研究者詳細 - 石川　佳治

論文 - 石川　佳治

分割表示 320 件中 61 - 80 件目／全件表示 >>

Finding Probabilistic Nearest Neighbors for Query Objects with Imprecise Locations 査読有り

Yuichi Iijima and Yoshiharu Ishikawa

Proceedings of the 10th International Conference on Mobile Data Management (MDM 2009) 頁： 52-61 2009年5月

記述言語：英語

A nearest neighbor query is an important notion in spatial databases and moving object databases. In the emerging application fields of moving object technologies, such as mobile sensors and mobile robotics, the location of an object is often imprecise due to noise and estimation errors. We propose techniques for processing a nearest neighbor query when the location of the query object is specified by an imprecise Gaussian distribution. First, we consider two query processing strategies for pruning candidate objects, which can reduce the number of objects that require numerical integration for computing the qualification probabilities. In addition, we consider a hybrid approach that combines the two strategies. The performance of the proposed methods is evaluated using test data.

トレーサブルなP2P レコード交換システムにおけるイベント駆動型問合せ査読有り

飯田卓也，李峰栄，石川佳治

日本データベース学会論文誌 8 巻 ( 1 ) 頁： 95-100 2009年6月

記述言語：日本語掲載種別：研究論文（学術雑誌）

我々の研究グループでは，情報の信頼性を確保するため，P2P
ネットワーク上でトレーサブルな情報交換をおこなうシステム
PI-REX を開発している．本稿では本システムにおけるイベント駆動型問合せについて述べる．イベント駆動型の問合せの導入により，ピアの行動監視，更新情報の監視などをネットワークに負荷をかけることなくおこなうことが可能となる．システムのイベント駆動型問合せの機能，および実装手法についても論じる．

Effective Top-k Keyword Search in Relational Databases Considering Query Semantics 査読有り

Yanwei Xu, Yoshiharu Ishikawa, and Jihong Guan

Proceedings of International Workshop on DataBase and Information Retrieval and Aspects in Evaluating Holistic Quality of Ontology-Based Information Retrieval (DBIR-ENQOIR 2009) (APWeb-WAIM 2009 Workshop) 頁： 172-184 2009年9月

記述言語：英語

Keyword search in relational databases has recently emerged
as a new research topic. As a search result is often assembled from multiple relational tables, existing IR-style ranking strategies can not be applied directly. In this paper, we propose a novel IR ranking strategy considering query semantics for eective keyword search. The experimental results on a large-scale real database emonstrate that our method results in signicant improvement in terms of retrieval eectiveness as compared to previous ranking strategies.

Skyline Queries Based on User Locations and Preferences for Making Location-Based Recommendations 査読有り

Kazuki Kodama, Yuichi Iijima, Xi Guo, Yoshiharu Ishikawa

Proceedings of 2009 International Workshop on Location Based Social Networks (LBSN 2009) 頁： 6-13 2009年11月

記述言語：英語

Due to the recent development of mobile computing and communication network technologies, information services for mobile phone users and car navigation systems have becomeof
some importance. Since these mobile devices have limited display sizes, we often need to select carefully the appropriate information to be presented to the user. However,
it is not easy to select the "appropriate" information because users have different contexts and preferences.
In this paper, we present an approach to recommending items such as restaurants to a mobile user taking into account his current location and preferences. In our framework,
a user initially provides a profile, which records preferences as relative orders within predefined categories such as food types and prices. We then select items to be recommended
from the database based on the user's profile as well as the current location. To select good items, we extend the notion of spatial skyline queries to incorporate not
only distance information but also categorical preference information.
Based on the proposed approach, a prototype system has been implemented in a small mobile PC containing a small embedded RDBMS. The facilities of the RDBMS, such as
spatial indexes, were used to process our skyline queries effectively.

データベース（特集：ロボットを進化させる最先端IT技術）

石川佳治，喜連川優

日本ロボット学会誌 28 巻 ( 3 ) 頁： 36-39 2010年3月

Efficient Continuous Top-k Keyword Search in Relational Databases 査読有り

Yanwei Xu, Yoshiharu Ishikawa, and Jihong Guan

Proceedings of 11th International Conference on Web-Age Information Management (WAIM 2010) 頁： 755-767 2010年4月

記述言語：英語

Keyword search in relational databases has been widely studied in recent years. Most of the previous studies focus on how to answer an instant keyword query. In this paper, we focus on how to find the top-k answers in relational databases for continuous keyword queries efficiently. As answering a keyword query involves a large number of join operations between relations, reevaluating the keyword query when the database is updated is rather expensive. We propose a method to compute a range for the future relevance score of query answers. For each keyword query, our method computes a state of the query evaluation process, which only contains a small amount of data and can be used to maintain top-k answers when the database is continually growing. The experimental results show that our method can be used to solve the problem of responding to continuous keyword searches for a relational database that is updated frequently.

Query Processing with Materialized Views in a Traceable P2P Record Exchange Framework 査読有り

Fengrong Li and Yoshiharu Ishikawa

Proceedings of WAIM 2010 International Workshops 頁： 246-257 2010年4月

記述言語：英語

Materialized views which are derived from base relations and stored in the database are often used to speed up query processing. In this paper, we leverage them in a traceable peer-to-peer (P2P) record exchange framework which was proposed to ensure reliability among the exchanged data in P2P networks where duplicates and modifications of data occur independently in autonomous peers. In our proposed framework, the provenance/lineage of the exchanged data can be available by issuing tracing queries. Processing for tracing queries was based on the "pay-as-you-go" approach. The framework can achieve low maintenance cost since each peer only maintains minimum amount of information for tracing. However, the user must pay relatively high query processing cost when he or she issues a query. We consider that the use of materialized views allows more efficient query execution plans. In this paper, we focus on how to incorporate query processing based on materialized views in our framework.

Query Processing in a Traceable P2P Record Exchange Framework 査読有り

Fengrong Li and Yoshiharu Ishikawa

IEICE Transactions on Information and Systems E93-D 巻 ( 6 ) 頁： 1433-1446 2010年6月

あいまいな位置情報に基づく最近傍問合せの処理手法査読有り

飯島裕一，石川佳治

電子情報通信学会論文誌 J93-D 巻 ( 6 ) 頁： 781-794 2010年6月

記述言語：日本語掲載種別：研究論文（学術雑誌）

移動ロボットやモバイルセンサネットワークなどの位置情報を利用したアプリケーションでは，最近傍問合せは重要な問合せとして位置づけられている．しかし，制御ノイズや測定誤差などの理由により問合せを行うオブジェクトの位置があいまいな位置情報としてしか得ることができない場合が存在する．そのため，位置のあいまい性を考慮した問合せ処理手法が必要とされている．そこで本論文では，問合せを行うオブジェクトの位置が正規分布の確率密度関数によってあいまいな位置情報として表現されている状況における最近傍問合せの処理手法を提案する．まず通常の最近傍問合せを拡張した確率的最近傍問合せを定義し，この問合せを効率的に処理するための戦略として二つの問合せ戦略を提案する．どちらの戦略も，問合せの対象となるオブジェクトの集合から明らかに問合せを満たさないといえるものを求めることで計算コストの削減を図るが，その方法が異なっている．これら二つの戦略にそれらのハイブリッド方式の戦略を加えた三つの戦略に対して，実験によって性能の比較を行った結果についても報告する．

Direction-Based Spatial Skylines 査読有り

Xi Guo, Yoshiharu Ishikawa, and Yunjun Gao

Proceedings of Ninth International ACM Workshop on Data Engineering for Wireless and Mobile Access (MobiDE 2010) 頁： 73-80 2010年6月

記述言語：英語

Traditional location-based services recommend nearest objects to the user by considering their spatial proximity. However, an object not only has its distance but also has its direction which originates from the user to it. In this paper, we study direction-based spatial skyline queries (DSS queries) which retrieve nearest objects around the user from different directions. The closer object is better than or dominates the further object if they are in the same direction. The objects that cannot be dominated by any other object are included in the direction-based spatial skyline (DSS). We propose algorithms to answer snapshot queries which find objects on the DSS according to the user's current position. We also develop algorithms to support continuous queries which retrieve objects on the DSS while the user is moving linearly. Extensive experiments verify the performance of our proposed algorithms using both real and synthetic datasets.

Query Processing with Materialized Views in a Traceable P2P Record Exchange Framework 査読有り

Fengrong Li and Yoshiharu Ishikawa

Proceedings of WAIM 2010 International Workshops 頁： 246-257 2010年7月

記述言語：英語

DOI： 10.1007/978-3-642-16720-1_25

Efficient Continuous Top-k Keyword Search in Relational Databases 査読有り

Yanwei Xu, Yoshiharu Ishikawa, Jihong Guan

Proceedings of 11th International Conference on Web-Age Information Management (WAIM 2010) 頁： 755-767 2010年7月

記述言語：英語

DOI： 10.1007/978-3-642-14246-8_71

Anonymizing User Location and Profile Information for Privacy-aware Mobile Services 査読有り

Masanori Mano and Yoshiharu Ishikawa

Proceedings of 2nd ACM SIGSPATIAL International Workshop on Location Based Social Networks (LBSN 2010) 頁： 68-75 2010年11月

記述言語：英語

Due to the growing use of mobile devices, location-based services have become popular. A location service often requires the user's exact location to provide appropriate services and this brings the risk of threats to privacy. In this paper, we propose an anonymization method for users of location-based services in mobile environments.

The anonymization approach is based on the well-known k-anonymity concept, but has additional features. We consider the situation that a mobile service (e.g., mobile advertisement) utilizes mobile users' profiles for its service. Since a profile contains privacy information such as the age and address of the user, the use of profile information brings another kind of privacy threat.

The anonymization method proposed in this paper considers not only location information but also privacy-related attributes in the user's profile. The location anonymizer, a trusted third-party placed between users and mobile application services, anonymizes the location and profile attributes based on the request. We define a similarity measure between mobile users for anonymization purposes. The similarity is used for related users in terms of their locations and profile attributes. We present the concept behind our method and the anonymization algorithm, and then show some experimental results.

Using Materialized Views to Enhance a Traceable P2P Record Exchange Framework 査読有り

Fengrong Li, Yoshiharu Ishikawa

Journal of Advances in Information Technology 2 巻 ( 1 ) 頁： 27-39 2011年2月

記述言語：英語掲載種別：研究論文（学術雑誌）

P2P technologies are drawing increasing attention nowadays, and have been widely deployed on the Internet for various purposes. Unlike the traditional client-server architecture, a P2P network allows all computers to communicate and share resources as equals and does not depend on a central server for control. In such an environment, tracing how data is copied between peers and how data modifications are performed are not easy because data replications and modifications are performed independently by autonomous peers. This creates inconsistencies in exchanged information and results in a lack of trustworthiness. To provide reliable and flexible information exchange facility in P2P networks, we have proposed a framework for enabling traceable record exchange. In this framework, a computer can exchange structured records with a predefined schema with other peers. The framework supports a tracing facility to query the lineage of the records obtained. A tracing query is described in Datalog and executed as a recursive query among cooperating peers in a P2P network. In the query execution process, the exchange and modification histories of the queried records are collected dynamically from relevant peers.
In this paper, we focus on how to enhance the traceable P2P record exchange framework using materialized views. First, we discuss how to construct materialized views in our framework. Then we present methods for reducing query processing cost and providing fault tolerance using the materialized views.

部分シーケンスマッチングのためのストリームアルゴリズム査読有り

豊田真智子，櫻井保志，石川佳治

電子情報通信学会論文誌D J94-D 巻 ( 7 ) 頁： 1058-1070 2011年7月

記述言語：日本語掲載種別：研究論文（学術雑誌）

本論文では，データストリームにおけるcross-similarityの問題を定義する．本論文の目的は，データストリームから類似する部分シーケンスペアを検出することである．シーケンス間の類似度を測定する距離尺度には，時間軸上でのスケーリングを考慮することができるダイナミックタイムワーピング（DTW：Dynamic Time Warping）距離を利用する．我々の提案するCrossMatchは，厳密にDTWに基づいた手法であり，データストリーム処理に適したワンパスアルゴリズムである．DTWを用いた純粋なアルゴリズムと比べて，CrossMatchは計算コストとメモリ使用量の大幅な低減化を実現する．理論的な分析を行い，提案アルゴリズムが精度を犠牲にすることなく類似する部分シーケンスペアを検出することを示す．また，実データと人工データを用いた実験から，CrossMatchがインクリメンタルにcross-similarityを検出することが確認された．

Multi-Objective Optimal Combination Queries 査読有り

Xi Guo and Yoshiharu Ishikawa

Proceedings of 22nd International Conference on Database and Expert Systems (DEXA 2011), Part I 頁： 47-61 2011年8月

記述言語：英語

Multi-objective optimization problem finds out optimal objects w.r.t. several objectives rather than a single objective. We propose a new problem called a multi-objective optimal combination problem (MOC problem) which finds out object combinations w.r.t. multiple objectives. A combination dominates another combination if it is not worse than anther one in all attributes and better than another one in one attribute at least. The combinations, which cannot be dominated by any other combinations, are optimal. We propose an efficient algorithm to find out optimal combinations by reducing the search space with a lower bound reduction method and an upper bound reduction method based on the R-tree index. We implemented the proposed algorithm and conducted experiments on synthetic data sets.

DOI： 10.1007/978-3-642-23088-2_4

An Index Structure for Spatial Range Querying on Gaussian Distributions 査読有り

Kazuki Kodama, Tingting Dong, Yoshiharu Ishikawa

Proceedings of the Fifth International Workshop on Management of Uncertain Data (MUD 2011) 頁： 1-7 2011年8月

記述言語：英語

In the research area of spatial databases, query processing based on uncertain location information has become an important research topic. In this paper, we propose an index structure for the case that the locations of a query object and target objects are imprecise and specified by Gaussian distributions with different parameters. The index structure efficiently supports probabilistic spatial range queries, which is an enhanced version of traditional spatial range queries, by considering the properties of Gaussian distributions. We implement the proposed index structure using GiST, a generalized index structure, and we evaluate its performance based on the experiments.

Simulation Based Analysis for a Traceable P2P Record Exchange Framework 査読有り

Fengrong Li, Yoshiharu Ishikawa

Proceedings of 4th International Conference on Data Management in Grid and P2P Systems (Globe 2011) 頁： 49-60 2011年9月

記述言語：英語

P2P technologies are getting more and more attention lately. However, unlike the traditional client-server architecture, a P2P network allows all computers to communicate and share resources as equals without central server control. This causes inconsistency in exchanged information and results in lack of trustworthiness. To provide trustful and flexible information exchange facility in P2P networks, we proposed a traceable P2P record exchange framework. In this framework, a peer can exchange structured records with a predefined schema among other peers. The framework supports a tracing facility to query the lineage of the obtained records based on database technologies. A tracing query is described in Datalog and executed as a recursive query among cooperating peers in a P2P network. In this paper, we focus on analyzing and verifying the traceable P2P record exchange framework based on simulation experiments in three different example P2P networks.

DOI： 10.1007/978-3-642-22947-3_5

Direction-Based Surrounder Queries for Mobile Recommendations 査読有り

Xi Guo, Baihua Zheng, Yoshiharu Ishikawa, Yunjun Gao

The VLDB Journal 20 巻 ( 5 ) 頁： 743-766 2011年10月

記述言語：英語掲載種別：研究論文（学術雑誌）

Location-based recommendation services recommend objects to the user based on the user's preferences. In general, the nearest objects are good choices considering their spatial proximity to the user. However, not only the distance of an object to the user but also their directional relationship are important. Motivated by these, we propose a new spatial query, namely a direction-based surrounder (DBS) query, which retrieves the nearest objects around the user from different directions. We define the DBS query not only in a two-dimensional Euclidean space E but also in a road network R. In the Euclidean space E, we consider two objects a and b are directional close w.r.t. a query point q iff the included angle aqb is bounded by a threshold specified by the user at the query time. In a road network R, we consider two objects a and b are directional close iff their shortest paths to q overlap. We say object a dominates object b iff they are directional close and meanwhile a is closer to q than b. All the objects that are not dominated by others based on the above dominance relationship constitute direction-based surrounders (DBSs). In this paper, we formalize the DBS query, study it in both the snapshot and continuous settings, and conduct extensive experiments with both real and synthetic datasets to evaluate our proposed algorithms. The experimental results demonstrate that the proposed algorithms can answer DBS queries efficiently.

DOI： 10.1007/s00778-011-0241-y

Efficient Continual Top-k Keyword Search in Relational Databases 査読有り

Yanwei Xu, Yoshiharu Ishikawa, and Jihong Guan

Journal of Information Processing (IPSJ) 20 巻 ( 1 ) 頁： 114-127 2012年1月

記述言語：英語掲載種別：研究論文（学術雑誌）

Keyword search in relational databases has been widely studied in recent years because it requires users neither to master a certain structured query language nor to know the complex underlying database schemas. Most existing methods focus on answering snapshot keyword queries in static databases. In practice, however, databases are updated frequently, and users may have long-term interests on specific topics. To deal with such situations, it is necessary to build effective and efficient facilities in a database system to support continual keyword queries. In this paper, we propose an efficient method for answering continual keyword queries over relational databases. The proposed method consists of two core algorithms. The first one computes a set of potential top-k results by evaluating the range of the future relevance score for every query result and creates a light-weight state for each keyword query. The second one uses these states to maintain the top-k results of keyword queries while the database is continually being updated. Experimental results validate the effectiveness and efficiency of the proposed method.

DOI： 10.2197/ipsjjip.20.114

　前のページ - 次のページ　

<石川 佳治>

論文 - 石川 佳治

<石川　佳治>

論文 - 石川　佳治