Measurement And Empirical Study Of Knowledge Representation Agreement Of Groups

In the process of information communication and delivery between two none-face-to-face groups, whether and to what extent the knowledge representations of an identical object or concept could reach an agreement among these two groups will substantially influence the performance of information communication and further the performance of cooperation between groups. During the information distribution process of unconventional events, whether there is agreement of knowledge representation between mainstream media and the public would have tremendous impact on public mental states and behaviors following unconventional events. During the scenario of product search in e-commerce websites, whether buyers and sellers could reach an agreement on the representation of product, would largely decide whether they could find the targeted opposites.Under this research context, this paper proposes a measurement of knowledge representation agreement within and across groups based on theory of cognitive categorization and theory of knowledge element. Meanwhile Information Entropy and Euclid Distance are chosen as benchmark criterions to validate measurements of within-group agreement and across-group agreement accordingly, which prove to be validated measurements. Based on the above, the proposed measure is applied to two classic information distribution and matching scenarios. Through empirical studies the agreement magnitudes are evaluated with proposed measurements. Further, potential factors which might have effect on the agreement within and across groups under each scenario are discussed and analyzed with details.This research of measurement of knowledge representation agreement within and across groups might contribute both theoretically and practically. Theoretically, the proposed within-group agreement measurement, oriented from domain of cognitive psychology and knowledge representation, reflects the agreement magnitude within groups with three constructs:size of feature pool, peak value of feature agreement rate and rate of features with over50%agreement rate. While the proposed three measurements are testified to have equal validity with Information Entropy, they could make up the singleness and complexity of Information Entropy, and at the same make it easier to incorporate with cognitive psychology research. The proposed across-group agreement measurement, oriented from domain of cognitive psychology and knowledge representation, reflects the agreement magnitude across groups by measuring the agreement of one group towards the other. While the proposed measurement is testified to have equal validity with Euclid Distance, they could make up the high dimension and wasted computation space used by Euclid Distance Practically, the proposed measurement could evaluate the agreement magnitude of knowledge representation of an identical object or concept between two none-face-to-face groups, thus offering suggestions and considerations as for improving the ways of knowledge representation, information communication and analyzing the potential factors that might lead to disagreement.