77 of 5 – based on 38 votes