2014 Fifth International Conference on Intelligent Systems, Modelling and Simulation

Framework of Indigenous Knowledge Representation Ng Boon Ding

Edwin Mit

Faculty of Computer Science and Information Technology Universiti Malaysia Sarawak 94300 Kota Samarahan, Sarawak. [email protected]

Faculty of Computer Science and Information Technology Universiti Malaysia Sarawak 94300 Kota Samarahan, Sarawak. [email protected]

Abstract—There are complex and multi-relationships between taboos, beliefs and its consequences. The truthfulness of the knowledge acquired from the relationships is very uncertain, vague and ambiguous. The heterogeneous consequences of similar taboo and belief of different communities affect the preciseness of the capturing, modeling and manipulating the indigenous knowledge. This paper presented a formal-fuzzy framework for the indigenous communities’ taboos and beliefs in Borneo which will be used to model the truth values of the relationships between taboos, beliefs and consequences. Formal-fuzzy approach is an integration of formal method and fuzzy logic. This approach is used as formal method is wellknown for specifying a precise, complete and consistent software model and fuzzy logic is a well-known approach dealing with uncertainties problem. This proposed framework is also expected to be used to preserve the knowledge behind the indigenous community taboos and beliefs. This knowledge is a baseline to derive the logical justification or explanation of the taboos and beliefs, related to current life style.

technologies and religion had a great impact on the remote indigenous taboos and beliefs. Causing some of the beliefs and taboos not being practiced anymore [1]. Although there are many members of the indigenous community are aware of the existence of the taboos and beliefs but, they do not know the reason behind it. The main motivation of this research is to study the logical explanation or justification and knowledge extracted from the taboos and beliefs. If the logical explanation and knowledge studied on the taboos and beliefs could be applied and integrated with the modern technologies this is expected to improve the daily life style of the indigenous community. For an example: Taboo: ‘children are not allow to eat sugar cane after dusk’ Consequence: ‘bird will cut their ears’ Truth value: ‘the degree of truth that this will happen’ Logical explanation: ‘Children are not allowed to eat sugar cane after dusk is because the children tend to throw the sugar cane waste everywhere, which will cause ants to come to the house’. Proving truth values for the knowledge are not easy. Knowledge is uncertain, vague and ambiguous [2]. It does not have a real concrete evidence to show the relationships of abstract and logical values. There are also no knowledge representations in the taboo and consequences for the indigenous communities. Thus, the study of knowledge acquired will be used to derive the truth. On the other hand, there is also no logical explanation for the relationships between set of taboos and its consequences, thus, this research will derive a set of logical explanation by mapping the beliefs, taboos and consequences to the logical explanations. The truth value of consequence is measured based on the logical explanation. The logical explanation will explain further in section III. This paper is organized as following: Section I which is the introduction and related works and some literature reviews. Section II describes the overall framework designed, section III, the methodology to carry out the research, section IV discussion of the approach, benefits and challenges towards modeling the beliefs and taboos, section V the conclusion of current stage of this research.

Keywords – formal-fuzzy, uncertain, fuzzy logic, formal method, truth value, knowledge repository

I. INTRODUCTION This is a framework to extract knowledge from the beliefs, taboos and consequences for the indigenous communities in Borneo. There are complex and multi-relationships between beliefs, taboos and its consequences. As an instance, capturing, modeling or manipulating the information and relationship between taboos, beliefs and the consequences are not easy. The difference between taboos and beliefs of different communities will also affect the preciseness of the modeling. This framework is trying to model a precise relationship between beliefs, taboos and its consequences by using formal-fuzzy approach. This framework is expected to be used to preserve the knowledge and good values behind the indigenous community taboos and beliefs. Thus, an expandable knowledge repository is required. VDM++ formal specification language will be used to do knowledge representation for the knowledge acquired from the sample data. To solve the uncertainties for the knowledge acquired, fuzzy logic will be applied to define the truth value of the consequences. Later, the study on logical justification of the taboos and beliefs is expected to be used to define the knowledge or good values of indigenous community’s taboos and beliefs. There are a number of indigenous community taboos and beliefs that had not properly recorded. Besides, the modern 2166-0662/14 $31.00 © 2014 IEEE DOI 10.1109/ISMS.2014.11

II. RELATED WORK This work is a part of the continuous research in modeling indigenous community model. The initial model defines the relationship between cultures and beliefs [1]. It 18

is not possible to see the logical relationship between beliefs and consequences. In regards to health belief systems, Western health professionals often experience difficulties in service delivery to Aboriginal people because of the disparity between Aboriginal and Western health belief systems [3]. In the background study, as an example, remote community in Long Lamai believe that if they hear “ketupong” bird on the way to farm, this indicates that an accident is going to happen to the person. It is difficult and quite impossible to prove it scientifically that the relationship between the bird chirping and an accident is happening to a person. There are a number of beliefs related to their daily lives and also the stages of their lives (e.g., starting from baby, growing children, teenager, looking for partner, having family, looking for place to stay, selecting leader, death, etc). Some cultures related to beliefs, and some are normal practices such as dance and rhyme. However some rhymes may relates with beliefs. The beliefs are triggered by some cultures and also the external factors or events such as bird, animal and natural events (e.g., falling tree, or branch of tree). The idea of this research is to study the good values behind the indigenous community taboos and beliefs based on the logical justification. There are complex and multi relationships between taboos and beliefs and its consequences. This study is trying to define precise relationships between taboos, beliefs and its consequences. Mit, E., et. al. in [1] had designed a new architecture of genealogy software to capture the information of cultures and beliefs into the family tree. The authors provided a platform to collect data in the scope of marriage process of the indigenous communities. The effort of preserving the remote community cultures, languages and beliefs and also the management of indigenous knowledge is insufficient without extracting the good values out of the knowledge by using modern technologies [5]. In a need analysis study carried out by Mit, E., et. al. in [1] in the year 2012 on the culture-based genealogy software for the indigenous communities in Borneo, showed that there are 76% of the survey respondents wanted their children to know about their culture, beliefs and taboos related to the marriage in their communities. The authors concluded that most of the communities wanted to preserve their cultures, and make it known to the outside world. The preservation of Borneo indigenous community’s cultures and beliefs are important. However, until today, the indigenous community still lack of documentations for their culture and beliefs information due to the traditional technique of knowledge pass down, which is through verbal communications [4]. Even though there are some newspaper articles and journal recording some of the taboos and beliefs however these resources are limited i.e., through interviewing the indigenous communities. Winschiers-Theophilus, et.al., had also put in the effort in the preservation of the indigenous knowledge in Borneo [4]. Here, the authors captured the cultural data related to the

daily life of the indigenous communities by using mobile technologies. A. Fuzzy logic to solve the uncertainties in knowledge acquired Knowledge usually gained through experiences or acquired from set of information. Knowledge is often abstract, incomplete and vague. [2]. Thus, pure inference system can hardly represent human knowledge. Truthfulness of knowledge usually difficult to determine, one of the methods to validate the truthfulness of the knowledge is by tracing the originality of the piece of knowledge, also known as knowledge provenance. Knowledge provenance is a method to trace the validity and origin of knowledge. Huang and Fox had also proposed 4 levels of knowledge provenance, to solve the uncertainties of the knowledge on the web [14]. Other than that, there are a few ways to handle uncertainties, namely probabilistic analysis, fuzzy analysis, Bayesian analysis, soft computing techniques and rule based classification technique. However, fuzzy analysis is one of the most effective methods [6] [7]. In normal remote community practice, the consequence is measured on either it happen or not. In traditional propositional logic, this is represented in the form of crisps value, which is either true or false, [0, 1]. However, in the real world the impact of the consequence is measured based on the “degree of truth”. This can be represented by using fuzzy logic. Fuzzy logic, founded by Lotfi A. Zadeh [8], is used widely in researches in order to solve the problems related to handling vagueness and uncertainties in knowledge [9]. It has also been used in many researches such artificial intelligence [21], genealogy research, bioinformatics [22], formal-fuzzy approach [23]. The reasoning in fuzzy logic is similar to human reasoning. It allows for approximate values and inferences as well as incomplete or ambiguous data (fuzzy data) as opposed to only relying on crisp data (binary yes or no choices). Fuzzy logic is a form of many-valued logic; it deals with reasoning that is approximate rather than fixed and exact. B. Knowledge representation Description logic plays an important role in knowledge representation, where concepts, roles and individuals are the foundation of those logics. However, when it comes to the verbally delivered language, it is not easy to represent the knowledge passed down by the ancestors. This is because description logics have no well-defined boundaries when it comes to the non-crisps value concepts [10]. For the knowledge representation in querying knowledge bases, Nguyen et. al. [11], gives a recommendation of using a novel method to give an efficient web-page recommendation through semantic enhancement by integrating domain and web usage knowledge of a website. They introduced two models to represent the domain knowledge, which are the ontology method and also an

19

automatically generated semantic network to represent the knowledge. In their paper, they had come out with a set of recommendation strategies for the prediction of next webpage requests of users in the process of querying the knowledge bases [11]. On the other hand, Goncalves et. al. [12] proposed a Formal Method in implementing fuzzy requirement. The authors introduced a fuzzy query language and by using formal specification techniques to avoid ambiguity of verbal language. By using formal method, they specified the properties and behavior of the system in a formal language. This method ensured a precise specification in the application development for the fuzzy queries. This solved the uncertainties in spoken language when queries are done in databases. As a summarization of the reviews, knowledge representation is widely used in representing uncertainties in knowledge. This comprises of the technique of developing and designing ontological model, semantic network model, and also formal model. Among all the models, formal method (formal model) is proposed to be used in this research, as it provides a precise specification to develop the knowledge repository. Other than that, formal model supports the logical and mathematical formulas and also set theories which it is much useful in doing knowledge representation. Thus, this research continues the work from [5] which it will be part of the continuous research in modeling the indigenous community model. This research will adopt the methods used in [14] to solve the uncertainties of knowledge by using [8] fuzzy logic membership theory and to model the specification by using formal method[12]. This paper describes the formal-fuzzy specification that models the relationships between beliefs and consequences. For example, a fuzzy set is a pair of (B, C) where B is a set of belief, and C is set of consequences, and C : B |-> [0,1]. For each x ∈ B, D(x) is called the degree of impact of x in (B, C). For a finite set B = {x1,...,xn}, the fuzzy set (B, C) is often denoted by {D(x1) / x1,...,D(xn) / xn}. Let x ∈ B, then x is called not included in the fuzzy set (B,C) if D(x) = 0, x is called fully included if D(x) = 1, and x is called a fuzzy member if 0 < D(x) < 1. The fuzzy values of consequences will be determined together with community leader or spiritual leader. The study on logical justification of the taboos and beliefs is expected to be used to define the knowledge or good values of indigenous community taboos/beliefs. In the later section, the framework also expected to be used to preserve the knowledge or good values behind the indigenous community taboos and beliefs, similar to “Feng Shui”, developed 3000 years ago, which consists of a complex body of knowledge [20]. The degree of consequence is differs based on the event. The same external events (e.g., hear bird, hear/see animal) may results in different degree of consequences based on the type of the events (e.g., type of bird - eagle, “ketupong”, “kepiotoki”,

type of animal - mouse, mouse deer, deer) or if the person offers the gift to the omen. The relationships between a set of beliefs and a set of consequences will be defined by using VDM++ formal specification language. Where, it will be used to model the relationships between taboos, beliefs and consequences based on its properties that will force software developer to identify the missing or incomplete requirements in order to produce a complete, precise and consistent software model [19]. III. FRAMEWORK OVERVIEW The first stage of this research is Data Collection. Data collected will be in a set of belief and taboos and a set of consequences. Then, next stage will be data mapping. Data mapping is done by using formal method. Here, the set of taboos, beliefs and consequences collected will be undergone through a process of data mapping. The last stage of this framework is to generate a fuzzy value for knowledge acquired in the knowledge repository, and then come out with a logical explanation. Fig. 1 shows the overall architecture of the framework for formal-fuzzy logic approach in defining the knowledge repository.

Figure 1: The architecture design for formal-fuzzy logic approach to solve the problem in this research

IV. METHODOLOGY Initial phase of this research will involve in the data collection activities. Here, set of taboos and beliefs and set of consequences will be collected. Taboos, beliefs and consequences will be collected and later categorize in the following sets: Taboos/Beliefs : T = {t1, t2,…, tn}. Consequences : C = {c1, c1,…, cn}. Logical Explanations: L = {l1, l2,…, ln }. The main technique of data collection is by interviewing the expert for the indigenous communities (such as the communities’ spiritual leader or “tuai rumah” which is the leader for the long houses). The sample data also obtain from newspapers articles and journals which are relevant, where those resources also the results from interviewing from the indigenous communities. Later, after the statement and fact of the set of taboos/beliefs and consequences are refined, the sets of data will undergo data mapping process, which is basis for a simple mapping rules. The knowledge, Ki is defined as a

20

V. DISCUSSION The knowledge representation will be carried out by using formal method which has the advantage of producing a consistent, precise and complete software model for the knowledge repository. Fuzzy logic will be applied in generating the truthfulness for solving the uncertainties of the knowledge acquired, by attaching a fuzzy membership to the knowledge acquired. The source of documents for the data collection for this research is very limited, as the knowledge passed from generation to another are done verbally by the communities. There are no proper or lack of documentations of the indigenous communities’ knowledge in Borneo. Therefore the data collection needs to be carried out by using the techniques of interview and questionnaire at the communities’ sites. However, most of the data sample will be collected from the domain expert in this field. Later, the knowledge acquired from the communities will be stored in the knowledge repository, such that it can be preserved for future references. On the other hand, the truth values and logical explanation generated from the designed framework which assigned to the knowledge can help the indigenous communities to appreciate their traditional belief system in relation to the current modern life style. This is particularly true when the traditional beliefs are mapped to the current logical explanation based on current practices. This whole process will be done together with the indigenous communities, and also the domain experts hoping that this technology could contribute to their daily life. With this framework, the implementation stage later will give a benefit of data collection and knowledge collection in the knowledge repository built. Later, the communities descendants will be able to refer to the repository through a user interface, which is also the prototype designed to retrieve and store knowledge and at the same time, having truth value and also logical explanation recommended attached to the knowledge as a reference. For example: If Ti Then {Ti }|-> {Ci |-> Li} and {Li |->  } Else {Ti }|-> { }.

mapping from Taboo, T n to Consequence, Cn, where ‘i’ and ‘n’ are finite positive integer. This representation will be done in the form of: (1) Kr:=Tn Cn After data mapping is done, the knowledge will be stored into a knowledge repository. This knowledge repository will enable the knowledge to be collected and retrieved through a user interface. In this stage, the relationships between a set of taboos, a set of beliefs and a set of consequences will be defined by using VDM++ formal specification language [17], which will help in identifying the missing or incomplete requirements in order to produce a complete, precise and consistent software model [18]. Later stage in the framework is the logical explanation and truth value generation. The knowledge stored in the knowledge repository will be retrieved and undergo a process of evaluation. Here, the fuzzy membership values (truth value) will be assigned to the knowledge acquired. The truth value extract will be obtained from a set of questionnaire in a number of samples, [1, N]. The truth value will be calculated according to the mean of the membership values assigned. For instance, take N = 30 samples from the indigenous communities, Xi, where i = [1, 30]. The truthfulness (membership values) will be assigned to the knowledge acquired. Typical architecture of the framework are based on Iban community, however the indigenous communities’ culture and beliefs are more or less similar [1]. Here, the average values for the membership values will be used to evaluate the truthfulness of the knowledge acquired.  =

∑ 

(2)



The value calculated for the knowledge acquired will be stored in a table such as table 1. Table 1: The summarization of the of the truth values for the knowledge acquired.

C1 C2 C3 Cn

T1  

     

T2  

     

T3  

     

Tn  

     

The example above indicates the rules for retrieving the data from the sets of knowledge stored in the repository. Where, when taboo is called, then, set of taboo which is related to Ti will be generated mapped to sets of Consequences of Ci where this Ci is mapped to the logical explanation. And at the same time, this logical explanation will mapped to the sets of membership values assigned. However, if the taboo is not found in the knowledge repository, thus, it will return an empty set.

In evaluation stage, knowledge repository and prototype built will be deployed at the indigenous communities’ sites. The feedback and set of responses will be collected through interviews from the communities’ leader, also known as the domain expert. The good values of the knowledge acquired will be identified, based on these feedbacks.

21

VI. CONCLUSION The main idea of this research is to come out with a better solution to solve the uncertainties of knowledge for the indigenous communities. Thus, a framework is designed by integrating the fuzzy logic and formal model. This framework inclusive three stages, data collection, data mapping and process of generating truth values and logical explanation stage. The data collection will be done together with the communities mainly the expert domain in this topic, which is the communities’ spiritual leader and also “tuai rumah” (long house leader). The framework designed will be implemented in later stage of the research and the result will be evaluated by the domain expert. The expected outcome from this research is a knowledge repository for indigenous communities that defined the relationships between taboos, beliefs and its consequences and also the good values of taboos which are defined based on the logical explanation/justification. The repository is designed is such a way so that it is expandable to accommodate taboos, beliefs and its good values from other indigenous communities. The application of formal-fuzzy logic models produce in this research not limited to taboos, beliefs and its consequences, but also can be used to represent the consequences of events in different domain such as the degree of consequences (or damage) of natural disasters (e.g., earthquake, flood, etc). The mapping model (e.g., disaster |-> damage) can be derived based on the historical records or a manipulation to a formula or input parameters.

[7]

[8] [9]

[10]

[11]

[12]

[13]

[14]

[15]

[16] [17] [18]

[19]

ACKNOWLEDGEMENT The authors would like to thank Universiti Malaysia Sarawak for providing the funding to publish and present this paper. This work is supported by Fundamental Research Grant Scheme: FRGS/02(30)/838/2012(78).

[20]

[21]

REFERENCES [1]

[2]

[3]

[4]

[5]

[6]

Mit, E., Borhan, N. H., and Khairuddin, M. A. (2012, 21-24 Oct. 2012). Need analysis of culture-based genealogy software for indigenous communities. Paper presented at the E-Learning, EManagement and E-Services (IS3e), 2012 IEEE Symposium. Achs, A. (2010, 10-11 Sept. 2010). Conclusion from uncertain knowledge. Paper presented at the Intelligent Systems and Informatics (SISY), 2010 8th International Symposium. Maher, P. (1999). A review of ‘traditional’ aboriginal health beliefs. Australian Journal of Rural Health, 7(4), 229-236. doi: 10.1046/j.1440-1584.1999.00264.x Winschiers-Theophilus, H., Rodil, K., Zaman, T., Yeo, A., and Jensen, K. (2013, 1-4 July 2013). Mobile technologies for preservation of indigenous knowledge in rural communities. Paper presented at the Information Technology in Asia (CITA), 2013 8th International Conference. Mit E., WaiShiang C., Asyraf M. and Hazlini N. (2011, November 29). Integrate cultures and beliefs into genealogy software for remote communities in Borneo. Paper presented at the IEEE 2011 International Conference on User Science and Engineering (i-USEr). doi: 10.1109/iUSEr.2011.6150570 Dutt, L. S., and Kurian, M. (2013). Handling of Uncertainty–A Survey. International Journal of Scientific and Research Publication 3(1), 4.

[22]

[23]

22

Chakraborty, S., (2012). Codd’s relational data model and fuzzy logic: a practical approach to find the computer solution. IJATER, Vol. 2(4), pp. 21-27. Zadeh, L.A. (1965). Fuzzy sets. Information and Control, Vol.8, pp.338-353. Yan, L., Zhang, F., and Ma, Z. M., (2012). f-SROIQ(G): An expressive fuzzy description logic supporting fuzzy data type group. SAC 2012, pp.320-325. doi: 10.1145/2245276.2245340 Zheng, S., Yiming, Z., and Martin, T. (2009, 15-18 Sept. 2009). A New Method for Fuzzy Formal Concept Analysis. Paper presented at the Web Intelligence and Intelligent Agent Technologies, 2009. WIIAT '09. IEEE/WIC/ACM International Joint Conferences on. Nguyen, T., Lu, H., and Lu, J. (2013). Web-Page Recommendation Based on Web Usage and Domain Knowledge. Knowledge and Data Engineering, IEEE Transactions on, PP(99), 1-1. doi: 10.1109/tkde.2013.78 Goncalves, M., Rodríguez, R. and Tineo, L. (2012). Formal method to implement fuzzy requirements. Dyna, 79(173) 15-24. Retrieved from: http://www.redalyc.org/articulo.oa?id=49623204003 Huang, J. and Fox M.S. (2005). Trust judgment in knowledge provenance. DEXA: 16th International workshop on Database and Expert Systems Applications. Copenhagen, Denmark. P.P. 524-528. Huang, J. and Fox M.S. (2004). Uncertainty in knowledge provenance. ESWS: Proceedings of the First European Semantic Web Symposium, C.J. Bussler, et. al., Editors. (2004), Springer: Heraklion, Crete, Greece. P.P. 372-387. Schlenoff, C., Foufou S., and Balakirsky S. (2012, March 20-22) Performance evaluation of robotic knowledge representation (PERK), Proceedings of the Workshop on Performance Metrics for Intelligent Systems, College Park, Maryland. The Overturetool Group (2006). Tool for Modelling VDM. Retrieved from: http://www.overturetool.org/, on November 11, 2013. CSK Corp. (2005). The VDM++ Language — Revised for V6.8.1. Abrial, J.-R. (2006). Formal methods in industry: achievements, problems, future. Paper presented at the Proceedings of the 28th international conference on Software engineering, Shanghai, China. Fitzgerald, J., and Larsen, P. (2007). Balancing Insight and Effort: The Industrial Uptake of Formal Methods. In C. Jones, Z. Liu and J. Woodcock (Eds.), Formal Methods and Hybrid Real-Time Systems (Vol. 4700, pp. 237-254): Springer Berlin Heidelberg. Rodika Tochi. (2012). What is Feng Shui, retrieved from:fengshui.about.com/od/thebasics/qt/fengshui.htm, on November 11, 2013. Bandemer, H. (1992). Fuzzy Local Inference in Fuzzy Knowledge Bases. In V. Novák, J. Ramík, M. Mareš, M. Černý and J. Nekola (Eds.), Fuzzy Approach to Reasoning and Decision-Making (Vol. 8, pp. 39-49): Springer Netherlands. Liang, L. R., Lu, S., Wang, X., Lu, Y., Mandal, V., Patacsil, D., and Kumar, D. (2006). FM-test: a fuzzy-set-theory-based approach to differential gene expression data analysis. BMC bioinformatics, 7(Suppl 4), S7. Běhounek, L., and Cintula, P. (2007). Features of mathematical theories in formal fuzzy logic. In Foundations of Fuzzy Logic and Soft Computing (pp. 523-532). Springer Berlin Heidelberg.