6

Conclusions

This thesis has first provided a brief survey about ontological engineering by mentioning the aggregates of the field and by reporting the latest progressions as well as the current pitfalls. As such, this survey can be used as a brief reference for today’s ontological engineering. In addition, it can aid the development of an ontology by means of providing information about suggested methodologies, ontology libraries, languages, editors and examples of current ontologies.

Second, this thesis has pointed out to the need for an ontology of socio-cultural time that formally describes the different time conceptions of various cultures and social groups. Subsequently, the thesis has described the characteristics of the two ontologies that have been developed with the aim to answer the need. In particular, the Ontology of Socio-Cultural Time Expressions Related to Nations, Religions, Business Life & Education and the Ontology of Socio-Cultural Time Concepts Related to Nations, Religions, Business Life & Education have been described, whereby the latter has comprised the core of the thesis.

The Ontology of Socio-Cultural Time Concepts can be used in several applications. It can provide semantic annotation for the Web pages, it can be integrated into a temporal type system to enable calendrical calculations about socio-cultural time and finally it can support a Web based automated appointment scheduling service. In the last chapter, we have discussed the means for the realization of these applications.

The following two subsections will both report the results of this thesis and will point out to the future research issues, which necessarily arise from the given results.

6.1 Results

We have partitioned the domain of socio-cultural time related to nations, religions business life & education into two levels, which are the lexical level and the conceptual level and for each level we have developed a separate ontology from scratch. This approach is along the same lines with those suggested by the NLP ontologists as discussed in [11],[12] and [13],[14].

Most current temporal and cultural ontologies, e.g. [48],[58],[39] are large projects that have been developed with different intentions in mind. Some temporal ontologies constitute a part of a far larger ontology, e.g. Cyc. Therefore, it has not been possible to reuse these in the development process of our two ontologies. Current technologies in ontological engineering, only allow the reuse of entire ontologies. That is, it is not yet possible to efficiently extract the relevant components of one ontology and plug them in another ontology. Therefore, it has not been possible either to reuse the relevant components of the ontologies we have mentioned above.

The two ontologies need to be extended in terms of their contents and in terms of the scope of the domain to be more representative. Both ontologies provide declarative knowledge only about a subset of the domain of socio-cultural time (i.e related to nations, religions, business life & education), which could be extended to include further socio-cultural groups such as sports, healthcare or others. The focal ontology, which is the Ontology of Socio-Cultural Time Concepts, can be primarily filled with more socio-cultural temporal data to upgrade it to a status of a more comprehensive knowledge base.

Currently, the temporal information provided in the ontology of Socio-Cultural Time Concepts solely concerns Gregorian calendar. In order to become more socio-cultural context sensitive, it requires information about other calendars such as the Islamic or the Hebrew calendar.

Reasoning about the ontology can only be done by means of external reasoning mechanisms. We have discussed the two current reasoning mechanisms that provide this service.

To conduct calendrical calculations about the time periods defined in the Ontology of Socio-Cultural Time Concepts, the ontology essentially needs to be connected to a temporal type system. Consequently, we have referred to a (socio-cultural context-sensitive) temporal type system called Towards a Multi-Calendar Temporal Type System for (Semantic) Web Query Languages in [80].

With the Ontology of Socio-Cultural Time Expressions we have referred to the area of Natural Language Processing (NLP) and ontologies, which we discussed in Chapter 3. Based on the fact that our motivation is not linguistic, we have not delved into the details of the expressions of socio-cultural time. We have merely stated the fact that the expressions of socio-cultural time, being a subset of natural language expressions, are not adequate for an unambiguous and systematic description of the domain because they are naturally vague. This issue needs to be referred to in more detail, so that the ontologies can provide linguistic support about the domain of socio-cultural time related to nations, religions, business life & education.

Currently, there are two concrete applications that use the Ontology of Socio-Cultural Time Concepts. For the first application we have uploaded the ontology to an ontology repository on the Web that supports query facilities. This enables the ontology to be accessed and processed online. Using the RDF query language RDFQL we have queried and extracted some instances of the ontology, for example the similar socio-cultural time periods or socio-cultural time periods that imply a general suspension of work.

Regarding the second application, we have used the extracted instances along with the others, to implement a small database in PROLOG that includes calendrical information about these time periods for the year 2004. If a calendar date in 2004, or duration, or an origin, or all of these specifications together are given to the PROLOG program, it delivers the associated socio-cultural time period. As such, we consider this application as the foundation of a socio-cultural calendar, which has been picturized.

Finally, we have illustrated a real world scenario as a possible future application for the Ontology of Socio-Cultural Time Concepts. The scenario concerns the issue of a Web based appointment scheduling process, which requires information about the socio-cultural context of the parties involved in the process. With this information, the Web-based appointment scheduling service can react in a more context sensitive manner to the culture specific temporal preferences of its users. We have reported that work is in progress for such Web-based automated appointment schedulers, for example the Carnegie Mellon University’s Retsina Calendar Web Agent. Such an application can use the ontology in order to extend its temporal information to include socio-cultural temporal information about the given nations, religions, business life and education.

6.2 Future Work

As a result of this thesis we have identified several research issues that require further investigation. Future investigations concern both ontologies, which have been developed within the scope of this thesis, however they will tackle different problems.

First ontology, which is the Ontology of Socio-Cultural Time Expressions, does not provide the right means for a systematic explicit and unambiguous description of socio-cultural time to support a temporal Web application system because of the previously discussed reasons. However, this does not imply that there can be no marriage between ontologies and natural language expressions. On the contrary, ontologies have started to become very popular in the Natural Language Processing area and as Bateman states [11],[12] today’s ontologists should by no means disregard the role of natural language in modelling knowledge about the real world. Parallels, a group of terminologists and datalinguists [13],[14] are working on systems that map conceptual data to lexical data with the objective of acquiring meaning and providing terminological consensus without being trapped by the ambiguity of natural language expressions.

Clearly, the Ontology of Socio-Cultural Time Expressions needs more attention in this direction. With a linguistic motivation, efficient methodologies can be sought for to map between the Ontology of Socio-Cultural Time Expressions and Ontology of Socio-Cultural Time Concepts.

A systematics for (semi)automatic extraction of socio-cultural time expressions is necessary. Acquisition of information about socio-cultural time expressions has been a highly time consuming procedure. Since such conceptions are mostly implicitly present, the information needs to be assembled from a variety of resources such as books, newspapers, the Web, human conversations, media and so forth. When searching for socio-cultural time expressions denoting time periods on the Web, we have used the following query pattern: “subject matter” + “calendar”. For example, “medical calendar”, “Turkish calendar”, “Jewish calendar” etc. In most of the cases this query pattern returned satisfactory results. We have retrieved pages that contain dates relevant to the subject matter such as various important dates in 2004 for Jewish religion. The surroundings of the date entries in the retrieved pages usually contained the necessary expressions, such as “Pessah”, “Passover”, “Jewish New Year”, etc. On the basis of this experience, information retrieval and information extraction methods can be investigated to retrieve Web pages matching the query pattern above and to extract the expressions in the surroundings of date entries, month names or in the surroundings of certain words such as the word “holiday”.

Finally, language support is necessary for the ontology. Currently, the ontology is conceptualized and implemented in the English language. However, due to the nature of the domain, the ontology refers to a variety of developers and users, who come from different socio-cultural contexts and who speak various languages. The ontology can be translated into other languages to increase user convenience.

 socio-cultural time can be captured by inspecting the expressions of socio-cultural time. In other words, it is possible to gain insight to the time conceptions of a culture or a social group by observing how their members talk about time. We will show the ways to represent this information by means of ontology.

Once information about socio-cultural time is formalized in a machine processable way, it can be used to support temporal Web application systems such as a Web based automated appointment scheduling service. We believe appointment scheduling is a task, where different conceptions of time can be frequently observed, when parties involved in appointment scheduling process come from different socio-cultural backgrounds. In [79] the characteristics of such a Web based automated appointment scheduling service is defined and it is stated that current Web based appointment scheduling services fall short in considering the notion of socio-cultural time. Being supported with this kind of temporal information, such services can react in a more sensitive way to their users’ socio-cultural context dependent temporal needs and preferences. Additionally, on the basis of the socio-cultural temporal information defined in the ontology, socio-cultural calendars can be developed for individual use or as a component of a Web based automated appointment scheduling service. Clearly, this information can also be used for the semantic annotation of Web pages

BACK TO MAIN PAGE