Presentation is loading. Please wait.

Presentation is loading. Please wait.

Efficient Type-Ahead Search on Relational Data: a TASTIER Approach Guoliang Li 1, Shengyue Ji 2, Chen Li 2, Jianhua Feng 1 1 Tsinghua University, Beijing,

Published by Modified over 9 years ago

Similar presentations

Presentation on theme: "Efficient Type-Ahead Search on Relational Data: a TASTIER Approach Guoliang Li 1, Shengyue Ji 2, Chen Li 2, Jianhua Feng 1 1 Tsinghua University, Beijing,"— Presentation transcript:

1 Efficient Type-Ahead Search on Relational Data: a TASTIER Approach Guoliang Li 1, Shengyue Ji 2, Chen Li 2, Jianhua Feng 1 1 Tsinghua University, Beijing, China 2 University of California, Irvine, CA, USA

2 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Traditional Keyword Search MUST Type in Complete keywords

3 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Type-Ahead Search Advantages:  Interactive: data exploration in relational databases  Full-text search: full-text search on-the-fly

4 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Challenges and Preliminaries  Efficiency requirement (milliseconds vs. seconds)  Client-side processing  Network delay  Server-side processing  Opportunities:  Subsequent queries can be answered incrementally

5 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Fundamentals  Data  R: a relational database with a set of tables  D: a set of distinct words tokenized from the data in R

6 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Fundamentals  Query  Q = {p 1, p 2, …, p l }: a set of prefixes  Query result  R Q : a set of subtrees (called Steiner trees) such that each subtree has all query prefixes, i.e., a set of relevant tuples connected through foreign keys such that each answer has all query prefixes (conjunctive)

7 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Traditional Keyword Search  Data Graph  database  search  sigmod  sigir  signature  Query: {database search sigmod}  Answers: Steiner trees(radius  r) a2a3 a5 a2a3 a5

8 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Type-Ahead Search  Data Graph  database  search  sigmod  sigir  signature  Query: {database search sig}  Answer: Steiner trees(radius  r) a2a3 a5 a2a3 a5

9 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Type-Ahead Search in Relational Data  Step 1  Incremental prefix matching  Step 2  Incrementally find relevant connected tuples that contain query prefixes  Contributions  Efficiently Finding answers using  -step forward index  Improving search efficiency graph partition query prediction

10 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Step 1: Incremental Prefix Matching  Example  D = {sigmod, search, spark, yu, graph}  Q = “graph s”  W s ={sigmod, search, spark}  Q’ = “graph sig”  W sig ={sigmod}

11 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Tire Index Graph

12 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Incremental Prefix Matching  sigmod, search, spark, yu, graph graph searchsigmodspark s

13 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Step 2: Finding answers  graph  How to efficiently find answers? yu Graph Yu

14 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Contributions  Step 1  Incremental prefix matching  Step 2  Efficiently Finding answers using  -step forward index  Improving search efficiency graph partition query prediction

15 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng  -step forward index Graph Yu Search

16 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Finding answers using  -step forward index Yu s

17 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Finding answers using  -step forward index p Yu s

18 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Contributions  Step 1  Incremental prefix matching  Step 2  Efficiently Finding answers using  -step forward index  Improving search efficiency graph partition query prediction

19 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Graph Partition  Step 1  Find subgraphs that contain query prefixes  Step 2  Find answers within subgraphs Graph

20 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Graph Partition  Q= “Graph Yu”  Step 1: find subgraphs S 2, S 3  Step 2: find answers within S 2, S 3

21 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng High-Quality Graph Partition  A: S 1,S 2  B: S 1,S 2  C: S 1,S 2 S1S1 S2S2 S3S3 S4S4  D: S 1,S 2  E: S 1,S 2  F: S 1,S 2  A: S 3  B: S 4  C: S 3  D: S 4  E: S 3,S 4  F: S 3,S 4 Advantages: 1.Shorten List 2.Subgraph Pruning

22 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Keyword-Sensitive Partition  Graph  Hypergraph  G(V, E)  G h (V h,E h ) V h =V if (u,v)  E, then (u,v)  E h, if u 1, u 2, …, u n contain a same keyword, then (u 1, u 2, …, u n )  E h  Hypergraph Partition B

23 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Contributions  Step 1  Incremental prefix matching  Step 2  Efficiently Finding answers using  -step forward index  improving search efficiency graph partition query prediction

24 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Query Prediction

25 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Previous Method vs. Query Prediction  Previous method Find all potential compute words of query prefixes and compute corresponding answers e.g., {sigmod, sigir, signature, …,} for sig  Query prediction Predict the complete keywords with maximal probabilities and compute corresponding answers using the predicted keywords E.g., predict 2 best keyword {sigmod, sigir} for sig

26 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Query Prediction  Query-prediction model  Bayesin network  Pr(k i ) = #of occurrences of k i / # of nodes  Pr(k i |k j, k n ) = Pr(k i |k n )

27 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Query Prediction  Q=“keyword s”keyword search  Q=“keyword search r”keyword search relation

28 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Experimental Results  Setting  C++, Gnu compiler, FastCGI,  Ubuntu, X5450 3.0GHz CPU, 3GB RAM  Datasets  DBLP  IMDB

29 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Search Efficiency

30 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Scalability: Index Size

31 Tsinghua & UC Irvine Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Efficient Type-Ahead Search on Relational Data Guoliang Li, Shengyue Ji, Chen Li, Jianhua Feng Scalability: Search Time

32 Questions ? Thank You! Questions? http://tastier.ics.uci.edu/ http://tastier.cs.tsinghua.edu.cn/ Search: tastier type-ahead search

Download ppt "Efficient Type-Ahead Search on Relational Data: a TASTIER Approach Guoliang Li 1, Shengyue Ji 2, Chen Li 2, Jianhua Feng 1 1 Tsinghua University, Beijing,"

Similar presentations

Efficient Interactive Fuzzy Keyword Search Shengyue Ji 1, Guoliang Li 2, Chen Li 1, Jianhua Feng 2 1 University of California, Irvine 2 Tsinghua University.

Efficient Interactive Fuzzy Keyword Search Shengyue Ji 1, Guoliang Li 2, Chen Li 1, Jianhua Feng 2 1 University of California, Irvine 2 Tsinghua University.
Chen Li ( 李晨 ) Chen Li Scalable Interactive Search NFIC August 14, 2010, San Jose, CA Joint work with colleagues at UC Irvine and Tsinghua University.

Chen Li ( 李晨 ) Chen Li Scalable Interactive Search NFIC August 14, 2010, San Jose, CA Joint work with colleagues at UC Irvine and Tsinghua University.
Computer Science and Engineering Inverted Linear Quadtree: Efficient Top K Spatial Keyword Search Chengyuan Zhang 1,Ying Zhang 1,Wenjie Zhang 1, Xuemin.

Computer Science and Engineering Inverted Linear Quadtree: Efficient Top K Spatial Keyword Search Chengyuan Zhang 1,Ying Zhang 1,Wenjie Zhang 1, Xuemin.
BY ANISH D. SARMA, XIN DONG, ALON HALEVY, PROCEEDINGS OF SIGMOD'08, VANCOUVER, BRITISH COLUMBIA, CANADA, JUNE 2008 Bootstrapping Pay-As-You-Go Data Integration.

BY ANISH D. SARMA, XIN DONG, ALON HALEVY, PROCEEDINGS OF SIGMOD'08, VANCOUVER, BRITISH COLUMBIA, CANADA, JUNE 2008 Bootstrapping Pay-As-You-Go Data Integration.
Jianxin Li, Chengfei Liu, Rui Zhou Swinburne University of Technology, Australia Wei Wang University of New South Wales, Australia Top-k Keyword Search.

Jianxin Li, Chengfei Liu, Rui Zhou Swinburne University of Technology, Australia Wei Wang University of New South Wales, Australia Top-k Keyword Search.
1 gStore: Answering SPARQL Queries Via Subgraph Matching Presented by Guan Wang Kent State University October 24, 2011.

1 gStore: Answering SPARQL Queries Via Subgraph Matching Presented by Guan Wang Kent State University October 24, 2011.
Lei Zou 1, Jinghui Mo 1, Lei Chen 2, M. Tamer Özsu 3, Dongyan Zhao 1 1 gStore: Answering SPARQL Queries Via Subgraph Matching 1 Peking University, 2 Hong.

Lei Zou 1, Jinghui Mo 1, Lei Chen 2, M. Tamer Özsu 3, Dongyan Zhao 1 1 gStore: Answering SPARQL Queries Via Subgraph Matching 1 Peking University, 2 Hong.
Effective Keyword Based Selection of Relational Databases Bei Yu, Guoliang Li, Karen Sollins, Anthony K.H Tung.

Effective Keyword Based Selection of Relational Databases Bei Yu, Guoliang Li, Karen Sollins, Anthony K.H Tung.
Jiannan Wang (Tsinghua, China) Guoliang Li (Tsinghua, China) Jianhua Feng (Tsinghua, China)

Jiannan Wang (Tsinghua, China) Guoliang Li (Tsinghua, China) Jianhua Feng (Tsinghua, China)
DISCOVER: Keyword Search in Relational Databases Vagelis Hristidis University of California, San Diego Yannis Papakonstantinou University of California,

DISCOVER: Keyword Search in Relational Databases Vagelis Hristidis University of California, San Diego Yannis Papakonstantinou University of California,
In Search of Influential Event Organizers in Online Social Networks

In Search of Influential Event Organizers in Online Social Networks
Connected Substructure Similarity Search Haichuan Shang The University of New South Wales & NICTA, Australia Joint Work: Xuemin Lin (The University of.

Connected Substructure Similarity Search Haichuan Shang The University of New South Wales & NICTA, Australia Joint Work: Xuemin Lin (The University of.
SPARK: Top-k Keyword Query in Relational Databases Yi Luo, Xuemin Lin, Wei Wang, Xiaofang Zhou Univ. of New South Wales, Univ. of Queensland SIGMOD 2007.

SPARK: Top-k Keyword Query in Relational Databases Yi Luo, Xuemin Lin, Wei Wang, Xiaofang Zhou Univ. of New South Wales, Univ. of Queensland SIGMOD 2007.
Dong Deng, Guoliang Li, Jianhua Feng Database Group, Tsinghua University Present by Dong Deng A Pivotal Prefix Based Filtering Algorithm for String Similarity.

Dong Deng, Guoliang Li, Jianhua Feng Database Group, Tsinghua University Present by Dong Deng A Pivotal Prefix Based Filtering Algorithm for String Similarity.
1 Efficient Subgraph Search over Large Uncertain Graphs Ye Yuan 1, Guoren Wang 1, Haixun Wang 2, Lei Chen 3 1. Northeastern University, China 2. Microsoft.

1 Efficient Subgraph Search over Large Uncertain Graphs Ye Yuan 1, Guoren Wang 1, Haixun Wang 2, Lei Chen 3 1. Northeastern University, China 2. Microsoft.
Indexing Network Voronoi Diagrams*

Indexing Network Voronoi Diagrams*
Speaker: Alexander Behm Space-Constrained Gram-Based Indexing for Efficient Approximate String Search Alexander Behm 1, Shengyue Ji 1, Chen Li 1, Jiaheng.

Speaker: Alexander Behm Space-Constrained Gram-Based Indexing for Efficient Approximate String Search Alexander Behm 1, Shengyue Ji 1, Chen Li 1, Jiaheng.
A Generic Framework for Handling Uncertain Data with Local Correlations Xiang Lian and Lei Chen Department of Computer Science and Engineering The Hong.

A Generic Framework for Handling Uncertain Data with Local Correlations Xiang Lian and Lei Chen Department of Computer Science and Engineering The Hong.
Reasoning and Identifying Relevant Matches for XML Keyword Search Yi Chen Ziyang Liu, Yi Chen Arizona State University.

Reasoning and Identifying Relevant Matches for XML Keyword Search Yi Chen Ziyang Liu, Yi Chen Arizona State University.
Suggestion of Promising Result Types for XML Keyword Search Joint work with Jianxin Li, Chengfei Liu and Rui Zhou ( Swinburne University of Technology,

Suggestion of Promising Result Types for XML Keyword Search Joint work with Jianxin Li, Chengfei Liu and Rui Zhou ( Swinburne University of Technology,

Similar presentations

Ads by Google