Development of a model to analyze the overall judgment of similarity and its evaluation

本研究首先发现，总体判断过程涉及到相似性判断、知觉过程、认知过程、语境效应、频率效应等多个方面。其次，在这些研究结果的基础上，提出了一个一般模型来分析的结构和那些方面涉及的总体判断。然后，提出了一种新的多维尺度方法，通过非线性优化方法估计模型参数。最后，通过蒙特卡洛研究，这表明该模型和缩放方法的实用性的算法进行了评估。这项研究是通过示范应用研究完成的。

In order to explore the complex aspects in the overall judgment of similarity or dissimilarity, this research aimed to develop a model and the scaling method and also to evaluate them.First it was found that the overall judgment process is related to many aspects such as (dis)similarity judgment, perceptual process, cognitive process, context effects, frequency effect. Next, on the basis of those findings, a general model was proposed to analyze the structure and those aspects involved in the overall judgment. A new method of multidimensional scaling was then developed to estimate the model parameters, by using nonlinear optimization methods. Finally, the algorithms were evaluated through a Monte Carlo study, which indicated practical applicability of the model and the scaling method. This research was accomplished by demonstrating application studies.