Above sections we have presented our proposed approaches: (1) content filtering based on keyword map and (2) collaborative filtering based on learner’s relationship. In this section, we explain our proposed approaches from educational aspects.
In chapter 3, we have described the learning environment and modeled individual learning process and group learning process. The design of our system is based on constructivist learning theories. In our system, the group learning is characterized by open ended generative tasks, collaborative decision making and problem solving. Moreover, instructional activities in the system which include such as creating stories, analogy making and etc. are based on learning strategies involving analogical reasoning, inquiry reasoning and etc.. Recommendations in this system act as guiders throughout the learning process to give learners guidance as needed and encourage learner independence in goal setting and decision making.
118 Chapter 5: Recommendation of Learning Resources
The design framework is as follows:
• Knowledge is acquired in context and builds on what is already know
• These facets such as learning process and social interaction aid knowledge retention, add interest, increase motivation and identify learners’ comparison
• Learners continually explore their knowledge, identify both their personal learning needs and strategies required to address them. This process helps to develop skills for constructivist learning
The inputs of recommendations include:
(1) Learning process is sequence of learning resources. Learning process reflect learners’ learning interests, intents, and experiences. In our view, learners’ learning process should not be separated from the knowledge base.
(2) Relationship inferred by social interactions represents a numerical account of the participation and learners’ situations in the collaborative and constructivist process.
The output is to identify learning issues to guide individual learning thus aid in reflective use of learning resources.
Based on above mentioned characters of our e-NOTEBOOK system, the purposed recommendation mechanisms as inquiry reasoning and collaborative decision making are used to guide the learner to more interesting or useful activities during learning. In other words, instead of allowing arbitrary searching and communication between learners on the web, the systems would be of great help if they were capable of suggesting learners or simply providing hints for their reflective use activity. In our e-NOTEBOOK system, group of learners take on roles, contribute ideas, critique each other’s work and solve aspects of problems together. A recommendation that creates opportunities for intentional reflection and provides the subject information and authentic problem can help to maximize the learning outcomes and knowledge transfer.
(1) Content filtering based on keyword map
Inquiry is the key part of constructivist learning. The mechanism of this approach is utilized to inquiry reasoning. System explains ratings in terms of informative features and explains features
Chapter 5: Recommendation of Learning Resources 119
in terms of examples based on properties. “The domain of learning”, types of instructional materials, and learners’ characteristics have been identified as features that can jointly contribute to the success of constructing knowledge. It mainly aids in inquiry reasoning based on transfer of learning. It assists to create stories and analogy making by using individuals’ learning processes.
The keyword map defined here is as a meta-level knowledge that indicates the structure of knowledge points in an inquiry reasoning concepts, and also learning process is the sequence in which these knowledge points has been delivered to learners. Their needs for resources change when the learning process changes, which calls for constructing new knowledge. On the other hand, the constructivist learning theory stresses the perspective that instruction needs to carefully consider learners’ prior knowledge and it emphasizes that human generate knowledge and meaning from their experiences. Social constructivism views each learner as a unique individual with unique needs and backgrounds, and it encourages the learner to arrive at his or her version of the truth, influenced by his or her background, culture or embedded worldview [wiki]. In this approaches, we use the module “learner’s keyword map based profile generator” to create stories based on transfer of learning to express individuals’ experiences and backgrounds. And analogy making is based on the module “Jaccard recommender module” to encourage the learner to arrive at his or her version of the truth.
(2) Collaborative filtering based on learner’s relationship
Social constructivism suggests that knowledge is first constructed in social context and is then appropriated by individuals [Bruning et al., 1999][Eggan and Kauchak, 2004]. The constructivist theory mentions that knowledge is actively constructed by each individual, and that social interactions with others are also influential in the constructive process [Brooks and Brooks, 1993]. In group, personal knowledge is correlated. Our approach is utilized to collaborative decision making based on collaborative learning processes. By using “estimate learners’
comparison module”, we can obtain the most advanced learner group. Learners’ comparison is important for knowledge sharing. In real world, when choosing educational learning resources, an individual’s decision is strongly influenced by advanced people with professional expertise in that field. Reflecting on learners’ experiences is via “inferring learner’s preference module”
based on implicit method. Finally, system use “learner’s relationship based collaborative
120 Chapter 5: Recommendation of Learning Resources
recommendation module” based on asymmetric interpersonal influence to make decisions.
Collaborative filtering based on learner’s relationship includes three facets which supports for brain storming, deep dives into topics and learners’ participation context to collaborative decision making process.
(i) Brain storming: Brainstorming is a group creativity technique designed to generate a large number of ideas for the solution of a problem. In proposed approach, the facet based on user dimension and item dimension belong to brain storming for enhancing either quantity or quality of ideas generated. According to [Osborn, 1963], there are four basic rules in brainstorming: focus on quantity, withhold criticism, welcome unusual ideas and combine and improve ideas. Here, brainstorming is individual brainstorming [Furnham and Yazdanpanahi, 1995] and includes such techniques as free writing, free speaking, word association, and drawing a mind map. Learners of group are peer collaborators. The rating inferred from learning process is learner’s contribution. Learners’ learning experience can be evaluated through direct data, such as their vote on the usefulness of a learning resources or it can be inferred from their learning processes, such as learning steps in learning a particular knowledge point. The preference should be closely connected with the learners’ learning processes.
(ii) Participations context: After interaction and brain storming, the ones who are interested in the same topic are formed as a segment. The motivation of such an arrangement is that a learner can learn and obtain useful information better from the advanced learners. [Vygotsky, 1978] proposes that a learner’s cognitive development is highly dependent on social interaction and collaboration with more capable and knowledgeable others. In zone of proximal development [Vygotsky, 1978], learners are challenged within close proximity to, yet slightly above, their current level of development. According to [Glasersfeld, 1989], sustaining motivation to learn is strongly dependent on the learner’s confidence in his or her potential for learning. Community can construct a model of education flexible enough for the way knowledge develops and changes today by producing a map of contextual knowledge. We can quantify the quality of the knowledge construction process in online collaborative learning through analysis of social interactions among the participants. Once the design is implemented
Chapter 5: Recommendation of Learning Resources 121
and the learning process starts, social relationships generate among participants. These social relationships control the learning outcomes. The relationship we proposed here takes into consideration both learners’ social interaction and their cognitive relations with the knowledge points in a group, so as to assess their level of interest in individual knowledge points, and the extent to which they grasp each knowledge point (e.g. their learning experience).
(iii) Deep dives into topic: On the other hand, according to [Scardamalia, 2002], learners are expected to make “constructive use of authoritative sources” such as books, websites, and experiments, treating them as potentially useful for informing their work. They are also expected to engage in progressive problem solving, reinvesting cognitive resources to deepen their understanding of problems and taking on more difficult problems over time [Bereiter and Scardamalia, 1993].
Learners collect information and data related to the interesting topic. The interaction of learners on the learning platform can influence the relatively inactive learners to trigger their interest which effective on some topics through the activities originated by their group participants. In section 4.2.2, we have defined some comparison indicators. These indicators imply that there is a community in which learners are joined together by mutual interests to intensively examine particular topics, and in so doing are able to learn together, exchange existing knowledge, and construct new knowledge. In this way, combining learners as a community enhances the acquisition of knowledge and understanding, and it satisfies the learning needs of its members. On the other hand, according to [Bednar et al., 1992] and [Black and McClintock, 1996], the constructivist epistemology focuses not only on learner individual inquiry or student-centered approach, but also on the cognitive apprenticeship provide by teachers or experienced peers. Thus, the advanced learners’ topics drive the activity of learners. This facet focuses on facilitating expert-like working with knowledge by guiding learners in a structured process of reflective use of learning resources.
We give the new definitions and impact of proposed approaches: content filtering based on keyword map and collaborative filtering based on learner’s relationship. Based on proposed mechanisms, system pushes learners toward useful resources for improving reflective use of learning resources.