Spatial Keyword Ranking (SKR) based Dominated Location with Safe Zone

261 views

Published on

Generally, spatial objects (e.g., Hotels) not only have spatial locations but also have quality attributes (e.g., star, price etc). In wild animal rehabilitation, an appropriate location is selected from a set of options for returning an animal to the nature, after its treatment in a rehabilitation center. A location is preferred if it is far away from (competing) animals with multiple better abilities (e.g., speed, weight, age). Since those animals are much stronger in fighting for essential resources, e.g., water and food, they will endanger the rehabilitated animal that is still unaccustomed to wildlife. By maximizing the distances to potential nearest dominators (NDs), the preferred location improves the survival chance of the rehabilitated animal in the nature. Hence the safe zone is created to retrieve result within the region. If the query moves from the safe zone another safe zone is created fot that query, the intersected part are purned. These reduce the database storage.

Published in: Education, Technology, Design
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
261
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
6
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Spatial Keyword Ranking (SKR) based Dominated Location with Safe Zone

  1. 1. International Journal of Computer Applications Technology and ResearchVolume 2– Issue 2, 176 - 179, 2013www.ijcat.com 176Spatial Keyword Ranking (SKR) based DominatedLocation with Safe ZoneKrishna Balan Karthiga Sakthi PriyaPondicherry Engineering College Pondicherry Engineering College Pondicherry Engineering CollegePondicherry, India Pondicherry, India Pondicherry, IndiaAbstract: Generally, spatial objects (e.g., Hotels) not only have spatial locations but also have quality attributes (e.g., star,price etc). In wild animal rehabilitation, an appropriate location is selected from a set of options for returning an animal to thenature, after its treatment in a rehabilitation center. A location is preferred if it is far away from (competing) animals withmultiple better abilities (e.g., speed, weight, age). Since those animals are much stronger in fighting for essential resources,e.g., water and food, they will endanger the rehabilitated animal that is still unaccustomed to wildlife. By maximizing thedistances to potential nearest dominators (NDs), the preferred location improves the survival chance of the rehabilitatedanimal in the nature. Hence the safe zone is created to retrieve result within the region. If the query moves from the safe zoneanother safe zone is created fot that query, the intersected part are purned. These reduce the database storage.Keywords: Anonymizer; Database Management; Safe zone; Spatial Database; Spatial Objects;__________________________________________________________________________________________________1. INTRODUCTIONIn reality, spatial objects (e.g., hotels) not only havespatial locations but also have quality attributes (e.g.,star, price). An object p is said to dominate another onep0, if p is no worse than p0 with respect to every qualityattribute and p is better on at least one quality attribute.Traditional spatial queries (e.g., nearest neighbor, closestpair) ignore quality attributes, whereas conventionaldominance-based queries (e.g., skyline) neglect spatiallocations. From this Farthest Dominated Location (FDL)[1] retrieves the results, includes both quality attributes,and spatial objects with sufficient R-Tree algorithm toretrieve the data. For each query, Location based Server(LBS) need to analyze the query objects and the querylocation. Then LBS search its entire database to giverespective queries result. This process is effective forstatic object. If it is for dynamic object means the serversends the result for one location point that is notsufficient if the object moves to another location. Forthis, the proposed system include safe zone. This zonecreates a circular zone with range for the query, andlocation will be analyzed. We propose an efficient indexcalled Spatial Keyword Ranking (SKR) Tree datastructure which performs 1) Spatial filtering, 2) Textualfiltering and 3) Object ranking in a fully integratedmanner.2. RELATED WORKSTerminal devices, which are the monitoredmoving objects, obtain their own location informationfrom the GPS system and transmit it to the server via awireless communication network. The whole system’stimeliness and efficiency is affected by the wirelesscommunication bandwidth. The location informationupdates are often the bottleneck because of the limitedwireless bandwidth and the high sampling rate in thetraditional uniform time/distance interval strategies.Theidea behind the rectangular safe region (RSR) algorithmis to define a rectangular safe region for every objectaccording to the registered query and the latest locationobtained. As long as the object’s motions do not exceedits safe region, all the query result sets of the objectremain unchanged.The terminal device is informed of the safe regionassignments dynamically. Hence when a terminal devicefinds that it has exceeded the safe region, it will report itsnew location information.2.1 Naive Iterative Incremental (NII)NN SearchNaive Iterative Incremental NN (nearest neighbour)search is for processing theFDL query. NII takes as input 1) a two dimensional R-tree Rp on the spatial object set P, 2) a set of locations L,and 3) a c-dimensional design competence [4]. Itexamines each location s of set L and computes thenearest dominator of s, by performing the incrementalnearest neighbour search [3] of s on the tree Rp.During the iterative search, the largest ndd (nearestdominated distance) is maintained together with thecorresponding location. After all locations are examined,the maintained location is returned as the query result.2.2 Best-First Search (BFS) AlgorithmThe Best-First Search Algorithm is a basicgraph-searching algorithm. Best-First makes use of aheuristic (or quickly computed estimate of the cost toreach the goal from each node), called ĥ, to guide itssearch. The idea behind using a heuristic to guide thesearch is that the algorithm will not waste time probingpaths that do not seem likely to lead to the goal state(node). However, this means that an inaccurate ĥ
  2. 2. International Journal of Computer Applications Technology and ResearchVolume 2– Issue 2, 176 - 179, 2013www.ijcat.com 177function can misguide the search, which can result in thesearch finding a path to the goal that is not optimal. Forthis reason, one of the greatest challenges in graphsearching is to develop a heuristic function that can bequickly computed and that will be a very close estimateof the actual cost to the goal.2.3Spatial Join Based (SJB) AlgorithmSpatial join algorithm can be classified intothree categories. Assume that it have spatial join relationon R1 and R22.3.3 Nested LoopIn this algorithm, for each tuple of R1, entireR2 is scanned; any pair of tuples of R1 and R2 whichsatisfies the spatial join predicate is added to the result.2.3.4 Tree MatchingTree matching algorithm can be applied whenindices are available on both the relations. For thisdiscussion, it will assume that R-tree index is available.When index exists for only one relation, the index on theother relation is built on the fly and tree-matchingtechnique is applied.2.3.5 Partition-Based Spatial Merge JoinIn this algorithm, first both of the relations aredivided into p partitions if both of them do not fit in mainmemory. After that partition i of R1, where pi 1 ,is compared with corresponding partition i of R2. Thisstrategy is very good when no indices are present on boththe relation.3. DESIGN FOR PROPOSED WORK4. SKR TREEWe propose an efficient indexing scheme called SKRtree (Spatial Keyword Range tree),which indexes boththe textual and spatial contents of objects to support dataretrievals based on their combine textual and spatialrelevance, which, in turn, can be adjusted with differentrelative weights. In fig.1 structure of SKR tree has nodeswhich have both spatial and non spatial information ofthe data object.Figure 1. Structure of SKR treeN1 and N2 are the child nodes,AN1 - keyword, and DN1 - spatial data.SKR Tree Construction AlgorithmInput: Set of Objects DOutput: Root of SKR treeProcedure:1: Ne ← 02: For each p €D do3: geo code p and represent Lp withMBB mp4: if for some e €Ne, me = mp then5: add p to e′s dataset De;6: else7: create a new entry e;8: set me ← mp and De ← {p};9: Ne ←Ne U {e};10: End if11: End for12: For each e € Ne do13: While Ne> n max do14: Cluster the data according to min/maxinto nodes15: Ne ← Ne′16: End while17: End for5. SAFE REGION BASED LOCATIONUPDATESQuerySnap shotqueryNearestNeighborQueryRangequerySearch forlocationRankedobjectsRetrievedataCreatingsafe zoneSKR-TREESpatialfilteringKeywordfilteringRanking
  3. 3. International Journal of Computer Applications Technology and ResearchVolume 2– Issue 2, 176 - 179, 2013www.ijcat.com 178Figure 2 demonstrates the infrastructure of a movingobject query system. The kernel of the system is thecontrol center (the main server of the system) in thecenter of the figure which runs the Moving ObjectDevice (MOD) engine, collects location information,handles continues queries and provides query results tothe application servers to the right of the figure.Therefore, the major computation workload is applied tothe main server/control center of a MOD system. Forsimplicity, we refer to the main server/control center asserver. Terminal devices, which are the monitoredmoving objects, obtain their own location informationfrom the GPS system and transmit it to the server via awireless communication network. The whole system’stimeliness and efficiency is affected by the wirelesscommunication bandwidth. The location informationupdates are often the bottleneck because of the limitedwireless bandwidth and the high sampling rate in thetraditional uniform time/distance interval strategies. Aslong as the object’s motions do not exceed its saferegion, all the query result sets of the object remainunchanged shown in Figure 3.Figure 2. Infra structure of object systemFigure 3. Location update of a moving object with circular safe regionunder continuous query.The safe region of object o (referred to as o.sr) as acircle with the center at the location of the object and theradius of o.r (Figure 3). By assuming the maximumspeed of the object is o.maxspd, then the continuousquery result of query q will not be affected within thetime interval of o.r/o. maxspd. Hence the server can issuea location report query to the terminal device at the timeof (o.r/o.maxspd -δ)whereδ is the sum of communication andcomputation delays.o.r is Radius of the object safe region.o.sr is Object safe region.o.maxspd is Maximum speed of the movingobject.5.1 Circular Safe Region Calculationand Updates for Continuous kNNQueriesFollowing is the formula to update the safe region radiusfor the i thobject in the object set ascending sorted bydistance to ordered kNN query q. The first object in theresult set is q and the extra object ok+1 is kept forcalculation of safe region radius of ok.where quar(q) is the radius of the quarantine region forquery q which surround and only surround the saferegions of all objects in the result set. Therefore, quar(q)= dist(ok, q) + ok.Figure 4. Quarantine region assignment for an ordered kNNquery.Figure 4 shows how such a quarantine region is assignedfor such an ordered kNN query q. Hence the followingproperty: either in or out of the kNN query result set(inside or outside the quarantine region), none of the saferegions of objects overlaps with each other. Therefore,when all the objects are moving inside their own saferegions, the result set and its order are not affected.6. CONCLUSIONSafe zone is created to get the optimal location whilecomparing with farthest dominated location. Safe zone iscreated so as to reduce the search space and time forserver because it retrieves the data within that regionspecified by the user. A quarantine region is alsoassigned for one or more objects present in the sameTerminaldeviceTrackingdeviceMainserverApplicationserverApplicationserver
  4. 4. International Journal of Computer Applications Technology and ResearchVolume 2– Issue 2, 176 - 179, 2013www.ijcat.com 179zone. This project proposes an efficient index SKR-treeand algorithms which perform spatial filtering, textualfiltering and object ranking in a fully integrated manneris expected to have reduce the database storage andreduction of CPU time.7. REFERENCES[1] Hua Lu and Man Lung Yiu, “On ComputingFarthest Dominated Locations”, IEEETransactions on Knowledge and DataEngineering, Vol. 23, No. 6, pp. 928-941,2011.[2] M.A. Cheema, L. Brankovic, X. Lin, W.Zhang, and W. Wang, “Multi-Guarded SafeZone: An Effective Technique to MonitorMoving Circular Range Queries”, Proceedingsof IEEE 26thInternational Conference on DataEngineering (ICDE), pp. 189-200, 2010.[3] G. Hjaltason and H. Samet, “DistanceBrowsing in Spatial Database”, ACMTransaction on Database Systems, Vol. 24, No.2, pp. 265-318, 1999.[4] N. Roussopoulos, S. Kelley, and F. Vincent,“Nearest Neighbour Queries”, Proceedings ofACM International Conference onManagement of Data, pp. 71-79, 1995.[5] Y. Du, D. Zhang, and T. Xia, “The Optimal-Location Query”, Proceedings of InternationalSymposium on Spatial and TemporalDatabases, pp. 163-180, 2005.[6] X. Huang and C.S. Jensen, “In-Route SkylineQuerying for Location-Based Services”,Proceedings international Workshop on Weband Wireless Geographical InformationSystems, pp. 120-135, 2004.[7] D. Papadias, Y. Tao, G. Fu, and B. Seeger, “AnOptimal and Progressive Algorithm for SkylineQueries”, Proceedings of ACM InternationalConference Management of Data, pp. 467-478,2003.

×