An Efficient and Novel Query Answering Support for Uncertain Location-Based Queries

advertisement
International Journal of Engineering Trends and Technology (IJETT) – Volume 6 Number 1- Dec 2013
An Efficient and Novel Query Answering Support
for Uncertain Location-Based Queries
1
S.Rama Sree,2Mrs.K.S.B.Ambika
Associate Professor,2M.Tech Scholar
1,2
Aditya Engineering College, Aditya Nagar, Surampalem, Andhra Pradesh
1
Abstract: Searching of spatial information over search
engines is still an important research issue in the field of
spatial data mining.We are proposing an empirical model of
mechanism for location based(Spatial) databases, Even
though various approaches are delivered by the various
researchers for finding the result based on the keywords, and
usually spatial query is a combination of a location and set of
features. In our approach we are handling the spatial queries
in two ways and returns the only user specified number of
optimal results, we implemented a cache based approach for
efficient results.
Instead, current systems use ad-hoc combinations of
nearestneighbor (NN) and keyword search techniques to
tackle theproblem. For instance, an R-Tree is used to find
the nearestneighbors and for each neighbor an inverted
index is used tocheck if the query keywords are contained.
We show that suchtwo-phase approaches are inefficient.
I.INTRODUCTION
II.RELATED WORK
The World-Wide Web has reached a size where it is
becoming increasinglychallenging to satisfy certain
information needs. Whilesearch engines are still able to
index a reasonable subset of the (surface)web, the pages a
user is really looking for are often buried underhundreds of
thousands of less interesting results. Thus, searchengine
users are in danger of drowning in information. Adding
additionalterms to standard keyword searches often fails to
narrowdown results in the desired direction. A natural
approach is to addadvanced features that allow users to
express other constraints orpreferences in an intuitive
manner, resulting in the desired documentsto be returned
among the first results. In fact, search engineshave added a
variety of such features, often under a specialadvanced
search interface, but mostly limited to fairly simple
conditionson domain, link structure, or modification date.
A spatial keyword query consists of a query area and a
setof keywords shown in below figure. The answer is a list
of objects ranked accordingto a combination of their
distance to the query area and therelevance of their text
description to the query keywords. Asimple yet popular
variant, which is used in our runningexample, is the
distance-first spatial keyword query, whereobjects are
ranked by distance and keywords are applied as
aconjunctive filter to eliminate objects that do not
containthem.
Unfortunately there is no efficient support for top-k
spatialkeyword queries, where a prefix of the results list is
required.
Spatial query processing has takes an intresting research
area in now adays of research in the field of spatial query
processing in terms of spatial object.Spatial object is a
combination of Spatial entity and feature,Initally it process
the query as query oriented results and the results can be
sorted according to their Euclidean distance
ISSN: 2231-5381
In this paper are proposing a cache based approach for
handling the queries jointly with corresponding geocoding
parameters.
In the traditional approaches various mechanism
introduced for the spatial query processing either by the
distance measures or either by the clustering approaches,
for grouping the similar type of object by measuring the
distance with the centroids .In a nutshell, given a user
query consisting of several keywords, aStandard search
engine ranks the pages in its collection in terms of Their
relevance to the keywords. This is done by using a text
indexStructure called an inverted index to retrieve the IDs
of pagescontaining the keywords, and then evaluating a
term-based rankingFunction on these pages to determine
the k highest-scoring pages.(Other factors such as
hyperlink structure and user behavior areAlso often used,
as discussed later). Query processing is highly optimizedTo
exploit the properties of inverted index structures, storedIn
an optimized compressed format, fetched from disk using
efficientScan operations, and cached in main memory.
Let D be a dataset in which each object p ∈ D is a
pair (λ, ψ)of a spatial location p.λand a textual description
p.ψ(e.g.,the facilities and menu of a restaurant).Similarly, a
spatial keyword query [3] q = _λ, ψ_ has twocomponents,
where q.λis a spatial location and q.ψis a setof keywords.
The answer to query q is a list of k objects thatare in
http://www.ijettjournal.org
Page 33
International Journal of Engineering Trends and Technology (IJETT) – Volume 6 Number 1- Dec 2013
ascending order of their distance to the query
locationq.λand whose descriptions contain the set of query
keywordsq.ψFormally, let the function dist(・, ・) denotes
the Euclideandistance between its argument locations, and
let D(q.ψ) ={p ∈ D | q.ψ⊆ p.ψ} be the objects in D that
contain allthe keywords in q. The result of the top-k spatial
keywordquery q, q(D), is a subset of D(q.ψ) containing k
objectssuch that ∀p ∈ q(D) (∀p_ ∈ D(q.ψ)−q(D) (dist(q.λ,
p.λ) ≤dist(q.λ, p_.λ))). The joint top-k spatial keyword
query Q isa set {qi} of such queries.We introduce the
following notion to capture useful informationon a joint
query Q: (i) Q.λ= MBRqi∈Qqi.λis the minimum bounding
rectangle (MBR) of the locations
of the subqueries in Q, (ii) Q.ψ= ∪qi∈Qqi.ψis theunion of
the keyword sets of the subqueries in Q, and (iii)
Q.m= minqi∈Q|qi.ψ| is the smallest keyword set size of
asubquery in Q.We later define a variable qi.τthat captures
the upper boundkthnearest neighbor distance of subqueryqi.
The value Q.τ=maxqi∈Qqi.τthen represents the maximum
upper bound kthnearest neighbor distance of all the
subqueries in Q.3Referring to Figure 1, the joint query Q
contains threesubqueriesq1, q2, and q3 (shown as shaded
dots). The objects(e.g., restaurants) are shown as white
dots. We have that:q1.ψ = {curry,sushi}, q2.ψ = {seafood,
sushi}, q3.ψ ={curry, seafood}. Note that Q.λdenotes the
MBR of theshaded dots in the figure. We also have: Q.ψ=
{curry,seafood, sushi} and Q.m= 2.
and input querynode and compare all features of the query
node with all node features until it meets the leaf node and
add the results to the list frequently and returns the result
Input: Query, Cache Queries
Output: Result set generated for query
Procedure:
If Query available in cache
Result related to query: =ForwardToTreeprocess
(Query)
Else
Result related to query: = GeocodingtreeProcess
(Query)
Geocoding process(Query):
Parameters
Qi—Input Spatial Query
Qj(j=1…n) ---Set of Queries contains same Location
Dist[j] (j=1…..n)-----Array for set of distances
Procedure:
(xi,yi)---Geocodings of Qi
(xj,yj)--- Geocodings of all queries with respect to location
Dist[i]=Euclidean distance between the geocodes
While not leafnode
Read nodes from tree For Q.features
If Q.features[i]==Q.features[j]
Add to list
End while
Sort list by feature and distance
Return list.
ForwardToTreeprocess ()
Figure1: A Dataset of Spatial Keyword Objects
III.PROPOSED SYSTEM
In this paper we proposed an efficient approach for
handling the spatial query based on geocodings(latitude
and longitudes) and for optimal performance we introduced
Cache mechanism. we implemented both approaches with
cache and withoutcache. Our entire process is divided in to
two phases based on cache and geocodes.
Geocoding process receives the query and find and the
geocodings of the respective user from where he made a
query and get the geocoding values from the all the queries
from the tree. Find the distance with all the query nodes
ISSN: 2231-5381
1. Build an empty list
2 .Make a root node
3. if Qi in cache and status=false
For j=0 to n
Compare features(Qi,Qj) status=true;
For Each child in tree
If(status==true)
Getnodebyfeature (Qi);
Getnodebyfeature (Qj);
End
Else
Empty list ()
End For Each
4.Add nodes to list
http://www.ijettjournal.org
Page 34
International Journal of Engineering Trends and Technology (IJETT) – Volume 6 Number 1- Dec 2013
5.Return list
Step By Step process:
1. User provides the spatial query which involves the
spatial object and feature
2. if data not available in cache
Server retrieves the Object oriented results,
Ex:((Object, Feature)
Else
Retrieves the Optimal results from the cache
3. filter the feature oriented results from the Object
oriented results
4. Retrieve the Geo codings of the end user and calculate
the Euclidean distance between the end user geocodings
the filtered results
5. Sort the results based on Euclidean distance
IV.CONCLUSION
Finally we proposed an efficient a novel search
implementation on spatial databases with simple
implementation than the complex tree constructions like
Rtrees, in both cache based and non cache based(with
geocodings),our algorithms shows an optimal results than
the traditional approaches.
REFERENCES
[1] Y.-Y. Chen, T. Suel, and A. Markowetz. Efficient
query processingin geographic web search engines. In
SIGMOD, pp. 277–288, 2006.
ISSN: 2231-5381
6. Return the sorted distance wise spatial results to the end
user
Experimental analysis:
For implementation purpose we had used C#.net and
Asp.net with sql server, process involves with a spatial
query which includes the spatial object and feature, Query
processed by the server initially retrieves the results based
on the spatial object and computes the Euclidean distance
and sort the objects based on the Euclidean distance
between the user location latitude and longitude parameters
and retrieved object wise results then filter the results based
on the feature set which involves the data. the following
screen shows an example of spatial query and retrieved
results as follows.
[2] G. Cong, C. S. Jensen, and D. Wu. Efficient retrieval of
thetop-k most relevant spatial web objects. In VLDB, pp.
337–348,2009.
[3] I. De Felipe, V. Hristidis, and N. Rishe. Keyword
search onspatial databases.In ICDE, pp. 656–665, 2008.
[4] M. Duckham and L. Kulik.A formal model of
obfuscation andnegotiation for location privacy. In
PERVASIVE, pp. 152–170,2005.
[5] A. Guttman. R-trees: a dynamic index structure for
spatialsearching. In SIGMOD, pp. 47–57, 1984.
[6] R. Hariharan, B. Hore, C. Li, and S.
Mehrotra.ProcessingSpatial-keyword (SK) queries in
geographic information retrieval(GIR) systems.In SSDBM,
p. 16, 2007.
[7] T. Brinkhoff, H. Kriegel, and B. Seeger. Efficient
processing of spatialjoins using R-trees. Proc. SIGMOD,
pages 237–246, 1993.
http://www.ijettjournal.org
Page 35
International Journal of Engineering Trends and Technology (IJETT) – Volume 6 Number 1- Dec 2013
[8] G. Cong, B. Ooi, K. Tan, and A. Tung. Go green:
recycle andreuse frequent patterns. In Data Engineering,
2004.Proceedings. 20thInternational Conference on, pages
128–139.
[9] A. Corral, Y. Manolopoulos, Y. Theodoridis, and M.
Vassilakopoulos.Closest pair queries in spatial databases.
Proc. SIGMOD, pages 189–200, 2000.
[10] I. D. Felipe, V. Hristidis, and N. Rishe. Keyword
search on spatialdatabases.In Proc. ICDE International
Conference on Data Engineering,2008.
[11] A. Guttman. R-trees: A dynamic index structure for
spatial searching.Proc. SIGMOD, pages 47–57, 1984.
[12] R. Hariharan, B. Hore, C. Li, and S. Mehrotra.
Processing spatialkeyword(sk) queries in geographic
information retrieval (gir) systems.In SSDBM, page 16,
2007.
[13] G. Hjaltason and H. Samet. Incremental distance join
algorithms forspatial databases. Proc. SIGMOD, pages
237–248, 1998.
[14] H. Jagadish, R. Ng, B. Ooi, and A. Tung.ItCompress:
An IterativeSemantic Compression Algorithm. In
Proceedings of the 20th International
Conference on Data Engineering (ICDE04), volume 1063,
pages20–00.
[15] H. V. Jagadish, N. Koudas, and D. Srivastava.On
effective multidimensionalindexing for strings.Proc.
SIGMOD, pages 403–414, 2000.
[16] K. Koperski and J. Han. Discovery of spatial
association rules ingeographic information databases. Proc.
SSD, pages 47–66, 1995.
[17] N. Mamoulis and D. Papadias.Multiway spatial joins.
Proc. TODS,26(4):424–475, 2001.
[18] B.-U. Pagel, H.-W.Six, H. Toben, and P.
Widmayer.Towards an analysisof range query performance
in spatial data structures. In PODS, pages214–221, New
York, NY, USA, 1993. ACM.
[19] D. Papadias and D. Arkoumanis. Approximate
processing of multiwayspatial joins in very large databases.
Proc. EDBT, pages 179–196, 2002.
[20] D. Papadias, N. Mamoulis, and B. Delis. Algorithms
for querying byspatial structure.Proc. VLDB, pages 546–
557, 1998.
[21] D. Papadias, N. Mamoulis, and Y. Theodoridis.
Processing and optimizationof multiway spatial joins using
R-trees.Proc. PODS, pages44–55, 1999.
[22] D. Papadias, Q. Shen, Y. Tao, and K. Mouratidis.
Group nearest neighborqueries.Proc. ICDE, pages 301–
312, 2004.
[23] N. Roussopoulos, S. Kelley, and F. Vincent.Nearest
neighbor queries.Proc. SIGMOD, pages 71–79, 1995.
[24] S. Shekhar and Y. Huang.Discovering spatial colocation patterns: Asummary of results.Proc. SSTD, pages
236–256, 2001.
[25] H. Shin, B. Moon, and S. Lee. Adaptive multi-stage
distance joinprocessing. Proc. SIGMOD, pages 343–354,
2000.
ISSN: 2231-5381
[26] Q. Vu, B. Ooi, D. Papadias, and A. Tung. A graph
method for keywordbasedselection of the top-K databases.
In Proceedings of the 2008 ACM
SIGMOD international conference on Management of
data, pages 915–926. ACM New York, NY, USA, 2008.
[27] N. Wang, S. Parthasarathy, K. Tan, and A. Tung.
CSV: visualizing andmining cohesive subgraphs. In
Proceedings of the 2008 ACM SIGMODinternational
conference on Management of data, pages 445–458.
ACMNew York, NY, USA, 2008.
[28] B. Yu, G. Li, K. Sollins, and A. K. H. Tung.ffective
keyword-basedselection of relational databases. In
Proceedings of SIGMOD, 2007.
[29] X. Zhang, N. Mamoulis, D. W. Cheung, and Y. Shou.
Fast mining ofspatial collocations.Proc. KDD, pages 384–
393, 2004.
BIOGRAPHIES
BIBILOGRAPHIES
K.S.B.Ambikais a student of Aditya
Engineering College, Surampalem.
Presently he is pursuing his M.Tech
[Computer Science] from this college
and he received his B.Tech from Sri
Prakash College of Engineering,
affiliated
to
JNT
University,
Hyderabad in the year 2006. His area
of interest includes Database Management Systems, Data
Mining, all current trends and techniques in Computer
Science.
S. Rama Sreeobtained her B.Tech.
Degree in Computer Science &
Engineering from AcharyaNagarjuna
University,
Guntur
and
M.Tech.Degree in Computer Science
from Jawaharlal Nehru Technological
University Kakinada, India. She is
currently a Research Scholar and
working as Associate Professor and Head of the
Department of Computer Science & Engineering at Aditya
Engineering College, Surampalem, India. She has 13
International Journal Papers and 5 National/International
Conferences to her credit. Her Research interests include
Software Engineering, Cost Estimation, Fuzzy Logic,
Neural Networks and Neuro Fuzzy Systems.
http://www.ijettjournal.org
Page 36
Download