Tuesday, May 28, 2024
HomeData ScienceAnalyzing the Bias of Expression in the Three Great Poetry Collections|Data Science...

Analyzing the Bias of Expression in the Three Great Poetry Collections|Data Science Accelerates Classical Research|Mr. Ikegami, Musashino University

In recent years, an increasing number of companies and educational institutions are adopting a learning method called PBL (problem-based learning) . The Musashino University Faculty of Data Science (MUDS) is one of them.

PBL at MUDS greatly contributes to student growth. At the 13th Forum on Data Engineering and Information Management (DEIM2021) held in March 2021, four MUDS students received the Student Presentation Award.

DEIM is an academic conference held for the purpose of discussing and exchanging opinions on research themes related to data engineering and information management.

How did MUDS students present their research results and receive awards at competitions attended by students, graduate students, and researchers from other universities?

This time, we interviewed Aiwa Ikegami, who won the Student Presentation Award at DEIM2021, about her research achievements and the secret behind how MUDS students are evaluated.

Table of Contents

  • Research that combines “classics” and “data science”
  • Changes in “Aesthetic Concepts” as Seen from Research on the Three Great Poetry Collections
  • Behind the scenes of MUDS-style PBL
    • Why are MUDS students evaluated?
    • Scrutinize through trial and error
  • Future prospects
  • lastly

Research that combines “classics” and “data science”

ーーIkegami-san, what kind of theme did you research?

Mr. Ikegami: I worked on research on the theme of waka poetry under the title ” Discovery of changes in aesthetic ideas through co-occurring word changes in poetry in three major anthologies .”

I was interested in psychology and wanted to study human sensibilities. Since human thoughts, knowledge, and emotions reside in “words,” I thought that by analyzing words, we could understand changes in human sensibilities and thoughts .

Taking the example of waka poems about flowers, I thought that by analyzing chronological changes in what states were perceived as beautiful, it would be possible to read the changes in aesthetic conception over time from waka poems. I worked on my research.

It was very interesting to see the changes, because the words and grammar used in ancient languages ​​differed depending on the era.

ーーPlease tell us how you came to conduct research on the theme of “classics x data science”.

Mr. Ikegami: I think that ” classics x data science ” is a field with less previous research than other fields, so I consulted with the professor about what we can do in this field. Also, I myself have been studying Japanese history and classics, so I started by thinking about what I could do with that knowledge and what I wanted to study.

While studying linguistics, I learned that there is a way of thinking that different languages ​​may lead to different views of the world . For example, in Japanese, “butterfly” and “moth” are distinguished, but in French they are called “papillon” without distinction.

Applying that idea to the classics, I thought that there might be some difference between the world we see in modern languages ​​and the world expressed in ancient languages . With this thought in mind, I established a research policy to interpret the changes in thinking and aesthetic concepts from differences in language.

Changes in “Aesthetic Concepts” as Seen from Research on the Three Great Poetry Collections

ーーWhat kind of research results did you obtain?

Mr. Ikegami: I carried out research focusing on the “poetry”, which is the core of waka poetry, through two experiments.

The first is the ratio of poetry used in the three major anthologies .

Many poems about “geography” and “time” were commonly found in all the anthologies, but in the “Manyoshu” there were many poems about “state”, “food, clothing and shelter”, and “animals and plants” in other anthologies. It turns out that there are more than Compared to other anthologies, the “Kokin Wakashu” uses more poetic words related to “heart” than other anthologies, and the “Shin Kokin Wakashu” uses many poems related to “heavenly phenomena” and “heart.” In the “Kokin Wakashu”, there were many poems related to “heart”, but in the “Shin Kokin Wakashu” there were fewer.

From these facts, it is thought that there were many realistic waka poems in the Manyoshu that expressed the unvarnished “state” of beautiful things and falling flowers .

Since the Kokin Wakashu contains many poems related to the “heart,” it is thought that it was an era when it became important to reflect the heart in waka.

In “Shin Kokin Wakashu”, the number of poems related to “heart” is decreasing, so it is thought that the poem changed to use euphemistic expressions instead of directly expressing the heart .

Mr. Ikegami: The second is the ” transition of co-occurring words in the three major anthologies of poetry .”

In this experiment, interesting results were obtained regarding the transition of co-occurrence words in “Haru”. The word “hal” co-occurred with “saku” in the Manyoshu, but it co-occurred with “chiru” in the Kokin Wakashu and Shin Kokin Wakashu.

From this result, I think it can be said that the aesthetic concept has changed greatly from “a state in which flowers are in full bloom” to “a state in which flowers are falling .”

Behind the scenes of MUDS-style PBL

Why are MUDS students evaluated?

ーーWhat do you think is the reason why 4 MUDS students won awards at DEIM2021 this time?

Mr. Ikegami: I think it is because there are many opportunities to present research results . In the MUDS class, I feel that there are quite a lot of opportunities to make presentations in front of people and create materials. One of the characteristics of MUDS is that there is a lot of group work, but an even bigger characteristic is the large number of opportunities for presentations .

Also, regarding the creation of materials, the more presentations you make, the more sophisticated presentation materials you will be able to create.

Since I have many opportunities to present my research results, I have an environment where I can learn from my seniors and people around me, such as how to create materials and how to communicate. I feel that this is one of the factors that led to the success of DEIM2021.

Scrutinize through trial and error

ーーCould you tell us about the flow of the “Mirai Creation Project” from identifying issues to compiling research results?

Mr. Ikegami: You start in earnest in the second semester of your first year, and first discuss your research theme with your professor. Early learners decide on specific themes and policies in a week or two, or a month at the latest, and then start learning technical matters.

I reported on the progress every week, and had the opportunity to consult with the professor about the reasons for selecting the theme, research methods, and expected results. I received feedback from professors and seniors, so I think it was a very important opportunity for first- and second-year students to feel the importance of vertical connections .

ーーWere there any difficulties or challenges in your research?

Mr. Ikegami: I had no programming experience, so there were technical issues. It was also difficult to balance lectures and assignments at university. Since the lectures were held online, I was unfamiliar with them, and at times I was worried about what to do if my research hit a dead end, and how to consult with them.

However, in the Future Creation Project, I was able to share my progress with the professor and listen to other students talk, so I was able to advance my research while receiving good stimulation.

ーーPlease tell us what you devised and what you were conscious of when summarizing the results.

Mr. Ikegami: What I was most conscious of was trying to find out how to convey waka poems in an easy-to-understand way to people who don’t have much knowledge about waka at the information-related academic conference “DEIM2021 . ”

Since the theme is waka poetry, which has not had much previous research, I thought it was necessary to carefully explain waka poems and the language of poetry. Therefore, since most of MUDS students are science students, I asked them if they could understand my talk and research.

Future prospects

ーーIkegami-san, please tell us what you would like to research in the future and what your future prospects are.

Mr. Ikegami: The research that was awarded at DEIM2021 focused on waka poetry. In the future, I would like to realize an analysis visualization tool that will increase the interest of more people in classical literature .

I have always liked Japanese history, so I would like to draw out new hypotheses and knowledge about historical backgrounds by deriving correlations from classical literary materials and Japanese history documents other than waka poems.

In addition, I believe that by conducting research on the background of the authors of unknown or anonymous works from the perspective of data science, we can make a significant contribution to historical and literary research .

Many works of classical literature have unknown authors. I believe that it is possible to make new discoveries about the author’s background by using features such as the words used, grammar, and content .


By repeating trial and error, MUDS students improve the accuracy of their research methods and the way they communicate their results. It is appreciated not only within the university, but also at a large academic conference called DEIM. Because MUDS is close to seniors and professors, and there are people you can rely on, I think you can output and receive feedback, which will lead to your growth.

Mr. Ikegami’s research area is a difficult area of ​​”words and expressions” with few precedents, but future research can be expected to have a major impact on research in Japanese literature. We will also pay attention to the future trends of MUDS.



Please enter your comment!
Please enter your name here

Recent Posts

Most Popular

Recent Comments