The post-visit learning stage

The student can continue learning via the Internet at any place and time after leaving the museum. The learning service recommends additional content to the student based on his preference and learning records tracked by the system during his onsite visit. The recommendation includes extra exhibition content that interests the student, but which he has not yet appreciated during his onsite visit. Other related content in collection and education knowledge bases are also recommended to the student. The system tracks and analyzes the student’s learning behavior from his rating and navigation. Hence, the automatic recommendation service for the student’s next pre-visit plan or post-visit learning is close to his requirements.

4.3 Learning service model

A knowledge-based ubiquitous learning service model in a museum (see Figure 4.2) can be sketched out according to the learning scenario in Section 4.3, and modularized into three layers, namely the ontological knowledge base layer, adaptive application layer and ubiquitous learning service layer.

Figure 4.2 Knowledge-based mobile learning service model

4.3.1 The ubiquitous learning service layer

The ubiquitous learning service layer provides pre-visit, onsite-visit and post-visit learning via a single service portal. The pre-visit learning service provides an ontological interface to determine the learning content chosen by a user or recommended by the system.

The learner can arrange a visiting plan, and register it before visiting. The learner can download a registered learning plan to choose package learning or free learning when visiting the museum. Our implementation discards infrared, ultrasound, WiFi (Wireless Fidelity) and RFID (Radio Frequency Identification) solutions, due to their problems with operating efficiency, roaming and usability in exhibitions. Instead, a context-aware system based on ZigBee/IEEE 802.15.4 technology [48] was developed to serve learners by determining their demography, preferences, interests, locations, stay duration and visiting behavior. The

post-learning service allows learners to continue learning from more recommended content related to past onsite visits.

4.3.2 The adaptive service and content layer

The adaptive service and content layer provides several services for learners in the pre-visit, onsite-visit, and post-visit stages. The learning planning service provides an ontology-based interface allowing a learner to create a learning plan during the pre-visit stage.

The mobile learning service provides plan-learning, package-leaning and free-learning modes for a learner to proceed with his learning activity during the onsite visit. The service collaborates with the context-aware service to transmit requests to the content service during the onsite visit, and reminds the learner of his time spent and his current learning status. The context-aware service senses the learner’s temporal and spatial context, and notifies the content service to deliver the appropriate guide maps, learning content and related activity messages to the learner. The user tracking service tracks the learning behavior of a learner during the onsite visit to capture the preference information for the personalization service.

The learner’s learning activities are dynamically tracked and recorded to refresh his learning preference of each learner. The personalization service adaptively and proactively recommends learning content to the learner in every stage. The content service delivers relevant content from requests of the learning planning service, the mobile learning service and the personalization service.

4.3.3 Ontological knowledge base layer

The ontological knowledge base layer generates and manages learning content, and maintains user context information and learning records for all learners. This layer consists of three components: ontology, unified knowledge base and user context & usage base. The ontology provides common and sharable concepts to denote the unified knowledge base and the user context & usage base. The unified knowledge base comprises a multimedia database

an ontology-based concept hierarchy for content creation, management and publication. The user context and usage base includes user profiles, learning records, usage profiles and usage patterns. Section 4.5 describes in detail the unified knowledge base and the user context &

usage base.

4.4 Ontology-Based content and user context modeling

Ontology has been applied in the past to digital archives, digital museums and museum-related e-learning projects to provide shared and reusable knowledge standards from user and system perspectives. The HowNet approach [23] is adopted herein to build a unified natural and cultural ontology for NMNS (see Figure 4.3). This study adopts the unified classification hierarchy from our previous work, and extends knowledge concepts about exhibition and education topics in natural science to establish the ontology. This ontology plays several significant roles in this study. First, this ontology serves as a sharable thesaurus describing entities, attributes, relationships and events for the unified knowledge base and user context. Second, the ontology maps content onto a simplified and standard specification, and processes it consistently for the learning service. Third, the concept tree of the ontology can be naturally employed to design access interface and calculate the similarity of concept definitions from the user context and the unified knowledge base in the personalization service.

Figure 4.3 A unified natural and cultural ontology

4.4.1 Modeling of Knowledge-based Learning Content

A global knowledge system with an ontology-based concept hierarchy and relationships in a domain or between domains was previously built in the Chapter 3. In order to support a mobile learning service, we extend the conceptual modeling of enterprise domain knowledge system and multi-layer reusable knowledge content structures in section 3.6 by applying ontology to connect the knowledge entity and user context entity. Figure 4.4 illustrates the conceptualization of knowledge elements in the unified knowledge base and the user context

Figure 4.4 Ontological content and user context modeling

The multi-layer reusable content structures defined previously in Section 3.4 can express and organize many classes of content. An entity called knowledge element represents the superclass of all content, and consists of core knowledge elements, advanced knowledge elements and innovative knowledge elements. A core knowledge element is the basis of knowledge content, and consists of a multimedia object and its semantic metadata. An advanced knowledge element is further structured from a set of core knowledge elements, and can be a multimedia document, a knowledge unit, a knowledge group or a knowledge network.

A multimedia document integrates several core knowledge elements to describe an item related to an object of nature or culture (one species of bird or plant, or an artifact), or an exhibition and education topic. A knowledge unit possesses a hierarchical structure, and is

adopted to structure all related multimedia documents to interpret a particular object or topic.

All knowledge units can be categorized into three subclasses, namely the collection knowledge unit, exhibition knowledge unit and education knowledge unit.

A new entity, knowledge package, is defined to support the package-learning mode. This entity comprises a set of knowledge units with the same properties to present a research, education or exhibition topic. The knowledge package consists of two subclasses, the original knowledge group entity and a newly defined entity, knowledge chain. A knowledge group is a set of learning content items, and a knowledge chain denotes a sequence of knowledge units in a learning path. The knowledge chain entity can be applied to design learning packages for the package-learning mode. Specialists in content domains can define the cross-relationships between any pair of the above various elements, whether from within the same domain or different domains, to form a knowledge network.

Each knowledge element is converted to a content profile according to the shared ontology. All knowledge elements can thus be converted to content profiles with a consistent specification and process for the personalization service. Section 4.6 describes the process for creating content profiles.

4.4.2 User context and usage information modeling

Context-awareness is a key issue in the area of mobile and ubiquitous computing. Mobile users often require information depending on their current context, e.g. their location, their environment or available resources. For a mobile information system, several aspects of context can be considered, such as the characteristics of the particular mobile device (storage and screen size) and network (bandwidth and peers), context of the application (requirements in storage, download and display capability), context of the user of the system (e.g., time, location, interests), context of information objects (e.g., location). The handling of the concept

adapt the networking mode (to gain efficiency in the system communication) or to adapt the information display (to gain effectiveness in the user communication) [37].

People interact with a rich and stimulating environment like a museum for intellectual and aesthetic pleasure. This activity is not structured: adapting choices to the contingent situations, visitors move in the physical space guided by their interests or stimulated by the context. In order to avoid breakdowns in the flow of the activity, the boundary between the physical space and the information space should .be seamless

Information on the user context of a learner is vital to provide intelligent and proactive services to support context-aware and personalized mobile learning. Each registered user has a user context, which includes the user profile and learning records. As well as demographic information such as ID, age, education, sex, address and e-mail, a user profile also includes static and dynamic preferences and environment context. The static preference includes interesting domains specified by a user, and the dynamic preference is updated from the historical learning records tracked by the system. The environment context contains a learner’s location, time used and time left, which are identified and measured by the system.

The learning records contain the learning plan and historical learning records, and are denoted by a usage profile with a set of concept definitions aggregated from the content profiles. Usage patterns of individuals, groups and global learners are further discovered from usage profiles and used by the personalization service.

This work models user context and usage information using ontology. This method has two major advantages. First, the recommendation results based on content accessed can be assured to be accurate if the concepts are consistently represented in learners’ usage profiles as knowledge bases. Second, learners can easily inquire about concepts when they do not precisely understand topics or know the titles of all learning content items.

4.5 Ubiquitous personalization service

In this section, we aim to develop a personalization technique to support ubiquitous learning during pre-visit, onsite-visit, and post-visit stages. Several techniques applied in personalization service are the following: content-based filtering, collaborative filtering, rule-based filtering and Web usage mining. More advanced data mining methods and algorithms for use in the Web domain include association rules (a technique for finding frequent patterns, associations and correlations among sets of items), sequential pattern discovery (an extension of associated rule mining in that it reveals patterns of co-occurrence incorporating the notion of time sequence), clustering (used to group together pages or users that have similar characteristics) and classification (a process that maps items into various classes such as different types of user profile) to discover interesting and valuable patterns [45].

We adopt Dai’s [17] web usage mining approach to design a personalization service for ubiquitous mobile learning. This approach can be split into three phases, semantic preprocessing, pattern discovery and online recommendation. The first two phases can be processed off-line, while the third must be processed on-line. Figure 4.5 shows the personalization learning service model that we modify from Dai’s to fit personalization service for a knowledge-based digital museum in our study.

Figure 4.5 Personalization learning service model (adapted from Dai’s)

4.5.1 Preprocessing phase

The preprocessing phase consists of two stages, content preprocessing and usage preprocessing. In the content preprocessing stage, the system generates a content profile for each knowledge element. A content profile consists of content ID, knowledge class, type and a set of concept definitions. For instance, two exhibition topics “Flying reptiles” and “Digging dinosaur fossils” belong to the exhibition area entitled “The age of dinosaurs.” The “Flying reptiles” can be represented by dinosaur, reptile, extinguish, and fly while the “Digging dinosaur fossils” can represented by dinosaur, fossil, and dig.

The content profiles of the highest N-frequent concept items in a learning activity are aggregated in the usage preprocessing stage to form a usage profile. A usage profile includes the learner ID, learning activity ID and preferred concept definition. The dynamic preference is then updated according to the usage profile. For instance, if both “Flying reptiles” and

“Digging dinosaur fossils” appear in a learner’s learning activity, then the preference concept definition comprises dinosaur, reptile, fossil, extinguish, fly, and dig. The usage profiles of all learners can be aggregated to generate a usage profile matrix, which is utilized to discover useful patterns, as discussed in the next section.

4.5.2 Pattern discovery phase

The pattern discovery phase discovers useful usage patterns utilized by the online recommendation phase. One content pattern and three learner usage patterns, namely individual usage pattern, group usage pattern, and global usage pattern, can be discovered in this phase. In this study, these patterns were discovered using association mining algorithms [34] based on usage profiles.

In the individual usage pattern discovering process, an N-frequent itemset is determined from the concept definitions of each learner’s historical learning records to represent his preference. In the group pattern discovering process, all learners are divided into groups according to their demography and static preferences. N-frequent itemsets are obtained from the learning records of all learners in the same group. In the global pattern discovering process, all N-frequent itemsets are found according to all users’ usage profiles to represent a set of popular learning content groups among all learners.

In order to recommend and deliver content efficiently, all content items are grouped into clusters based on their similarity in their content profiles. Each cluster can be represented by a cluster content pattern with an aggregated concept definition from all content items in the cluster using clustering algorithms [34].

4.5.3 Online recommendation phase

The online recommendation phase provides an intelligent and proactive learning service during the pre-visit, onsite-visit, and post-visit stages. The recommendation service has two modes, user specification and system recommendation. The user specification mode makes

The system recommendation mode makes recommendations by matching the individual usage pattern (dynamic preference), group usage patterns and global usage patterns with cluster content patterns. The similarity between the usage patterns and the cluster content patterns is measured by calculating the distribution of concept definitions in the concept tree according to the shared ontology. The algorithm for measuring the similarity between a usage pattern and content pattern is shown below:

U1, U2 represent the preference concept definition list specified by a learner and concept definition list specified by content experts U1i，U2j are two nodes in the same concept tree

A(U1i,U2j) is U1i & U2j’s co-ancestor If U1i=U2j then

During the pre-visit stage, a learner can specify topics from the recommendation list, and arrange a plan to meet his requirements for both modes. The recommended content focuses on exhibition items and education activities that the learner plans to see. During the onsite-visit learning stage, the learner can choose the plan-learning mode by downloading a registered plan, or choose the package-learning or free-learning mode. The extended content is proposed dynamically, while the refreshed individual usage pattern is similar to cluster content patterns.

The recommended content focuses on collection knowledge units, exhibition units and messages of education activities according to the up-to-date individual usage pattern. After visiting the museum, the learner can obtain further recommendations from the system based on the last learning activity or matching individual, group or global usage pattern with the cluster content patterns. The recommendations are extended to relevant areas in collection, exhibition and education, and are categorized according to content type in the multi-layer reusable content structures defined in Section 4.5.1. Figure 4.6 shows an example of the personalization service during the pre-visit, onsite-visit and post-visit stages. Figure 4.7 shows the implementation system to fulfill this ubiquitous personalization service model.

Figure 4.6 Example of ubiquitous personalization service

Figure 4.7 The implementation system for ubiquitous personalization service

Chapter 5

Ontological techniques for reuse and sharing knowledge

5.1 Background and Purpose

In the previous two chapters, the KBDM framework was constructed through the UCKM model and the knowledge-based ubiquitous learning service model. A unified knowledge system is created by integrating knowledge bases from various individuals, domains, projects, and applications. A knowledge-based digital museum is created by reusing the unified knowledge content; these contents and applications are diffused to various users proactively, adaptively and ubiquitously based on common and sharable ontological knowledge concepts.

However, knowledge bases should be reused and shared as broadly as possible among communities and institutes using common knowledge acquisition and representation tools.

Furthermore, implicit and innovative knowledge should further be automatically discovered and generated based on these knowledge bases. Consequently, knowledge sharing and discovery services can be incorporate into the KBDM framework to achieve these objectives.

Like most digital museums, NMNS aims to construct a virtual museum for the public by utilizing various information technologies. Though the content can be manually represented via query or through metadata schemas or hyperlinks, this study argues that the digital archive represents a promising model for providing "knowledge." Restated, the usability of current NMNS only focuses on providing explicit and static information. The current systems thus are insufficient for supporting advanced knowledge engineering, for example knowledge inference processes. To make provision for the future knowledge era, NMNS has surveyed

pioneered project based on ontological techniques for restructuring current digital contents into corresponding knowledge bases. Three main tasks involved in this project are:

(1) Constructing a vascular plant ontology which represents a prototype of the knowledge base of NMNS,

(2) constructing an herbal drug ontology which pretends another knowledge base from outside communities, and

(3) developing a knowledge system for verifying the usability of correlative knowledge bases.

5.2 XML-based and Ontological techniques

Various studies have utilized KBSs in various fields, including expert systems, artificial intelligence, and decision support systems. Additionally, various approaches and tools have been used over the years to develop KBSs. However, many studies have noted challenges of sharing and reuse among KBSs [12, 31, 81]. Swartout [76] have concluded that KBSs are generally based on embedded systems and utilize their own special terminology system.

Consequently, traditional knowledge technology and engineering restrict interoperable abilities among KBSs. Recently; XML technology has been introduced for data exchange and integration purposes in various application areas. Moreover, the XML-based plus ontological technique is an emerging technology that has been considered as an important solution for addressing the problem of reuse and sharing in KBSs.

The ontologies have long been used to express a shared understanding of information by human beings. Gruber [30] defined ontology as "a specification of a conceptualization." A conceptualization is an abstract, simplified view of the world that is used for representational purposes. That is, the ontology is a formal description of the concepts, attributes, and

relationships involved in constructing common understanding for cognitions of real world events. A conceptualization is an abstract, simplified view of the world that is used for representational purposes. ‘Explicit’ means that the type of concepts used and the constraints on their use are explicitly defined. The ontology is a formal description of the concepts, attributes, and relationships involved and should be machine-readable. The ‘shared’ means ontology is used in constructing common understanding of a domain that can be

relationships involved in constructing common understanding for cognitions of real world events. A conceptualization is an abstract, simplified view of the world that is used for representational purposes. ‘Explicit’ means that the type of concepts used and the constraints on their use are explicitly defined. The ontology is a formal description of the concepts, attributes, and relationships involved and should be machine-readable. The ‘shared’ means ontology is used in constructing common understanding of a domain that can be