Knowledge management systems (KMSs) are important and useful tools for an enterprise to manage business knowledge effectively. In addition, KMSs provide adequate support for improving decision making and gaining competitive advantage.
In an organization, workers are assigned to carry out various tasks, and workers need to apply their experience and professional knowledge to complete the tasks to achieve organizational goal [12][13]. For this reason, an organization has the urgency to make use of the knowledge assets which are emerging from organizational operations and management activities to increase its profitability and productivity with the support of KMSs. For KMSs, Information Technology (IT) plays the role of facilitating the access, reuse and sharing of knowledge assets within and across an organization to assist knowledge workers in executing their tasks [9][17]. In addition, workers need to access lots of textual documents in conducting knowledge-intensive tasks. Accordingly, Information Filtering (IF) techniques are often employed to model users’ information needs as profiles and provide relevant information based on the modeled profiles.
Effective knowledge management relies on understanding workers’ information needs on tasks, for brevity, task-needs. As the operations and management activities of enterprises are mainly task-based, KMSs focus on providing task-relevant knowledge to workers engaged in knowledge-intensive tasks [1][2][11][12][13]. The Kabiria system supports knowledge-based document retrieval in office environments, allowing users to conduct document retrieval according to the operational context of task-associated procedures [8]. The KnowMore system maintains task specifications (profiles) to enumerate the process-context of tasks and associated knowledge items [1]. Context-aware delivery of task-specific knowledge can then be facilitated based on the task specifications and current execution context of the process. The above works provide an appropriate perspective for designing task-based knowledge support.
However, they focus on specifying the process-context of a task to support context-aware or process-aware knowledge retrieval, rather than on a systematic method for constructing a task profile that models a worker’s task needs. Moreover, very few researches address the issue of profile adaptation to track workers’ dynamic information needs.
This research proposes a novel adaptive task-profiling technique to model worker’s information needs on tasks, i.e., task-needs, as follows.
First, a task profile specifies the key concept terms of a worker’s current task (task at hand), and models the information needs of the worker during the task’s execution. The proposed technique adjusts task profiles to model workers’ dynamic task-needs based on the documents accessed by workers without considering user feedback (positive or negative opinion).
Second, in order to model users' task needs, a proper task-based topic taxonomy is used to conceptualize the domain information of organizational activities. Note that the main subjects of organizational activities defined by domain experts and previously (executed) representative tasks form the task-based topic taxonomy. The topics and their corresponding topic profiles are used as references to adjust task profiles according to their relevance (similarity) to the documents accessed by the workers.
Third, generally, the more recent the document accessed the more important it is to reflect a work’s current task needs. Thus, the effect of time factor is considered in profile adaptation.
Fourth, the proposed profiling technique adopts a novel collaborative profile adaptation approach to adjust task profiles. Knowledge workers usually require a substantial amount of time to accomplish knowledge-intensive tasks. For such long-term tasks, the information needs of the workers may vary according to their progress during the performance of tasks. For example, a graduate student is seeking adequate knowledge documents for her research. Her research topics may vary as the following: "Event detection" => "Mining event change" => "Mining Patent change"
=> "Patent Mining", where the symbol '=>' denotes that the left-hand side occurs before the right-hand side. Without explicitly specifying the change of information needs directly by workers, we try to capture the variations of workers’ information needs through those documents accessed by the workers. This work uses the variation of workers’ topic needs to model the variation of the workers’ information needs.
Conventional user profiling approaches, which are based on relevance feedback on documents, can only reflect the information needs accumulated, and lack consideration of possible change of information needs. This work measures the variation of workers’ topic needs according to their knowledge activities (e.g. access documents). Workers with similar variations of topic needs over time are identified. A novel collaborative profile adaptation approach is proposed to adjust task profiles via
using similar workers’ variations of topic needs to predict the target worker’s future possible variations of topic needs.
The proposed approach enhances knowledge retrieval through collaboration from similar workers. The codified knowledge that is relevant to the current task can be retrieved based on the adjusted task profile to fit the worker’s dynamic task needs.
Empirical experiments demonstrate that the proposed approach models workers’
task-needs effectively and helps provide task-relevant knowledge.
The rest of this thesis is organized as follows. Section 2 surveys the related work.
The overview of proposed methodology is described in Section 3. The approach of task needs modeling is detailed in Section 4. The details of measuring and maintaining variation of topic needs over time are presented in Section 5. Section 6 details how to identify similar workers with similar variations of topic needs and the proposed collaborative profile adaptation approach. The experiments of our proposed approaches are illustrated in Section 7. The conclusion and future works are concluded in Section 8.