Report on Knowledge Representation & Semantic Web Technologies

3.1       Introduction of Knowledge Representation & Semantic Web Technologies

The literature review in this chapter of the thesis is written with the intent to introduce knowledgebase systems and semantic web technologies. It includes an overview of the research, concepts, as well as technologies from several scholars. It also establishes information into a machine-readable format. The technologies covered in this chapter will investigate ways they can be used as the start-up point to explore the development of the Smart-SCCS infostructure.

3.2       Knowledge Representation & Semantic Web Technologies

It needs a framework in the workplace in which information can be made explicit for machines to help with information management issues. Such type of framework will allow machines to automatically process knowledge, and be able to share knowledge. However, a common problem of knowledge representation then becomes the development of succinctly accurate notation for representing knowledge. The knowledgebase is a collection of facts generated from large volumes of data which can be incomplete or imprecise. Therefore, the meaning, as well as the procedures, is deemed to be better as compared to the database and the reason is that it provides the power of reasoning to solve complicated questions. Knowledge representation is used as an encoding method that is mentioned as the following: knowledge, beliefs, actions, feelings, goals, preferences, desires, and other mental states in the Knowledgebase. By exchanging knowledge to reasoned Knowledgebase, Semantic web can define the principles. Therefore, intending to build up a good Knowledgebase, it is also necessary to make better knowledge representation. In this situation, the main consideration is knowledge representation.

Knowledge representation is a topic under-developed in both cognitive science and artificial intelligence (AI). Cognitive science is concerned with the way people process and store information. It is the theory base behind KR, because studying human thinking, particularly mental states, representational structures, and computational procedures, i.e., thinking, reasoning and operating on them in the mind, is a key issue in this discipline. It is assumed in cognitive science that the human mind can be mentally represented similar to computer data structures. Thus, the mind function can be simulated by computational algorithms.

Artificial Intelligence is the field of computer science that investigates the nature of human understanding, knowledge and mental abilities, and applies theories to the software. AI focuses to develop software programs to perform tasks that simulate human behaviors. In other words, it tries to approximate human reasoning outcomes by running programs to structure, represent, encode and manipulate heuristic and factual knowledge.

There is no single ideal KR technique appropriate for all applications. This indicates that developers of intelligent systems should choose the KR technique that best suits the application being developed. In other words, the technique that can end up driving the application should be selected as the technique chosen can define outcomes through the way that it works. This highlights the importance of being aware of the various techniques to select from. The rationale behind selecting ontology rather than other technologies in this research is discussed in the following section.

3.3       Data, Information, Knowledge Concepts

Data, Knowledge, and information will be us throughout this thesis. The meaning of these three terms can easily be confused by the reader; therefore, we will start by defining their meaning in the simplest context possible. Figure 3.x illustrates the relationships between information, data, and knowledge. The black arrows are pointing up to show all the important roles in the knowledge regarding the data interpretation, learning of knowledge, and information elaboration. The following model illustrated will be investigated accurately throughout this section’s remainder.

 According to (Aamodt & Nygård, 1995) the concepts of data, information, and knowledge are characterized as follows;

• Data: Data is mainly a syntactic entity that consists of meaningless patterns; it is also considered as an input for the interpretation process. The use of data is the initial step for the decision-making process.

• Information: Information can be described as interpreted data about the information used in the data is regarded as a meaningful process. Information is extracted as the yield from data analysis and it is also the input to the process. in case of output from information regarded as the  procedure for the process of decision-making.

• Knowledge: Knowledge is the information taken from the data and included within the reasoning resources agent. It is prepared completely as ready for dynamic application in the process of making a decision; it is the productivity of a knowledge progression.

Several authors who have looked into the relationship between data and information have agreed on the differences. Silver, for example, considered different perspectives about the production process in which rare substances along with information are considered as the data as the product. Researchers’ concluded with a similar distinction to ours. The only dissimilarity is that, here, we stress the importance of the  difference. Throughout an elucidation procedure, the structure of syntactic is further converted into the semantic, important unit, making the notion of interpretation central. The role of knowledge is needed to perform the interpretation. The interpretation is given through the dynamic element within the converting data processes into information, acquiring new knowledge as well as deriving other information, specifically to learn. The process further goes through the following summary about knowledge roles:

3.3.1    Data interpretation of Knowledge Representation & Semantic Web Technologies

Data interpretation defined as differences between information and data. The data consist of different parameters including uninterpreted characters, signs, signals, and patterns. The data consider all the system concerned. Data converts to information once it has been construed to present the significance. This can be seen in figure 3.x by the arrow of Data Interpretation. For data to be construed into information, knowledge is required by a system. The type of data is semantics and it requires process and information to interpret data and the information in the decision process and it produces strong relation to pragmatics, i.e., explanation of real-world framework as well as a specific objective, and it is not different from the language-syntactical semantics. The interaction is between methods of data interpretation and the knowledge determines a process within the system’s ability to recognize the data. Related to the process of data interpretation from the human perspective, the decision-maker might consider the cultural setting, recollections of related interpretation previously, unconscious bias, a prospect that is generated by particular point of view, in addition to textbook information and domain reliant heuristic policies, with a purpose to resolve the related connotation of information. Meanwhile, the  computer systems, unfortunately, have not satisfied the basic requirements and degree of sophistication yet, with the consideration that they never will. There are some issues regarding the systems physical recognitions, and another that is diverse from the theory issue and the computational procedures that we are investigating now.

3.3.2    Elaboration of information of Knowledge Representation & Semantic Web Technologies

Initially, the information is specified by some initial interpretation and the information is given by the process described above, it delves further into detail so that it is better implicit and designed for developing innovative data. This can be seen at the  in Figure 3.x. The data is further collected by considering the additional problem features, consequences of hypotheses, generated hypotheses, solutions from the problems, justifications along with suggestions and critiquing arguments. The elaboration process can be recognized based on the actual problem-solving process, in which the central decision-making takes place. The data interpretation and information could be simplified through pre-processing based on the core decision-making process. However, considering the realistic approaches in the decision-making process it can be used for elaboration as well as data interpretation. The initial interpretations adapt to changes during elaboration, this leads elaboration to question the environment, which in turn leads to revised or new data being entered. Elaboration includes object-level inference and control level strategic reasoning. The elaboration processes depend on the implicate methods, for instance, spreading activation, rule chaining, associative memory retrieval, etc and tends to end in a termination. This could either be an ultimate decision or else an impermanent stop so that a hypothesis or question to the environment can be raised for further information. When considering both situations and cases of a system it will return from one point of view, based on the information, but it is different from the perspective of receiver and data. If the receiver can interpret the data, then the delivering agent will turn it into appropriate information for the receiver as intended. Therefore, a common body of knowledge is important between agents.

Information and knowledge play two different roles:

1.      Information that has been interpreted from data is provided as input for the elaboration process, and it is also used to produce relative output from it.

2.      The knowledge is used to have a better decision process. It is a further inherent resource to provide for the reasoning agent and it enables inferring existing and the new information.

A computational method is used based on inference and the knowledge and it indicates several conjectures based on the fundamental processing that has completed. Non-deterministic processing defines as a central and can sometimes be the only resource available when the optimal ways of processing input data are not well understood. Due to the category of issues addressed, the processing techniques described are complex and require a flexibility level that would not effortlessly be realized through the stringent algorithmic methods. In such a category of computation, there is a cognition-stimulated language that is based on the knowledge structures as well as the inference techniques have. The process becomes suitable while characterizing and realizing all the processing techniques.