コロキアムB発表

日時: 9月14日(火)3限(13:30~15:00)


会場: L1

司会: 大内 啓樹
安西 崇 M, 2回目発表 ソーシャル・コンピューティング 荒牧 英治, 中村 哲, 若宮 翔子
title: Naramachi Happy Map: Crowd-sourced Urban Atmosphere Estimation
abstract: Understanding an area's atmosphere is essential for making various geographical decisions. This kind of atmosphere (e.g., “beautiful,” “quiet,” ”happy,” “historical,” etc.) is formed not only by physical features such as scenery and functionality, but also by subjective information based on people's perceptions and experiences. Although scholars in various fields have attempted to quantitatively measure subjective human experiences, the amount of location-based subjective data varies depending on the area. Thus, we propose a method to quantify individuals' perception of urban atmospheres, as well as a means to anticipate these perceptions from landscape images. In this study, we focus on the ``historical'' atmosphere of old Japanese cities. First, we use crowdsourcing to assign atmospheric scores to the cities of Nara and Kyoto, the ancient capitals of Japan, based on how much history people perceive in the landscape images. Then, we train a deep learning model to estimate the atmospheric score of unknown landscape images using the previously labeled data. Finally, we discuss the utility of our method for routing applications and explore its generalizability across different cities.
language of the presentation: Japanese
 
伊藤 英里 M, 2回目発表 ソーシャル・コンピューティング 荒牧 英治, 中村 哲, 若宮 翔子, 田中 宏季
title: Extraction of expressions specific to nursing care by using nursing care records
abstract: Recently, in the healthcare industry, Real World Data (RWD), such as electronic medical records, nursing records, and medication diaries, which are generated through daily medical operations and personal health management, are getting widespread international attention. In Japan, the nursing care recording system are expected to be utilized for care, and we need to improve the efficiency of nursing care operations with declining birth rate and increasing population of the elder. In this study, we attempted to automatically extract physical condition expressions of patients from nursing care recording system written by nursing home staff.
language of the presentation: Japanese
 
廣田 一輝 M, 2回目発表 ソーシャル・コンピューティング 荒牧 英治, 中村 哲, 若宮 翔子
title: Temporal Analysis of Posting Content in Long-Term Twitter Users
abstract: Social media was born in the early 2000s and has become a tool used by everyone. Social media has become a tool that everyone uses, and some of them have been in operation for more than ten years, making them a valuable source of information about people's daily lives. However, there are few studies that focus on the changes of people from social media. Therefore, this study focuses on Twitter, one of the social media, and analyzes the content of posts by long-term Twitter users over time. In this study, we analyzed the similarity rate of postings among long-term Twitter users, which is expected to change with time. (1) the similarity rate of posted contents, (2) the similarity rate of users in each cluster, and (3) the similarity rate of posted contents. (2) Distance traveled by users in each cluster, (3) Meaning of words (3) Meaning of words (3) word meanings. For these three changes, we conducted experiments on the similarity rate of topics using a classification model, experiments on the distance traveled by users, and experiments on the semantic changes of words, and obtained some results suggesting changes in people. As a result of these experiments, we found several results suggesting that people are changing. For example, after a long period of time, the contents of users' postings become inconsistent about 60 percent of the time, about 70 percent of users show cluster transitions, and semantic changes in words such as ``corona'' are observed. In addition to reporting these results, this paper discusses the semantic changes in words obtained from the third experiment.
language of the presentation: Japanese
発表題目: 長期Twitterユーザにおける投稿内容の経時的分析
発表概要: Social mediaは2000年代初頭に生まれ,今では誰もが使用するツールとなった.また,中には10年以上の長期にわたって運営されているものもあり,人々の日常が蓄積された貴重な情報源である. しかしながら,実際にそのようなSocial mediaから人々の変化に焦点を当てて分析した研究はあまりない. したがって,本研究ではSocial mediaの1つであるTwitterを取り上げ,長期Twitterユーザにおける投稿内容の経時的分析を行った. また,本研究では時間と共に変化が見られると考えられるユーザの (1)投稿内容の類似率, (2)クラスタごとのユーザの移動距離, (3)単語の意味 に関する観点から変化を検討する. そして,これらの3つの変化に関して,分類モデルを用いた話題の類似率に関する実験,ユーザの移動距離に関する実験,単語の意味的変化に関する実験を行った結果,人々の変化を示唆する結果がいくつか得られた.それは,長期間経つとユーザの投稿内容は60\%程度一致しなくなり,70\%程度のユーザがクラスタの遷移をみせ,``コロナ''などをはじめとした単語に意味的変化が見られることである.本稿ではそれらについて報告することに加え,3つ目の実験により得られた意味的変化が見られた単語について考察する.
 
SUMAILA NIGO D, 中間発表 ソーシャル・コンピューティング 荒牧 英治, 中村 哲, 若宮 翔子
Sumaila Nigo

Gamified Participatory vector surveillance

Sumaila Nigo( 2021080 )


Emerging vector-borne diseases are a crucial issue in global health. Last two decades nearly a third of all recorded events related to emerging infectious diseases were vector-borne. Among these, mosquito-transmitted pathogens are a prime concern. vector surveillance is crucial component of vector control and can be used as a proxy for disease surveillance. Prevailing vector surveillance approaches are time consuming and expensive. Thus, this project introduces a complementary vector surveillance approach. This project proposesed an approach that aid communities to participate in vector surveillance efforts to lessen the burden of expert knowledge and costs. We implemented and tested in the field a crowdsourced vector surveillance approach through a gamified web app for mosquito-image collection. Our results show areas of high-density mosquito infestation, the geographical distribution, and a glimpse the vector behavior.