Analogical Classification: Handling Numerical Data

Bounhas, Myriam; Prade, Henri; Richard, Gilles

doi:10.1007/978-3-319-11508-5_6

Cited by 8 publications

(11 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several authors have studied analogical proportions, i.e. statements of the form "salmon tartare is to grilled salmon what beef carpaccio is to grilled steak", and their use in classification in recent years [29,30,31,32]. Most of these approaches are restricted to binary or nominal attributes, although recently some promising results have been obtained for numerical attributes as well [32].…”

Section: Formalizing Commonsense Reasoningmentioning

confidence: 99%

“…statements of the form "salmon tartare is to grilled salmon what beef carpaccio is to grilled steak", and their use in classification in recent years [29,30,31,32]. Most of these approaches are restricted to binary or nominal attributes, although recently some promising results have been obtained for numerical attributes as well [32]. The use of analogical-proportion based reasoning in logic has been considered in the approach from [25], where the more general notion of extrapolative reasoning was studied.…”

Section: Formalizing Commonsense Reasoningmentioning

confidence: 99%

“…However, while promising, the results that were obtained are not competitive with standard methods such as k-NN and SVM. Recently, somewhat better results have been reported in [32], although only datasets with clearly defined and relatively few attributes have been considered (the largest considered dataset has 36 attributes).…”

Section: Classification Based On Relational Similaritymentioning

confidence: 99%

See 2 more Smart Citations

JKR adhesive contact for a transversely isotropic layer of finite thickness

Argatov

Borodich

Popov

2015

J. Phys. D: Appl. Phys.

View full text Add to dashboard Cite

Commonsense reasoning patterns such as interpolation and a fortiori inference have proven useful for dealing with gaps in structured knowledge bases. An important di culty in applying these reasoning patterns in practice is that they rely on fine-grained knowledge of how di↵erent concepts and entities are semantically related. In this paper, we show how the required semantic relations can be learned from a large collection of text documents. To this end, we first induce a conceptual space from the text documents, using multi-dimensional scaling. We then rely on the key insight that the required semantic relations correspond to qualitative spatial relations in this conceptual space. Among others, in an entirely unsupervised way, we identify salient directions in the conceptual space which correspond to interpretable relative properties such as 'more fruity than' (in a space of wines), resulting in a symbolic and interpretable representation of the conceptual space. To evaluate the quality of our semantic relations, we show how they can be exploited by a number of commonsense reasoning based classifiers. We experimentally show that these classifiers can outperform standard approaches, while being able to provide intuitive explanations of classification decisions. A number of crowdsourcing experiments provide further insights into the nature of the extracted semantic relations.

show abstract

Section: Formalizing Commonsense Reasoningmentioning

confidence: 99%

Section: Formalizing Commonsense Reasoningmentioning

confidence: 99%

Section: Classification Based On Relational Similaritymentioning

confidence: 99%

See 1 more Smart Citation

JKR adhesive contact for a transversely isotropic layer of finite thickness

Argatov

Borodich

Popov

2015

J. Phys. D: Appl. Phys.

View full text Add to dashboard Cite

show abstract

“…

O d d_{2} (N N, S t d)

has close classification results to those of analogy‐based classifier in the numerical case for most datasets. In the Boolean case, both

o d d n e s s

‐based and

a n a l o g y

‐based classifiers achieve good results for “Balance,” “Car,” “Monk1,” and “Monk3.” For “Monk2” data set, Analogy‐based classifier significantly outperforms

O d d_{2} (N N, S t d)

while for “Spect” and “Voting” the converse is observed.…”

Section: Experimentations and Preliminary Discussionmentioning

confidence: 81%

Oddness-based classification: A new way of exploiting neighbors

Bounhas

Prade

Richard

2018

Int. J. Intell. Syst.

Self Cite

View full text Add to dashboard Cite

The classification of a new item may be viewed as a matter of associating it with the class where it is the least at odds w.r.t. the elements already in the class. An oddness measure of an item with respect to a multiset, applicable to Boolean features as well as to numerical ones, has been recently proposed. It has been shown that cumulating this measure over pairs or triples (rather than larger subsets) of elements in a class could provide an accurate estimate of the global oddness of an item with respect to a class. This idea is confirmed and refined in the present paper. Rather than considering all the pairs in a class, one can only deal with the pairs whose an element is one of the nearest neighbors of the item, in the target class. The oddness evaluation computed on this basis still leads to good results in terms of accuracy. One can take a step further and choose the second element in the pair also as another nearest neighbor in the class. Although the method relies on the notion of neighbors, the resulting algorithm is far from being a variant of the classical k-nearest neighbors approach. The oddness with respect to a class computed only on the basis of pairs made of two nearest neighbors leads to a low complexity algorithm. Experiments on a set of UCI benchmarks show that the classifier obtained can compete with other well-known approaches.

show abstract

“…Then one looks for triples (a, b, c) of items with a known class, for which the class equation cl(a) : cl(b) :: cl(c) : cl(x) is solvable, and for which analogical proportions hold with x on the attributes describing the items. It has been first successively applied to Boolean attributes [4,21] and then extended to nominal and to numerical ones [5]. Recent formal studies have shown that analogical classifiers always give exact predictions in the special cases where the classification process is governed by an affine Boolean function (which includes x-or functions) and only in this case, which does not prevent to get good results in other cases (as observed in practice), but which is still to be better understood [7,16].…”

Section: Applicationsmentioning

confidence: 99%

Analogical Proportions and Analogical Reasoning - An Introduction

Prade

Richard

2017

Case-Based Reasoning Research and Development

Self Cite

View full text Add to dashboard Cite

Analogical proportions are statements of the form "a is to b as c is to d". For more than a decade now, their formalization and use have raised the interest of a number of researchers. In this talk we shall primarily focus on their modeling in logical settings, both in the Boolean and in the multiple-valued cases. This logical view makes clear that analogy is as much a matter of dissimilarity as a matter of similarity. Moreover analogical proportions emerge as being especially remarkable in the framework of logical proportions. The analogical proportion and seven other code independent logical proportions can be shown as being of particular interest. Besides, analogical proportions are at the basis of an inference mechanism which enables us to complete or create a fourth item from three other items. The relation with case-based reasoning and case-based decision is emphasized. Potential applications and current developments are also discussed.

show abstract

Analogical Classification: Handling Numerical Data

Cited by 8 publications

References 10 publications

JKR adhesive contact for a transversely isotropic layer of finite thickness

JKR adhesive contact for a transversely isotropic layer of finite thickness

Oddness-based classification: A new way of exploiting neighbors

Analogical Proportions and Analogical Reasoning - An Introduction

Contact Info

Product

Resources

About