The metric and dimensional assumptions that underlie the geometric. In such cases, similarity will be reduced more by the distinctive features of the subject than by the distinctive features of the predicate. Semantic web 0 0 1 1 ios press tverskys featurebased similarity and beyond silvia likaveca, ilaria lombardia. Similarity, confusability, and the density hypothesis. The tversky index, named after amos tversky, is an asymmetric similarity measure on sets that. In a fingerprint the presence or absence of a structural fragment is represented by the presence or absence of a set bit. Amos tversky 19371996, a towering figure in cognitive and mathematical psychology, devoted his professional life to the study of similarity, judgment, and decision making. In this work we wanted to gain some insights into tverskys and more specifically jaccards featurebased semantic similarity measure on instances in a specific ontology. For, first, the set of features that represents an object will in general be only a subset of the set of all the properties the object has. Tversky has related degree of judged similarity between among objects to the degree of overlap of matching features. This will in turn influence the judged similarity between objects that share these diagnostic features. Our proposed algorithm yields the best result in terms of precisionrecall, fmeasure and dcg compared to the traditional tversky similarity measure and geometry similarity algorithm. Because this gives equal weight to all parts of the encoding thence to different substructures in the molecule it may not be. Using tverskys contrast model to investigate how features of.
While i have an interest in behavioural economics, i have no expertise in psychology. Data were collected from 150 participants who performed an image description and a similarity judgment task. Nevertheless, two general properties, irregularity and local representativeness, seem to capture the intuitive notion of randomness. Studies of similarity amos tversky itamar gati hebrew university, jerusalem any event in the history of the organism is, in a sense, unique. Computes the tversky index between two sequences the tversky index is an asymmetric similarity measure between two sets, one defined the prototype and the other the variant. Tversky 1977, which provided compelling demonstrations against all the. He had a unique ability to master the technicalities of normative ideals and then to intuit and demonstrate experimentally their systematic violation due to the vagaries and consequences of human information processing. The contrast model can account for violations of minimality and asymmetry, as well as certain other phenomena associated with proximity data, such as systematic differences between ratings of similarity and of dissimilarity and changes in similarity with context tversky, 1977. Relationships between perceived features and similarity of. Doctor of philosophy information science, may 2005, 125 pp. It serves as an organizing principle by which individuals classify objects, form concepts, and make generalizations.
The metric and dimensional assumptions that underlie the geometric represen. We develop a similarity measure, based on fuzzy logic, that exhibits several features that match experimental findings in humans. The model is dubbed fuzzy feature contrast ffc and is an extension to a more general domain of the feature contrast model due to tversky 1977. In this work we wanted to gain some insights into tverskys featurebased semantic similarity measure on instances in a specific ontology. Features of similarity and categorybased induction evan heit department of psychology university of warwick coventry cv4 7al united kingdom e. Additive clustering relies on the common features measure for item similarities tversky, 1977. Tversky s 1977 contrast model, unlike geometricspatial models of similarity, does not represent stimuli as points in a multidimensional space. Tverskys contrast model 1977 systematizes this feature approach.
The observed similarity of object a to object b, s a, b, is a function of their common features, those that are shared by both a and b, and their distinctive features, those that belong to one. Tverskys 1977 contrast model, which defines similarity as a feature contrast task and equates the degree of similarity of two stimuli to a linear combination of their common and distinctive features, explains human similarity judgments much better than the geometric models. Corter teachers college, columbia university amos tversky stanford university proximity data can be represented by an extended tree, which generalizes traditional trees by including marked segments that correspond to overlapping dusters. The representativeness heuristic suggests that similarity judgments provide a basis for judgments of likelihood.
Features of similarity amos tversky hebrew university jerusalem, israel the metric and dimensional assumptions that underlie the geometric representation of similarity are questioned on both theoretical and empirical grounds. The triangle inequality differs from minimality and symmetry in that it cannot be formulated in ordinal terms. A quantum geometric model of similarity psychological sciences. Similarity the percepts and concepts laboratory at indiana. A central assumption of the model is that the similarity of object a to object b is a function of the features common to a and b a and b, those in a but not in b symbolized ab and those in b but not in a ba. The tversky index, named after amos tversky, is an asymmetric similarity measure on sets that compares a variant to a prototype. The tversky index can be seen as a generalization of dices coefficient and tanimoto coefficient aka jaccard index.
A unique feature of the quantum similarity model is that, whereas previous models. To avoid a combinatorial explosion, the valueofclairvoyance computation employs the assumption that at most one additional feature will be reported by the user. According to tverskys diagnosticity principle, features that are used to cluster stimuli into subgroups have a higher diagnostic value and will therefore affect similarity judgments more than features that are not used to create clusters. Lehmann graduate school of business, columbia uniciersity, new york, ny10027 usa. For alpha beta 1, the index is equal to the tanimoto. A new settheoretical approach to similarity is developed in which objects are.
An experimental note on tverskys features of similarity. Rorissa, abebe, perceived features and similarity of images. The contrast model expresses similarity between objects as a weighted. As is true of the similarity of sample to population, the speci. This cited by count includes citations to the following articles in scholar.
Tversky index is a generalization of the dice similarity coe cient and the f scores. Ec features that determine apparent randomness differ depending on context. When we compare something we judge features, not a simple physical distance of resemblance on a mental mapchart. Similarity is one of the most straightforward ways to relate two objects and guide the human perception of the world. Then, the user observes one or more of these features. Selected writings amos tversky edited by eldar shafir amos tversky 19371996, a towering figure in cognitive and mathematical psychology, devoted his profes sional life to the study of similarity, judgment, and decision making. It is shown here that the extent to which mentally retarded children judge planometric figures to be alike is related positively to the number of matching cues. Our method was deliberately designed to find a point of contact between the two traditions of research in heuristics the early tradition exemplified by kahneman and tverskys work, and the later tradition exemplified by the work of gigerenzer and goldstein 1996. Johnson school of business admintitration, university of michigan, ann arbor, mi 48109, usa donald r. As will be shown below, this approach to the judgment of likelihood leads to serious biases, because several of the factors that should be considered in assessing likelihood play no role in judgments of similarity. Indeed, the concept of similarity is ubiquitous in psychological theory. I have read some of his work in other collections particularly judgement under uncertainty and it. Features of similarity amos tversky similarity plays a fundamental role in theories of knowledge and behavior.
In addition to being present or absent, a feature can also be missing, in the sense that the value of that feature is not known. Using tverskys contrast model to investigate how features. Alternatively, one may estimate probability by assessing availability, or associative distance. Semantic web 0 0 1 1 ios press tverskys featurebased. A similarity comparison involves comparing the feature lists that represent the concepts. A new model is constructed that is based on the alcove model kruschke, 1992, but it uses a feature matching similarity measure see, e. It is possible to account for peoples intuitions or ratings of the similarities. Previous studies compared the molecular similarity of marketed drugs and endogenous human metabolites endogenites, using a series of fingerprinttype encodings, variously ranked and clustered using the tanimoto jaccard similarity coefficient ts. It asserts that one distance must be smaller than the sum of two others, and hence it cannot be readily refuted with ordinal or even interval data. Tverskys featurebased similarity and beyond semantic web journal. Hebrew university jerusalem, israel the metric and dimensional assumptions that underlie the geometric represen tation of similarity are questioned on both theoretical and empirical grounds. Nevertheless, the works of amos tversky are immediately accessible to the outsider. This theory provided an explanation for a number of judgmental anomalies that had been observed by other researchers.
Similarity in categorization 381 as additive features, so there is no loss of generality if we restrict attention to the latter. Features of similarity 329 sessment of similarity are briefly discussed in the final section. Computational models of word similarity two types of models for predicting human word similarity ratings dominate the literature. Features that are shared in the feature lists are commonalities of the pair and features that are contained in one feature set but not the other are differences of the pair. We show how adjusting the hyperparameters of this index allow placing emphasis on false negatives in training a network that generalizes and performs well in highly imbalanced data as it leads to high sensitivity, dice, f. The ones marked may be different from the article in the profile. Tverskys approach can also easily deal with the contextdependence of similarity judgments that goodman brought to bear against the geometrical model. A heuristic for judging frequency and probability122.
91 1488 591 1372 762 67 124 547 894 1390 1405 1042 592 411 602 778 96 1134 1637 153 73 1396 933 181 896 926 1008 632 1275 106 1081 1439 1258 1203 1467 827 1116 347 748 939 321