Web-Based Evaluation System to Measure Learning Effectiveness in Kampo Medicine

Figure 2

MSs who scored higher than the mean + 1SD () and those who scored lower than the mean − 1SD () were defined as high and low achievers, respectively. Values on the vertical and horizontal axes were computed arbitrarily using a nonlinear mapping tool (see [8] and [9]). Briefly, we have a set of medical students (MSs) in the 8-dimensional space corresponding to the correct answer rate of the 8-type fields of KM. Initially we choose at random a set of MSs in the 2-dimensional space. The set is the initial configuration of the 2-dimensional space. Next we compute all distances between the MSs in the 2-dimensional space. The next step in the algorithm is to adjust the vectors of MSs in the 2-dimensional space so that the configuration of MSs in the 2-dimensional space can well fit that of the MSs in the 8-dimensional space. This is achieved by carrying out a steepest descent procedure. As a result, we get the optimal configuration of MSs in the 2-dimensional space in this figure. In Sammon’s algorithm, the configuration of MSs is essential. Hence, vertical axis and horizontal axis in the figure have no physical means. High (blue) and low (red) achievers classified according to cluster analysis using correct answer rates for 8-type fields of KM.