=Paper=
{{Paper
|id=Vol-1510/paper10
|storemode=property
|title=Property-based Semantic Similarity: What Counts?
|pdfUrl=https://ceur-ws.org/Vol-1510/paper10.pdf
|volume=Vol-1510
|dblpUrl=https://dblp.org/rec/conf/aic/LikavecC15
}}
==Property-based Semantic Similarity: What Counts?==
https://ceur-ws.org/Vol-1510/paper10.pdf
Property-based Semantic Similarity: What Counts?
Silvia Likavec and Federica Cena
Università di Torino, Dipartimento di Informatica, Torino, Italy
{likavec,cena}@di.unito.it
Abstract. Similarity, one of six Gestalt principles, is one of the most intuitive
ways to perceive the world and categorise the objects surrounding us. The notion
of similarity plays an important role in many areas, and it is important to simu-
late human perception of similarity in order to obtain satisfying results in various
applications. We draw our inspiration from Tversky’s work on similarity and de-
fine property-based similarity for ontological concepts taking into account their
common and distinctive features and their values. We also discuss some possible
ways to improve the property-based similarity.
Keywords: Similarity, properties, ontology
1 Introduction
It is inherent to human nature to try to categorize objects surrounding us, finding pat-
terns and forms they have in common. One of the most intuitive ways to relate two
objects is through their similarity. Similarity is one of the six Gestalt principles which
guide the human perception of the world, the remaining ones being: Proximity, Closure,
Good Continuation, Common Fate, and Good Form.
According to Merriam Webster “similarity” is a quality that makes one person or
thing like another and “similar” means having characteristics in common. There are
many ways in which objects can be perceived as similar, such as having similar color,
shape, size, texture etc. But if we move away from just visual stimuli, we can apply
the same principles to define the semantic similarity of two objects. This leads to a
similarity based on features these two objects have in common, and consequently, the
lack of distinctive features characterising each object.
The concept of semantic similarity can be encountered in various fields, from Nat-
ural Language Processing (NLP) and Information Retrieval to Semantic Web. In this
work we deal with the semantic similarity of concepts in domain ontologies (Gruber,
1993, Guarino and Poli, 1995), where concepts are distinguished by the properties as-
sociated to them. The usage of ontologies to represent various domains accounts for
both similarities and differences among domain objects as well as generic objects and
very specific ones.
Our inspiration comes from Tversky’s work on Features of Similarity (Tversky,
1977) and we try to apply his ideas to similarity among ontological objects. More pre-
cisely, two objects are similar if they both are defined having the same properties with
the same values. In addition to this simple notion of similarity, we explore how this
similarity can be improved by considering relevance for properties or relevance for val-
ues or hierarchical relationships among values, Throughout this work we use domain
of recipes to provide examples and explain our approach and reflections.
The rest of the paper is organised as follows. In Section 2, we provide a brief back-
ground on ontologies for knowledge representation and on the treatment of properties
in OWL. We give the details of how to calculate the property-based similarity for in-
stances in the domain ontology in Section 3 and then we look into some possible ways
to improve the property-based similarity in the ontology in Section 4. We summarise the
most relevant related work which regards the semantic similarity in Section 5. Finally,
we conclude in Section 6.
2 OWL ontologies and knowledge representation
In various fields, from e-commerce and e-learning to cultural heritage, medicine, digital
libraries etc., it is possible to describe the concepts of the domain by using the properties
of these concepts and their respective values. The ones that immediately come to mind
are ontologies (Antoniou and van Harmelen, 2008, Allemang and Hendler, 2008) and
linked open data (Bizer et al., 2009), where properties are prominent elements of the
domain and contribute to the description of domain concepts.
In this work we deal with ontologies, powerful and expressive formalisms which
make it possible to explicitly specify domain elements and their properties, as well as re-
lationships which exist among domain elements. Also, rigorous reasoning mechanisms
are associated with ontologies. One standard formalism for representing ontologies is
OWL.1
Throughout this work we would use domain of recipes to provide examples and
explain our approach and reflections.
2.1 Properties in OWL
In ontologies expressed in OWL properties are used to describe domain elements and
express their features. There are two kinds of properties in OWL:
(i) object properties describing relations among individuals and
(ii) data type properties providing relations among individuals and data type values.
Object properties and datatype properties are defined as instances of the built-in OWL
classes owl:ObjectProperty and owl:DatatypeProperty, respectively. Both are subclasses
of the RDF class rdf:Property. Here, we only consider object properties, and leave the
treatment of data type properties (such as literal values) for future analysis, since it is
more complex.
The property axiom is used to define the characteristics of a property. Usually, it de-
fines its domain and range. rdfs:domain links a property to a class description, whereas
rdfs:range links a property to either a class description or a data range. For example:
1
http://www.w3.org/TR/owl-ref
defines a property has ingredient which connects the elements of Recipe class to
the elements of Food class.
Equivalent properties are defined with owl:equivalentProperty.
Properties can be explicitly defined for the classes and can be used to define classes
with property restrictions. Our approach to similarity is best illustrated when consider-
ing instances in the ontology, hence we will provide here a brief description of proper-
ties for instances.
2.2 Instances and their properties
An instance in the ontology are characterised by its class membership, individual iden-
tity and property values. An instance inherits its properties from the classes it is an
instance of and it has a specific value associated to each property. For example:
defines a recipe Herbed Asparagus which has ingredients: asparagus, parmesan and
herbs, originates from Italy and is suitable for vegetarians.
3 Property-based similarity
First of all, let us have a look at an example which should clarify the basics of our ap-
proach. We consider the domain of recipes where properties such as has ingredient,
has origin, suitable for diet are defined. These properties have one or more val-
ues assigned to them. Intuitively, the similarity among recipes depends on the property-
value pairs they have in common. Consider for example the following recipes: Asparagus
Parmigiana and Herbed Asparagus With Parmesan Cheese. They both have in-
gredients: Asparagus, Butter, Parmesan, Pepper among others and are both suit-
able for vegetarian diet. On the other hand, Indian Style Chicken has only Butter
in common with any of them and is not suitable for vegetarians. So the asparagus dishes
are definitely more similar among themselves than any of them with the chicken dish.
Hence, in order to determine similarity among two objects, we want to consider
both, their common features and distinctive features for each of them. To this aim we
use Tversky’s feature-based model of similarity (Tversky, 1977):
α(ψ(O1 ) ∩ ψ(O2 ))
simT (O1 , O2 ) = . (1)
β(ψ(O1 ) \ ψ(O2 )) + γ(ψ(O2 ) \ ψ(O1 )) + α(ψ(O1 ) ∩ ψ(O2 ))
where ψ(O) is the function describing all the relevant features of the object O, and
α, β, γ ∈ R are constants which permit different treatment of the various components.
For α = 1 common features of the two objects have maximal importance and for β = γ
non-directional similarity measure is obtained. In our approach we have α = β = γ = 1.
We will be using the following notation:
– common features of O1 and O2 : cf(O1 , O2 ) = ψ(O1 ) ∩ ψ(O2 ),
– distinctive features of O1 : df(O1 ) = ψ(O1 ) \ ψ(O2 ) and
– distinctive features of O2 : df(O2 ) = ψ(O2 ) \ ψ(O1 ).
Using this notation and setting α = β = γ = 1 the formula (1) becomes:
cf(O1 , O2 )
simT (O1 , O2 ) = . (2)
df(O1 ) + df(O2 ) + cf(O1 , O2 )
Since each of the domain objects has a number of property-value pairs describing
it, for each property p we will have to calculate how much it is responsible for common
features among these objects, as well as for distinctive features of each of them. We
denote these values by cf p , df1p and df2p . We consider equal the properties defined with
owl:EquivalentProperty.
3.1 Similarity among instances
In this work we present our approach only for instances of classes, although it can be
extended to classes defined with their properties and to classes defined as property re-
strictions (see Cena et al. (2012)). The essence of property-based similarity calculation
lies in simple comparison of the property-value pairs for each instance. Let us assume
that the property p has h0 different values in O1 and h00 different values in O2 , and k is
the number of times O1 and O2 have the same value for p, then
k2 h0 − k h00 − k
cf p = 0 00
, df1p = 0
and df2p = .
hh h h00
Let us assume that the objects O1 and O2 have properties p1 , . . . , pn in common. We
can repeat the above process for each property pi , i = 1, . . . , n.
Now, there are two possible ways to calculate similarity between O1 and O2 .
First, we can obtain all common and distinctive features of O1 and O2 :
cf(O1 , O2 ) = Σi=1
n
cf pi df(O1 ) = Σi=1
n
df1pi df(O2 ) = Σi=1
n
df2pi
where n is the number of properties O1 and O2 have in common. The similarity between
two instances O1 and O2 is then calculated using the formula (2):
cf(O1 , O2 )
sim(O1 , O2 ) = .
df(O1 ) + df(O2 ) + cf(O1 , O2 )
This method for property-based similarity of objects in the ontology was first intro-
duced in (Cena et al., 2012) classes defined with property restrictions but only for value
restrictions. It was further developed to include cardinality restrictions and applied to
categorization of shapes in (Likavec, 2013).
Second, we can calculate partial similarities w.r.t. each property pi , i = 1, . . . , n:
cf pi
sim pi =
df1pi + df2pi + cf pi
and then use these similarities to calculate the total similarity between O1 and O2 as:
sim(O1 , O2 ) = Σi=1
n
sim pi .
4 Improving property-based similarity
The above presented base case property-based similarity provides high rates of simi-
larity among objects which can be used in many applications. We still did not perform
the thorough evaluation but we evaluated it in the field of user interest propagation and
obtained very satisfying results (Cena et al., 2012). But, while performing the second
evaluation in this field, we became aware that in certain domains, this property-based
similarity of domain objects can be improved w.r.t. various aspects. We will discuss
here some of them.
4.1 Relevance of properties
When defining the concepts of a domain, not all the properties play an equal role.
Hence, it is possible to introduce the relevance of properties and assign different im-
portance to different properties in the domain. Actually, the relevance of a property can
be considered as the capacity of the property to determine the similarity between two
entities. For example, in the recipe domain, the property has ingredient is far more
important than has author and the two recipes with the same ingredients would be
considered more similar than the two recipes with the same author. So, the property
has ingredient would have a higher relevance factor than prophas author.
There are various approaches to calculation of property relevance in a domain. It
can be declared a priori and although effective, this solution may not be very feasible
for a huge domain. Also, it is possible to introduce an automatic method to determine
the relevance of properties. One possibility is to compute the similarity of concepts and
then to calculate the relevance factor for each property as the square of the average
similarity between concepts with the same value for that property.
4.2 Property or underlying hierarchy?
First of all, some aspects of the domain can be seen as properties, as well as underlying
hierarchy. So the question is, which way of modelling of the domain would provide bet-
ter similarity with human judgement. For example, in the recipes domain, the concepts
corresponding to dish type can be easily organised into a hierarchy and we can have
all the instances be instances of certain Dish Type classes. On the other hand, we can
simply have a property dish type and have all the recipes be instances of Recipe.
4.3 Relevance of values
One of the problems with the approach in which all the values for properties are treated
equally is that they might not contribute to the overall similarity with the same degree,
since some values might be more important in a certain context than the other. For ex-
ample, if we consider recipes and the has ingredient the values beef or asparagus
would be more important than salt or pepper. Hence, we come to the point where we
might need to introduce relevance for values, along the lines for relevance for proper-
ties. These would have to be proposed by domain experts or calculated by an algorithm
designed for this purpose.
4.4 Hierarchy of values
Another possible improvement of property based similarity is to take into account the
underlying hierarchy which might exist among the concepts used as values for proper-
ties. For example, if we consider recipes and the property has ingredient, one recipe
can have ingredient Fusilli and the other one Spaghetti. Although these two con-
cepts are not equal, they could be considered equal or equal to a certain degree (e.g. 80%
equal), since they are both types of pasta, and are descendants of Pasta concept. So
it might be possible to consider “almost equal” direct descendants of a certain concept
and even less equal second degree descendants of a certain concept.
5 Related work
There are various approaches to calculating similarity among concepts, depending on
the data structure used to represent the domain and on the amount and type of data
available about the concepts of the domain. The principal approaches to similarity cal-
culation are the following: (i) information content-based methods, (ii) distance-based
methods and (iii) feature-based methods. Various hybrid methods combine some of the
above methods.
In his seminal paper, Resnik (1999) proposes to calculate the semantic similarity of
concepts by calculating the information content in an is-a taxonomy of the closest class
subsuming both compared concepts. This similarity measure is given by the negative
logarithm of the probability of occurrence of the class in a text corpus. Another impor-
tant information-theoretic definition of similarity is introduced by Lin (1998) where the
similarity among concepts is calculated taking into account the shared information for
the two concepts and the amount of information needed to fully describe them.
The origins of the distance-based approach go back to Rada et al. (1989) where the
ontology graph structure is used to calculate the distance between nodes (i.e., the num-
ber of edges or the number of nodes between the two nodes) as a measure of their simi-
larity. Leacock and Chodorow (1998) use the normalised path length in WordNet (Fell-
baum, 1998) between all the senses of the concepts being compared. The semantic
similarity is computed as a negative logarithm of the ratio between the number of nodes
in the path which connects the given concepts and the maximum depth of the taxonomy.
Wu and Palmer (1994) take into account the depths of the given words in the taxonomy
and the depth of their common subsumer in their similarity measure.
Pirrò and Euzenat (2010) introduced a FaITH semantic similarity measure which
uses Tversky’s feature-based model and calculates the saliency of the features using a
new information content approach based on the ontology structure. This new framework
permits to calculate semantic similarity, as well as semantic relatedness and can be
used to rewrite the existing similarity measures so that they can also compute semantic
relatedness.
Smyth (2007) calculates the similarity by taking into account individual features of
concepts and by assigning to each feature its own similarity function and the weight
which helps distinguish the importance of individual features..
A semantic similarity measure for OWL objects introduced by Hau et al. (2005) is
defined as a ratio between the shared and total information content of the two objects.
The information content is calculated from the objects’ description sets containing all
the statements describing the given objects and is based on the number of new RDF
statements that can be generated by applying a certain set of inference rules to the
predicate.
The similarity measure introduced by Zadeh and Reformat (2013) is similar to ours
in the sense that it uses Tversky’s feature-based model for calculating similarity and
then calculates object’s common and distinctive features by observing all the relations
the objects have in the given ontology.
In the realm of “Conceptual Spaces” proposed by Gärdenfors (2004) the concepts
can be seen as convex regions in a conceptual space, whereas instances correspond
to points. The conceptual spaces are constructed using primitive quality dimensions
which represent various qualities of objects (e.g., color, shape, size). These dimensions
of conceptual spaces provide the means for determining similarity between concepts
and instances which can be defined as the inverse of their distance in the space.
Recently, Conceptual Spaces have been integrated with ontological formalisms to
form hybrid knowledge bases by Lieto et al. (2015). Since the points are represented as
vectors of the point coordinates (representing various object dimensions), their mutual
similarity is calculated as cosine similarity.
6 Conclusions and future work
In this work we present an approach to calculate similarity based on properties defined
in an ontology, as well as insights on which other factors can be included to improve
this similarity in different contexts. We limited ourselves to presenting the approach
only for the instances in the ontology, although the approach can be applied to classes
and classes defined as property restrictions as well. In addition, this approach can be
applied to linked open data Bizer et al. (2009) or any other structure where the objects
are described by means of their properties. For example, it would be interesting to apply
our measure of similarity to ConceptNet Speer and Havasi (2013), where an edge which
connects two nodes can be seen as a property and a target concept as its value.
In the case presented here, the prerequisite is the ontology with explicitly defined
properties for classes, rather than only a simple taxonomy of concepts. We only dealt
with object type properties in this work, since data type properties, such as literals,
require more complex analysis.
One of the limitations of the present approach, known for Tversky’s notion of simi-
larity, is that in the case of concepts with few properties defined for them, is is possible
that some concepts would be equally similar to the concepts which in reality have dif-
ferent degrees of similarity with them. This problem can be overcome by enlarging the
knowledge base with as many properties as possible for each concept. Also, by assign-
ing relevance to certain properties, the more important features would be taken into
account.
The evaluation of the approach on different datasets is being carried out and would
be published elsewhere.
Bibliography
Allemang, D. and Hendler, J. (2008). Semantic Web for the Working Ontologist: Effec-
tive Modeling in RDFS and OWL. Morgan Kaufmann Publishers.
Antoniou, G. and van Harmelen, F. (2008). A Semantic Web Primer, second edition.
The MIT Press.
Bizer, C., Heath, T., and Berners-Lee, T. (2009). Linked Data - The Story So Far.
International Journal on Semantic Web and Information Systems, 5(3):1–22.
Cena, F., Likavec, S., and Osborne, F. (2012). Property-based interest propagation in
ontology-based user model. In 20th Conference on User Modeling, Adaptation, and
Personalization, UMAP 2012, volume 7379 of LNCS, pages 38–50. Springer.
Fellbaum, C., editor (1998). WordNet: An Electronic Lexical Database. MIT Press.
Gärdenfors, P. (2004). Conceptual spaces: The geometry of thought. MIT Press.
Gruber, T. R. (1993). A translation approach to portable ontology specifications. Knowl-
edge Acquisition Journal, 5(2):199–220.
Guarino, N. and Poli, R. (1995). Editorial: The role of formal ontology in the informa-
tion technology. International Journal of Human-Computer Studies, 43(5-6):623–
624.
Hau, J., Lee, W., and Darlington, J. (2005). A semantic similarity measure for semantic
web services. In Web Service Semantics Workshop at WWW (2005.
Leacock, C. and Chodorow, M. (1998). Combining local context and WordNet similarity
for word sense identification, pages 305–332. In C. Fellbaum (Ed.), MIT Press.
Lieto, A., Minieri, A., Piana, A., and Radicioni, D. P. (2015). A knowledge-based
system for prototypical reasoning. Connection Science, 27(2):137–152.
Likavec, S. (2013). Shapes as property restrictions and property-based similarity. In
Kutz, O., Bhatt, M., Borgo, S., and Santos, P., editors, 2nd Interdisciplinary Workshop
The Shape of Things, volume 1007 of CEUR Workshop Proceedings, pages 95–105.
CEUR-WS.org.
Lin, D. (1998). An information-theoretic definition of similarity. In 15th Interna-
tional Conference on Machine Learning ICML ’98, pages 296–304. Morgan Kauf-
mann Publishers Inc.
Pirrò, G. and Euzenat, J. (2010). A feature and information theoretic framework for
semantic similarity and relatedness. In 9th International Semantic Web Conference,
ISWC ’10, volume 6496 of LNCS, pages 615–630. Springer.
Rada, R., Mili, H., Bicknell, E., and Blettner, M. (1989). Development and application
of a metric on semantic nets. IEEE Trans. on Systems Management and Cybernetics,
19(1):17–30.
Resnik, P. (1999). Semantic similarity in a taxonomy: An information-based measure
and its application to problems of ambiguity in natural language. Journal of Artificial
Intelligence Research, 11:95–130.
Smyth, B. (2007). Case-based recommendation. In The Adaptive Web, Methods and
Strategies of Web Personalization, volume 4321 of LNCS, pages 342–376. Springer.
Speer, R. and Havasi, C. (2013). The peoples web meets nlp. In Gurevych, I. and Kim,
J., editors, ConceptNet 5: A large semantic network for relational knowledge, pages
161–176. Springer Berlin Heidelberg.
Tversky, A. (1977). Features of similarity. Psychological Review, 84(4):327–352.
Wu, Z. and Palmer, M. (1994). Verbs semantics and lexical selection. In 32nd Annual
Meeting on Association for Computational Linguistics, pages 133–138. Association
for Computational Linguistics.
Zadeh, P. D. H. and Reformat, M. (2013). Assessment of semantic similarity of concepts
defined in ontology. Information Sciences, 250:21–39.