Faculty Profiles - KAZAMA Kazuhiro

Misc - KAZAMA Kazuhiro

Division display41 - 60 of about 79／ All the affair displays >>

Information Diffusion on Twitter(<Special Issue>Twitter and Social Media)

KAZAMA Kazuhiro, Kazuhiro Kazama, NTT Network Innovation Laboratories Nippon Telegraph and Telephone Corporation

人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence ( 人工知能学会 ) 27 ( 1 ) 35 - 42 2012.01

Time-Series Variation Analysis of Relationships between Terms Using Tweets in East Japan Earthquake

風間一洋, 鳥海不二夫, 榊剛史

人工知能学会全国大会論文集 ( 人工知能学会 ) 26 1 - 4 2012

Space-Time Analysis of Geographical Search Targets using Query Logs

風間一洋

研究報告データベースシステム（DBS） 2011 ( 9 ) 1 - 7 2011.07

地理情報検索において，ユーザの地理的検索対象が時間と共にどのように変動しているかについて調べる．そこで，商用サーチエンジンのクエリログから，東京の首都圏の各地名と共起語を抽出し，頻度の時間的・空間的な変動を分析した．その結果，地理的クエリは一般的なクエリとは異なる時間的変動を示し，現実の人間の行動の影響があること，地名はそれぞれ固有の時間的変動パターンを持ち，共起語集合がその地域の特徴を示すことがわかった．さらに，地図上に特に頻繁に検索されている地名を可視化して．注目されている地名の分散や移動パターンに，現実の人間の行動との関連性があることを示した．I examined the spatial and time variation of users' geographical search targets in geographical searches. I extracted geographical names of Tokyo metropolitan district and their cooccurrence terms from commercial search engine query logs and analyzed the spatial and time variation of their frequency. I found that the variation of geographical queries was different from one of general queries and it was influenced by human activities in the real world. Each geographical name had its own characteristics and its cooccurrence terms showed regional features. Furthermore, I displayed frequently-searched geographical names on the map. I showed that their distribution and moving patterns were related to real human movement.

Analysis of Relativity between Press Releases and Twitter Comments concerning Fukushima Daiichi Nuclear Power Plants Disaster

Shibata Yugo, Kimura Hiroshi, Sakaki Takeshi, Toriumi Fujio, Kazama Kazuhiro

Proceedings of Annual / Fall Meetings of Atomic Energy Society of Japan ( Atomic Energy Society of Japan ) 2011 ( 0 ) 12 - 12 2011

2011年5月現在いまだ進行中である福島第一原子力発電所の事故を受けて、政府や事業者等の各種機関からは原子力発電所について様々なプレスリリースが出され続けている。プレスリリースは報道機関や一般の人々に対する情報提供を目的として発信される資料であるが、今日、人々は情報を受け取るだけでなく、原子力発電所事故に関する自身の考えをTwitter上で発信し続けている。本研究では原子力発電所事故に関するプレスリリースとTwitter上における発言との関連性を分析し、原子力事故発生時にプレスリリースが人々にどのような影響を及ぼしたのかについて考察することを目的とする。

DOI

Distribution Pattern Analysis of Associated Geographical Names on Transportation Networks

風間一洋

人工知能学会全国大会論文集 ( 人工知能学会 ) 25 1 - 4 2011

Sensing Network : Mining Real World Information from the Web

KAZAMA Kazuhiro, KURIHARA Satoshi

IPSJ Magazine ( Information Processing Society of Japan (IPSJ) ) 51 ( 9 ) 1171 - 1180 2010.09

Extraction of Overlapping Community Structures from Human Relationships

風間一洋, 佐藤進也, 斉藤和巳

日本ソフトウェア科学会大会論文集 ( [日本ソフトウェア科学会] ) 27 10p 2010.09

Multi-faceted Ranking of Information Sources Using Information Diffusion Properties in Blogspace

風間一洋, 今田美幸, 柏木啓一郎

情報処理学会論文誌データベース（TOD） ( 情報処理学会 ) 3 ( 2 ) 102 - 110 2010.06

本論文では，3 種類の情報伝播特性を用いた，ブロガが注目している情報源の多面的ランキング手法を提案する．ブログ検索結果から抽出した情報伝播ネットワークを各情報源から到達可能な複数の部分ネットワークに分割し，部分ネットワーク中の 3 種類の有向 2 エッジ連結部分グラフの数から定義した情報拡散度，情報集約度および情報転送度を情報源のランキングに用いる．さらに，実際の情報伝播ネットワークに適用し，ランキング結果の順位相関係数を調べるとともに，オフィシャルサイト，ニュース，CGM という 3 種類の情報源の MRR と MAP を用いてランキング手法の特性を明らかにし，情報伝播特性を用いた多面的ランキングの有効性を示す．This paper proposes a multi-faceted ranking method of information sources attracting bloggers by using three information diffusion properties. The information diffusion network extracted from searched blog entries is divided into subnetworks, each of which is reachable from an information source. We define the information scatter degree, the information gather degree, and information transmit degree as information diffusion properties by using the number of three types of directed 2-edge connected subgraphs in a subnetwork and use them for the ranking of information sources. We apply this ranking method to real information diffusion networks and analyze rank-order correlation coefficient of ranking results. Furthermore, we characterize ranking methods by using MRR and MAP of three-types of information sources: official sites, news sites and CGMs. In the result, we demonstrate the effectiveness of a multi-faceted ranking method by using information diffusion properties.

Twitter の情報伝播ネットワークの分析

風間一洋

第24回人工知能学会全国大会, 2010 2010

Quantification of Information Diffusion Characteristics in Blogspace

風間一洋, 今田美幸, 柏木啓一郎

論文集 ( 人工知能学会 ) 23 1 - 4 2009

A Propagation Feature of Personal Information on Social Networking Services

IMADA Miyuki, KAZAMA Kazuhiro

IEICE technical report ( The Institute of Electronics, Information and Communication Engineers ) 108 ( 289 ) 31 - 36 2008.11

A important privacy problem of future ubiquitous services lies on network aria that is possible to combine private information from ubiquitous devices with private information from network services such as weblog and social networking services (SNS). In this paper, we propose a model of privacy violation using anonymity quantitative method based information entropy on network space. This model focuses on privacy violation brought about results from personal data combinations. The private information propagation is took account of network models. Then we construct a simulation model using multi-agent simulator and discuss the effect of privacy violation that aspects on private information propagation on the network aria and actual aria.

A Detection Method of Privacy Violation on Social Networking Services

IMADA Miyuki, KAZAMA Kazuhiro

IEICE technical report ( The Institute of Electronics, Information and Communication Engineers ) 108 ( 258 ) 71 - 76 2008.10

In this paper, we propose a detection method of privacy violation using information entropy on social networking services (SNS). This method prevents personal identification by private information combination on each SNS domain. The basic function of the method quantitatively evaluates the anonymity level of a person who discloses his/her data, and controls the personal-data disclosure according to the level. Then we evaluate the method from the aspect of private information violation using multi-agent simulator, and we show the effectiveness and clarify the applied aria.

ブログ空間における情報伝播ネットワークの抽出と分析

風間一洋

Web とデータベースに関するフォーラム, 2008 2008

ブログサービスにおける読者ネットワークの分析. ブログサービスにおける読者ネットワークの分析

風間一洋

ネットワークが創発する知能研究会第2回国内ワークショップ(JWEIN06) 72 - 79 2006

ブログ空間の主要トピック抽出 (特集「Web Intelligence」および一般)

斉藤和巳, 木村昌弘, 風間一洋

知識ベ-スシステム研究会 ( 人工知能学会 ) 70 5 - 10 2005.08

Web community discovery using personal names

K Kazama, S Sato, K Fukuda, H Kawakami, O Katai

INTERNATIONAL WORKSHOP ON CHALLENGES IN WEB INFORMATION RETRIEVAL AND INTEGRATION, PROCEEDINGS ( IEEE COMPUTER SOC ) 2005 113 - 121 2005

These days, web access is essential for our daily lives, and many communities are formed by the exchange of information. We propose a method for discovering web communities by using personal names. It extracts personal names from search results that are relevant to a specific topic and finds appropriate web communities by analyzing the bipartite graph between personal names and websites. In general, web pages are too noisy to extract such social network structures automatically. We introduce a measure called "effectiveness" and extract only key people or key websites to make their bipartite graph. We also demonstrate an application called "Community Navigator" and visualize web communities for a specific topic by using it. Moreover, we analyze the measure for an actual example and show that it is suitable for selecting key people and key websites.

DOI

Web Community Analysis by Personal Names

KAZAMA KAZUHIRO, HARADA MASANORI, SATO SHIN-YA, FUKUDA KENSUKE, KAWAKAMI KOJI, KATAI OSAMU

IPSJ SIG Notes. ICS ( Information Processing Society of Japan (IPSJ) ) 2004 ( 85 ) 131 - 138 2004.08

With the rapid growth of the web, huge and various documents are opened to public on the web and it is essential for your life and business to search desired information by web search engines. Nevertheless user often fails to search desired information which surely exist on the web because of too many information. So new search assistance methods are desired. In this paper, we try to assist user's search process by using a real-world oriented searching which combines the web space and the real world. We uses personal names for this purpose. We assume that there is a relation between a person and a web site in the case that his name appears in the web site. And we try to extract communities which belong to the same topic by analyzing the bipartite graph between personal names and web sites in consideration of their effectiveness and visualize them.

人名を用いたWeb空間のコミュニティの解析 (特集:「ネットワークが創発する知能」および一般)

風間一洋, 原田昌紀, 佐藤進也

知識ベ-スシステム研究会 ( 人工知能学会 ) 65 127 - 134 2004.08

Figuring out People by their Workspaces on the Web

SATO Shin-ya, HARADA Masanori, KAZAMA Kazuhiro

IPSJ SIG Notes ( Information Processing Society of Japan (IPSJ) ) 2004 ( 45 ) 75 - 82 2004.05

Although the Web is an artificial space for information utilization, many of the pages on the Web are mentioning things or concepts in the real world. As the Web is going to play more important role in our daily lives, it will more precisely reflect the real world. In consideration of these facts and the trend, an approach for analyzing the Web that makes good use of knowledge on the real world, so to say "real world oriented Web mining", is considered to be promising. Based on this idea, this paper presents a method to extract workspaces for persons with a given name on the web. The method is derived by applying knowledge on the relationship between people and their workspaces in the real world to the Web. The validity of the method is verified through experiments. A set of workspaces for a person extracted by the method from the web is expected to be put to use for characterizing the person. As an example, the paper introduce a technique to visualize relationships between the workspaces, which can be used for roughly distinguishing between persons with the same family and personal name.

Development and operations of the distributed system for the Ja - Jakarta Project

KOYAMA Hiroshi, KAZAMA Kazuhiro, OKAMOTO Takashi, YOKOTA Takehiko

情報処理学会研究報告インターネットと運用技術（IOT） ( Information Processing Society of Japan (IPSJ) ) 2004 ( 46 ) 17 - 22 2004.05

The Ja-Jakarta project is a volunteer project to promote the japanization of open source products, and to translate their documents into Japanese. We have developed a distributed system constructed mainly with open source products and operates them to offer a cooperating work environment to the contributors scattered all over the country. We also offer a translation support service based on XML. We evaluated our system design and operation management by the analysis of the information in project mailing lists, system logs, and the qestionnaire to our contributors. As the results, we confirmed that our services such as information exchange, information disclosure, and translation support have been effectively utilized. This paper thus reports that the open source model is effective for translation works.

　 PREV - NEXT 　

<KAZAMA Kazuhiro>

Misc - KAZAMA Kazuhiro