Why collective inference improves relational classification

Jensen, David; Neville, Jennifer; Gallagher, Brian

doi:10.1145/1014052.1014125

Cited by 207 publications

(203 citation statements)

References 8 publications

Supporting

Mentioning

197

Contrasting

Unclassified

Order By: Relevance

“…The relational learning addresses the relational representation (first order logic) that relational learning has seen many applications such as relational classification [21], link based clustering of web search [22], and link prediction in relational data [23].…”

Section: Statistical Relational Learningmentioning

confidence: 99%

Learning and prediction of relational time series

Tan

Darken

2015

Comput Math Organ Theory

View full text Add to dashboard Cite

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instruction, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to 12a. DISTRIBUTION / AVAILABILITY STATEMENTApproved for public release; distribution is unlimited 12b. DISTRIBUTION CODE ABSTRACT (maximum 200 words)Prediction of events is fundamental to both human and artificial agents. The main problem with previous prediction techniques is that they cannot predict events that have never been experienced before. This dissertation addresses the problem of predicting such novelty by developing algorithms and computational models inspired from recent cognitive science theories: conceptual blending theory and event segmentation theory. We were able to show that prediction accuracy for event or state prediction can be significantly improved using these methods.The main contribution of this dissertation is a new class of prediction techniques inspired by conceptual blending that improves prediction accuracy overall and has the ability to predict even events that have never been experienced before. We also show that event segmentation theory, when integrated with these techniques, results in greater computational efficiency. We implemented the new prediction techniques, and more traditional alternatives such as Markov and Bayesian techniques, and compared their prediction accuracy quantitatively for three domains: a roleplaying game, intrusion-system alerts, and event prediction of maritime paths in a discrete-event simulator. Other contributions include two new unification algorithms that improve over a naïve one, and an exploration of ways to maintain a minimum-size knowledge base without affecting prediction accuracy. SUBJECT TERMS ABSTRACTPrediction of events is fundamental to both human and artificial agents. The main problem with previous prediction techniques is that they cannot predict events that have never been experienced before. This dissertation addresses the problem of predicting such novelty by developing algorithms and computational models inspired from recent cognitive science theories: conceptual blending theory and event segmentation theory.We were able to show that prediction accuracy for event or state prediction can be significantly improved using these methods.The main contribution of this dissertation is a new class of prediction techniques inspired by conceptual blending that improves prediction accuracy overall and has the ability to predict even events that have never been experienced before. We also show that event segmentation theory, when integrated with these techniques, results in greater computational efficiency. We implemented the new prediction techniques, and more traditional alternatives such as Markov and Bayesian t...

show abstract

Section: Statistical Relational Learningmentioning

confidence: 99%

Learning and prediction of relational time series

Tan

Darken

2015

Comput Math Organ Theory

View full text Add to dashboard Cite

show abstract

“…Within-network classification, for which training entities are connected directly to entities whose labels are to be classified, stays in contrast to across-network classification, where models learnt from one network are applied to another similar network [4]. Across-network classification can be understood as a transfer learning approach that accomplishes relational classification [5].…”

Section: Relational Influence Propagationmentioning

confidence: 99%

MapReduce approach to relational influence propagation in complex networks

Kajdanowicz

Indyk

Kazienko

2012

Pattern Anal Applic

View full text Add to dashboard Cite

The relational label propagation problem for large data sets using MapReduce programming model was considered in the paper. The method we propose estimates class probability in relational domain in the networks. The method was examined on large real telecommunication data set. The results indicated that it could be used successfully to classify networks' nodes and, thanks to that, new offerings or tariffs might be proposed to customers who belong to other providers. Moreover, basic properties of relational label propagation were examined and reported.

show abstract

“…Increasing complexity and structure of data on the Web revealed limitations of the traditional attribute based (content) classi cation based solely on own content of data objects. In search for advanced methods capable to exploit structure of interconnected data instances more intensively, single relational classi cation [7,5] originated as more e cient alternative to content classi cation.…”

Section: Introductionmentioning

confidence: 99%

Moderated Class–Membership Interchange in Iterative Multi–relational Graph Classifier

Vojtek

Bieliková

2010

Advances in Soft Computing

View full text Add to dashboard Cite

Organizing information resources into classes helps signi cantly in searching in massive volumes of on line documents available through the Web or other information sources such as electronic mail, digital libraries, corporate databases. Existing classi cation methods are often based only on own content of document, i.e. its attributes. Considering relations in the web document space brings better results. We adopt multi relational classi cation that interconnects attribute based classi ers with iterative optimization based on relational heterogeneous graph structures, while di erent types of instances and various relation types can be classi ed together. We establish moderated class membership spreading mechanism in multi relational graphs and compare the impact of various levels of regulation in collective inference classi er. The experiments based on large scale graphs originated in MAPEKUS research project data set (web portals of scienti c libraries) demonstrate that moderated class membership spreading signi cantly increases accuracy of the relational classi er (up to 10%) and protects instances with heterophilic neighborhood to be misclassi ed.

show abstract

Why collective inference improves relational classification

Cited by 207 publications

References 8 publications

Learning and prediction of relational time series

Learning and prediction of relational time series

MapReduce approach to relational influence propagation in complex networks

Moderated Class–Membership Interchange in Iterative Multi–relational Graph Classifier

Contact Info

Product

Resources

About