Autocorrelation and Relational Learning: Challenges and Opportunities

作者: Jennifer Neville , Ozgur Simsek , David Jensen

DOI: 10.21236/ADA472226

关键词: Machine learningMathematicsIndependence (mathematical logic)Artificial intelligenceInferenceModel learningVariable (computer science)Statistical relational learningAutocorrelation

摘要: Abstract : Autocorrelation, a common characteristic of many datasets, refers to correlation between values the same variable on related objects. It violates critical assumption instance independence that underlies most conventional models. In this paper, we provide an overview research autocorrelation in number fields with emphasis implications for relational learning, and outline challenges opportunities model learning inference.

参考文章(25)
Julian Besag, Statistical Analysis of Non-Lattice Data The Statistician. ,vol. 24, pp. 179- 195 ,(1975) , 10.2307/2987782
Robert F. Engle, ARCH: Selected Readings ,(1995)
Charles J. Geyer, Elizabeth A. Thompson, Constrained Monte Carlo Maximum Likelihood for Dependent Data Journal of the royal statistical society series b-methodological. ,vol. 54, pp. 657- 683 ,(1992) , 10.1111/J.2517-6161.1992.TB01443.X
Yiming Yang, Seán Slattery, Rayid Ghani, A Study of Approaches to Hypertext Categorization intelligent information systems. ,vol. 18, pp. 219- 241 ,(2002) , 10.1023/A:1013685612819
Sofus A. Macskassy, Foster Provost, A Simple Relational Classifier Defense Technical Information Center. ,(2003) , 10.21236/ADA452802
James P. LeSage, R. Kelley Pace, Spatial Dependence in Data Mining Springer, Boston, MA. pp. 439- 460 ,(2001) , 10.1007/978-1-4615-1733-7_24
David Jensen, Jennifer Neville, Linkage and Autocorrelation Cause Feature Selection Bias in Relational Learning international conference on machine learning. pp. 259- 266 ,(2002)
Jeffrey M. Wooldridge, Introductory Econometrics: A Modern Approach ,(1999)
PETER V. MARSDEN, NOAH E. FRIEDKIN, Network studies of social influence Sociological Methods & Research. ,vol. 22, pp. 127- 151 ,(1993) , 10.1177/0049124193022001006