SEMANTIC (ARTICLE REVIEW)

     

     ARTICLE REVIEW

   Title of Journal : Understanding Semantics Relationship

   Veda.C Storey

\

•    Background of Study

Semantic relationships are the associations that there exist between the meanings of words (semantic relationships at word level), between the meanings of phrases, or between the meanings of sentences (semantic relationships at phrase or sentence level).One advance needed in database management systems (DBMS) is the capture of some of the semantics of an application for which a database is developed. In particular, there is a need within the database community to extend the relational model to accommodate more real world knowledge (Reiter, 1984). Some initial steps have been taken through the incorporation of certain semantic relationships into DBMS designs, commonly referred to as data abstractions. In general, an abstraction is a simplified description, or specification, of a system that emphasizes some of the system's details or properties while suppressing others (Shaw, 1984).

•     Explanation

I.        Semantic Relationships
Figure 1 presents a taxonomy of the seven types of semantic relationships analyzed here (inclusion, possession, attachment, attribution, antonym, synonym, and case). The taxonomy is based on the work of Winston et al. (1987) and Chaffin et al. (1988), but is expanded to include the other classes of semantic relationships identified by Landis et al. (1987). An interesting feature of the taxonomy is that it places the well-known data abstractions within the context of a broader set of semantic relationships. In particular, inclusion as found in the database literature corresponds to class inclusion in the taxonomy; aggregation corresponds to component-object; and association to member-collection.
II.      Database Design
This relationship is interpreted to mean that each company employs between one and many (*) employees; each employee is employed by one and only one company. In general, when an entity-relationship model is converted to a relational model, each entity becomes a separate entity relation (Teorey et al., 1986) that looks exactly the same as the entity (i.e., it has the same key and non-key attributes.) This entity relation may then be modified if a relationship is represented by adding another entity type's key to it as a foreign key. Teorey et al. (1986) refer to this modified entity relation as an extended entity relation. Alternatively, a relationship may be represented by creating a separate relationship relation whose key is the concatenation of the keys of the involved entity types and whose non-keys are the relationship's attributes. In general, when the min/max cardinalities of one entity type are (1,1) in a relationship, the key of the other entity type is added as a foreign key to the entity type with the (1,1) cardinalities.    
Meronymic Inclusion. Meronymic (from the Greek word "meros" for part) relationships occur between something and its parts (Winston et al., 1987). Seven different types of meronymic relationships, based on Storey (1991b), are discussed below
1.    Component-Object
2.    Feature-Event.
3.    Member-Collection.
4.    Portion-Mass
5.    Phase-Activity
6.    Place-Area
7.    Stuff-Object

III.   General Relationships
Some relationships are very general, leading to ambiguity in a design, particularly, have~has relationships of the form A have B or A has B. Although this relationship is often used in the entity-relationship model, it can easily be seen to have multiple interpretations (Storey, 1988). The verb phase have~has can be used to convey either an attribute, or a relationship in which the best interpretation could be almost any of the semantic relationships previously discussed. It would, therefore, be beneficial to use the interpretation that best reflects the semantics of the application and reserve the general form of this relationship for situations where it is not possible to clarify the semantics further. Various interpretations and types of difficulties associated with the have~has relationship are outlined below. These also illustrate why it is important to understand the semantics of an application and the sometimes subtle differences among the relationships.
IV.   The Semantic Relationship Analyzer
A prototype system, the Semantic Relationship Analyzer, has been developed that implements the preceding analysis. 8 The system elicits relationships of the form A verb phrase B from a user who is either a database designer or an end-user. For each relationship, the Semantic Relationship Analyzer obtains the min/max cardinalities and then tries to capture the best interpretation for the relationship. For example, because a have~has relationship is ambiguous, the system determines whether such a relationship would be better expressed bypart-of or one of the other relationships that can easily be confused with have~has.

    

    Summary 

      So data abstractions are often used in data modeling, research in linguistics, logic, and cognitive psychology has identified many more semantic relationships. The objective of this article has been to analyze a number of these lesser-known relationships in terms of their design implications and ways they could be employed by database design systems to capture some of the semantics of an application. It is hoped that the Semantic Relationship Analyzer, or its knowledge base, could eventually be incorporated into a database design system to assist the user in providing the best input. Ideally, this would result in more of the semantics of the real world being captured in the final design.

   Source


(1)http://www.vldb.org/journal/VLDBJ2/P455.pdf