Abstract
This paper presents a novel approach for lazy classification based on the notion of analogical proportions. Our starting point is a method from the literature based on a measure of analogical dissimilarity. Based on some observations about the effectiveness of different analogical proportion situations for classification purposes, we optimize this method, considerably reducing the size of the training set. These results raise some questions about the reasons of the effectiveness of the analogical approach, which are briefly discussed at the end of the paper.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Correa Beltran, W., Jaudoin, H., Pivert, O.: Analogical prediction of null values: The numerical attribute case. In: Proc. of the 18th East-European Conference on Advances in Databases and Information Systems (ADBIS 2014) (2014)
Correa Beltran, W., Jaudoin, H., Pivert, O.: Estimating null values in relational databases using analogical proportions. In: Laurent, A., Strauss, O., Bouchon-Meunier, B., Yager, R.R. (eds.) IPMU 2014, Part III. CCIS, vol. 444, pp. 110–119. Springer, Heidelberg (2014)
Miclet, L., Prade, H.: Handling analogical proportions in classical logic and fuzzy logics settings. In: Sossai, C., Chemello, G. (eds.) ECSQARU 2009. LNCS, vol. 5590, pp. 638–650. Springer, Heidelberg (2009)
Prade, H., Richard, G., Yao, B.: Enforcing regularity by means of analogy-related proportions — a new approach to classification. IJCISIM 4, 648–658 (2012)
Prade, H., Richard, G.: Reasoning with logical proportions. In: Lin, F., Sattler, U., Truszczynski, M. (eds.) KR. AAAI Press (2010)
Prade, H., Richard, G.: Analogical proportions and multiple-valued logics. In: van der Gaag, L.C. (ed.) ECSQARU 2013. LNCS, vol. 7958, pp. 497–509. Springer, Heidelberg (2013)
Bayoudh, S., Miclet, L., Delhay, A.: Learning by analogy: A classification rule for binary and nominal data. In: IJCAI, pp. 678–683 (2007)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this paper
Cite this paper
Correa Beltran, W., Jaudoin, H., Pivert, O. (2014). Lazy Analogical Classification: Optimization and Precision Issues. In: Straccia, U., Calì, A. (eds) Scalable Uncertainty Management. SUM 2014. Lecture Notes in Computer Science(), vol 8720. Springer, Cham. https://doi.org/10.1007/978-3-319-11508-5_7
Download citation
DOI: https://doi.org/10.1007/978-3-319-11508-5_7
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-11507-8
Online ISBN: 978-3-319-11508-5
eBook Packages: Computer ScienceComputer Science (R0)