公开(公告)号：CZ20001552A3

公开(公告)日：2000-08-16

申请号：CZ20001552

申请日：1998-10-27

Applicant: IBM

Inventor： CASTELLI VITTORIO ,  LI CHUNG-SHENG ,  THAMASIAN ALEXANDER

IPC: G06F20060101 ,  G06F17/30 ,  G06K9/62

Abstract: An improved multidimensional data indexing technique that generates compact indexes such that most or all of the index can reside in main memory at any time. During the clustering and dimensionality reduction, clustering information and dimensionality reduction information are generated for use in a subsequent search phase. The indexing technique can be effective even in the presence of variables which are not highly correlated. Other features provide for efficiently performing exact and nearest neighbor searches using the clustering information and dimensionality reduction information. One example of the dimensionality reduction uses a singular value decomposition technique. The method can also be recursively applied to each of the reduced-dimensionality clusters. The dimensionality reduction can also be applied to the entire database as a first step of the index generation.

公开(公告)号：CZ297222B6

公开(公告)日：2006-10-11

申请号：CZ20001552

申请日：1998-10-27

Applicant: IBM

Inventor： CASTELLI VITTORIO ,  LI CHUNG-SHENG ,  THAMASIAN ALEXANDER

IPC: G06F17/30 ,  G06F20060101 ,  G06K9/62

Abstract: Predlozený vynález se týká vylepseného zpusobu indexování vícerozmerných dat, který vytvárí kompaktní indexy takovým zpusobem, ze vsechny nebo vetsina indexu muze být rezidentní v hlavní pameti v libovolném okamziku. Behem shlukování a snízení pocturozmeru jsou vytvoreny informace o shlukování (111) a informace o snízení poctu rozmeru (112), které jsou vyuzity v následné vyhledávací fázi. Zpusobindexování muze být efektivní i za prítomnosti promenných, které nejsou vysoce korelovány. Dalsí nástroje slouzí pro efektivní provádení exaktního vyhledávání a vyhledávání nejblizsího souseda s vyuzitím informace o shlukování (111) a informace o snízení poctu rozmeru (112). Jeden príklad snízení poctu rozmeru vyuzívá postupu dekompozice singulární hodnoty. Zpusob muze být rovnez aplikován rekurzivne na kazdý ze shluku o snízeném poctu rozmeru. Snízení poctu rozmeru muze být rovnez aplikováno na celou databázi jako prvotní krok pri vytvárení indexu.

公开(公告)号：DE69802960D1

公开(公告)日：2002-01-24

申请号：DE69802960

申请日：1998-10-27

Applicant: IBM

Inventor： CASTELLI VITTORIO ,  LI CHUNG-SHENG ,  THAMASIAN ALEXANDER

IPC: G06F20060101 ,  G06F17/30 ,  G06K9/62

Abstract: An improved multidimensional data indexing technique that generates compact indexes such that most or all of the index can reside in main memory at any time. During the clustering and dimensionality reduction, clustering information and dimensionality reduction information are generated for use in a subsequent search phase. The indexing technique can be effective even in the presence of variables which are not highly correlated. Other features provide for efficiently performing exact and nearest neighbor searches using the clustering information and dimensionality reduction information. One example of the dimensionality reduction uses a singular value decomposition technique. The method can also be recursively applied to each of the reduced-dimensionality clusters. The dimensionality reduction can also be applied to the entire database as a first step of the index generation.

公开(公告)号：DE69802960T2

公开(公告)日：2002-08-29

申请号：DE69802960

申请日：1998-10-27

Applicant: IBM

Inventor： CASTELLI VITTORIO ,  LI CHUNG-SHENG ,  THAMASIAN ALEXANDER

IPC: G06F20060101 ,  G06F17/30 ,  G06K9/62

Abstract: An improved multidimensional data indexing technique that generates compact indexes such that most or all of the index can reside in main memory at any time. During the clustering and dimensionality reduction, clustering information and dimensionality reduction information are generated for use in a subsequent search phase. The indexing technique can be effective even in the presence of variables which are not highly correlated. Other features provide for efficiently performing exact and nearest neighbor searches using the clustering information and dimensionality reduction information. One example of the dimensionality reduction uses a singular value decomposition technique. The method can also be recursively applied to each of the reduced-dimensionality clusters. The dimensionality reduction can also be applied to the entire database as a first step of the index generation.

公开(公告)号：HU0100581A2

公开(公告)日：2001-06-28

申请号：HU0100581

申请日：1998-10-27

Applicant: IBM

Inventor： CASTELLI VITTORIO ,  LI CHUNG-SHENG ,  THAMASIAN ALEXANDER

IPC: G06F20060101 ,  G06F17/30 ,  G06K9/62

Abstract: An improved multidimensional data indexing technique that generates compact indexes such that most or all of the index can reside in main memory at any time. During the clustering and dimensionality reduction, clustering information and dimensionality reduction information are generated for use in a subsequent search phase. The indexing technique can be effective even in the presence of variables which are not highly correlated. Other features provide for efficiently performing exact and nearest neighbor searches using the clustering information and dimensionality reduction information. One example of the dimensionality reduction uses a singular value decomposition technique. The method can also be recursively applied to each of the reduced-dimensionality clusters. The dimensionality reduction can also be applied to the entire database as a first step of the index generation.