A Chinese time ontology for the Semantic Web

Abstract

Representation of and reasoning with temporal knowledge are fundamental in information systems that involve changes and actions. To build such systems, a time ontology is demanded. The development of a time ontology is also an indispensable part of effort to realize the Semantic Web. Nevertheless, our practice shows that any practical time ontology is closely related with a specific calendar, culture or history. To this end, this paper presents a Chinese time ontology for knowledge systems and web services which involve temporal entities or temporal properties. First, we define a base time ontology. As a core component, it consists of a time system, a timing system, a Gregorian timing system, and a timing ontology. Upon this base ontology, other parts of the Chinese time ontology are finally constructed, including the traditional Chinese timing system, temporal representation in Chinese idiosyncratic ways, and transformation between temporal entities in the Gregorian timing system and temporal entities in the traditional Chinese timing system. We will argue that the base time ontology is not only a basic and integral part of the Chinese time ontology, but also a base for constructing other time ontologies.

Introduction

One of the crucial problems in information and knowledge systems that involve action and change is the representation of and reasoning with time. To develop such systems, a time ontology is demanded. Building a time ontology is also necessary to realize the Semantic Web, which aims to provide automated web services based on the descriptions of the contents and capabilities of web resources [27], [24]. A time ontology is a specification of a conceptualization for temporal knowledge. Currently, much effort has been made on building explicit time ontologies, such as the DAML ontology of time [17], [18], [42], the time ontology in OWL [47], KSL time ontology [44], the time ontology in KIF [48], Times and Dates in Cyc knowledge base [50], the time of DAML-S [43], the temporal portion of IEEE Standard Upper Ontology [49], and other works [11], [15], [28], [32].

In working on time ontologies, our experience indicates that time ontologies are closely related with specific nations or cultures, though they may share a common part. This is especially true if nations, e.g. the Chinese nation, have a long history. For example, in our work on knowledge processing of Chinese historical knowledge [4], timing methods are associated with various historical events; in fact, each Chinese dynasty or kingdom had its own particular timing method. As another example, in our work on agricultural knowledge acquisition [5], we found that ancient farmers used a timing method of 24 Solar Terms for farming (e.g. what crops and animals are planted and raised during what solar terms?). Although no one exactly knows when this approach was created, it is still one of the dominant timing methods in Chinese societies.

In spite of the great progresses in the past years, there are still two important issues about time ontologies that need to be further addressed, which constitute the motivation of this paper. One is that most existing ontologies count and express time according to the Gregorian calendar, and those ontologies are incapable of computing and representing time for people in societies with their own calendars. In other words, the effect of calendars in constructing time ontologies is simply ignored. However, calendars play an important role in building time ontologies. Actually, calendars are the basis of computing durations of and relationships between time units, and different calendars constitute different timing systems. In addition, to realize the Semantic Web, it is necessary for machines to be able to process and understand the semantics of temporal entities in various calendars on the World Wide Web [26]. Hence, it is indispensable to build up mappings between temporal entities in different timing systems within the same or different calendars.

Another issue is the diversity of temporal computing and representation. Current time ontologies compute and express time mainly in a manner of the calendar date. However, there are other various approaches of computing and representing time that are related to a country’s culture, history, and agriculture and so on. People in different social-cultural contexts use different approaches to count and express time. This knowledge should also be addressed when building time ontologies.

In this paper, we present a Chinese time ontology based on the current Chinese calendars: the Gregorian calendar and the traditional Chinese calendar. The Gregorian calendar is a kind of solar calendar, and is used almost everywhere in the world. The traditional Chinese calendar is a type of lunisolar calendar [39], [51]. The ontology is developed for web services and knowledge systems that involve temporal entities or temporal properties. We build our ontology in two steps. First, we build a core component, called the base time ontology. Specifically, it consists of a time system, a timing system, a Gregorian timing system, and a timing ontology. Here, the Gregorian timing system is employed as a timing system based on the Gregorian calendar. Second, upon this base time ontology, we develop the other parts of the Chinese time ontology, including a traditional Chinese timing system, temporal representation in Chinese idiosyncratic ways, and transformation (mapping) between temporal entities in the Gregorian timing system and temporal entities in the traditional Chinese timing system. Here, the traditional Chinese timing system is a timing system founded on the traditional Chinese calendar. Furthermore, we will also argue that the base time ontology is not only a basic and integral part of the Chinese time ontology, but also a base for constructing other time ontologies.

The remainder of this paper is organized as follows. Section 2 reviews the related work of time ontologies. Section 3 presents our Chinese time ontology. Section 4 compares our Chinese time ontology with other time ontologies, and introduces its applications in question answering and web services. The discussion is given in Section 5. Section 6 concludes this paper.