The importance of knowledge and cognition in business intelligence and decision support systems (DSS) is indisputable. However two major issues, a) biases in cognition, and b) knowledge integration overhead in knowledge warehousing, hinder their optimum utility in such systems. We address the issue of biases by proposing semantic de-biased associations (SDA) model, which is an improvement over the conventional causal map representation of mental models. SDA model incorporates semantics and contextual information to implement automated de-biasing by employing de-biasing techniques and algorithm into the inherent process of mental model elicitation, storage and retrieval. An elicitation process customised for SDA-based representation was also proposed namely SDA articulation and elicitation cycle. SDA model automates the process of mental model validation and integration, so as to prevent any espoused theories to be stored in the system. It also provides faster access to relevant knowledge, while creating a knowledge cycle between user and the system, which provides learning and knowledge growth opportunities to the system users, promoting organizational learning.
The issue of knowledge integration overhead is dealt with by proposing a unified, standard storage structure for knowledge warehousing in subject-oriented semantic knowledge warehouse (SSKW). The unified storage structure is achieved through categorising knowledge on syntactic level, and creating universal templates of these categories. In addition, the rules of how they can be connected together are outlined. The categories of knowledge, formalised, are object, process, and event. The connections between them are implemented through semantic relationships. The SSKW provides a domain-independent knowledge warehousing architecture to store knowledge in a subject-oriented, semantic, integrated, systematic and meaningful manner. It incorporates object-oriented, semantic, and human-centric approaches to facilitate an intuitive and efficient communication. It prevents loss of knowledge, improves precision of output, and ensures efficient delivery of knowledge when required.
The SDA model and SSKW are integrated together in this research to form a human-centric DSS, semantic-driven knowledge-enabled cognitive decision support system (SCDSS). SCDSS accumulates knowledge of many decision makers over time, thus if a decision maker leaves the organisation, his/her knowledge is retained through this system. Moreover, it automates the dissemination of knowledge across the organisation. Two evaluations were conducted to measure the performance of SCDSS against selected criteria. The results of the evaluations show that SCDSS successfully mitigates availability, framing, contextual and group biases, and generates new knowledge during decision making process. The results also demonstrate the effectiveness of SCDSS in knowledge sharing and enhancement, efficiency in producing output; and the relevance of knowledge in the output.
The system can be accessed at http://tasneememon.com/SCDSS/index.php.