Abstract
After applying discrete wavelets multi-resolution analysis to literary works, it is verified that this method considerably removes the analyzed results from the dependency of analyst. Nevertheless, it is difficult to remove the choice and collection of analysis data depending on personal experience exactly. By introducing a general idea of Orthnormal Analysis, duplicated information in analysis data gets objectivity and each data gets same weight to the work. This paper applies Orthonormal Linear Space System Analysis to the selected data comprising the vectors of "The Tale of Genji", after that multi-resolution analysis of discrete wavelets is applied to these vectors. The term vector is built by three keywords of feeling "Inclination", "Disinclination" and "Adoration". As a result, it is revealed that the story is constituted by "Inclination" and "Adoration", and the visualized curve of "Adoration" has common part to "Inclination" in every level in one episode in "The tale of Genji "Utsusemi story". Further, it is visualized that the same tendency of conventional opinion that the hero's emotion fluctuates due to triflingness of the heartless attitude of the heroin and loss feeling of youth of hero. On the other hand, it is extracted the characteristic more in detail.
