Download AI 2011: Advances in Artificial Intelligence: 24th by Mani Abedini, Michael Kirley (auth.), Dianhui Wang, Mark PDF

By Mani Abedini, Michael Kirley (auth.), Dianhui Wang, Mark Reynolds (eds.)

This publication constitutes the refereed complaints of the twenty fourth Australasian Joint convention on synthetic Intelligence, AI 2011, held in Perth, Australia, in December 2011. The eighty two revised complete papers offered have been rigorously reviewed and chosen from 193 submissions. The papers are equipped in topical sections on information mining and data discovery, computer studying, evolutionary computation and optimization, clever agent structures, good judgment and reasoning, imaginative and prescient and snap shots, snapshot processing, usual language processing, cognitive modeling and simulation know-how, and AI applications.

Show description

Read Online or Download AI 2011: Advances in Artificial Intelligence: 24th Australasian Joint Conference, Perth, Australia, December 5-8, 2011. Proceedings PDF

Similar nonfiction_5 books

Reaction and resistance: feminism, law, and social change

During this well timed quantity, members from a variety of disciplines examine response and resistance to feminism in different parts of legislation and coverage – baby custody, baby poverty, sexual harassment, and sexual attack – and in a couple of institutional websites, resembling courts, legislatures, households, the mainstream media, and the academy.

Nanotechnology: Content and Context

This can be an NSF-funded (Grant No. EEC-0407237) Rice collage path referred to as "Nanotechnology: Context and content material. " The target is to coach scholars a few uncomplicated nanoscience/nanotechnology by means of placing it in a social and cultural context. scholars are anticipated to profit either a few easy technological know-how and expertise and while, a few innovations for figuring out the social and cultural importance, function, and attainable results of this rising technological know-how.

Additional resources for AI 2011: Advances in Artificial Intelligence: 24th Australasian Joint Conference, Perth, Australia, December 5-8, 2011. Proceedings

Example text

2 Results There was no general statistical difference across the datasets (in a win/loss sense) between the ordinal (S. Ord) and non-ordinal discretized (Nom) methods. However, on four of the datasets the ordinal method (S. Ord) significantly outperformed the unordered method (Nom), and on no datasets was it Table 1. Discretized Numeric Order Discovery : Classifier Accuracy Data Set Nom UD S. Ord R. Ord S. O+N R. 22 Win/Draw/Loss vs Nominal 10/0/8 9/0/9 4/0/14 10/0/8 7/2/9 Win/Draw/Loss vs Random Ordinal 17/0/1 14/0/4 16/0/2 16/0/2 15/0/3 ◦ significant improvement against Nominal • significant degradation against Nominal 34 A.

These results add value to investigating if it is possible to recover or discover orders, since those orders can be of significant benefit in classification. 3 Discovering Orders We developed three related methods for finding attribute orders on typical nominal attributes. To narrow the scope and focus on improving classification we made a fundamental assumption that the order of an attribute would be related to the class value, which is reasonable because if an attribute is present in a dataset it is likely there because of an initial suspicion that it is related to the solution.

Semi-Supervised Classification Using Tree-Based SOMs 5 27 Results Comparison to Other Classifiers: The results of the performance of the different classifiers (columns) across all the dataset (rows) is summarized in Table 2. Specifically, we are interested in the performance of our classifier on problems across a diversity of domains in which labeled and unlabeled data is available2 . 33% the correct label of the instances belonging the wine dataset. 98%. One possibility for quantifying the quality of our method is to consider the family of classifiers inheriting the VQ mechanism.

Download PDF sample

Rated 4.26 of 5 – based on 22 votes