Lad Evaluato Supported by MDS 1 XUE Fegchag School of Remote Sesg,Nag Uversty of Iformato Scece & Techology, Nag,Jag Su Provce,cha. 10044 xfc9800@16.com Abstract GIS-MCE s the ma method lad evaluato,but t s a lear method ad eglects multdmesoal complexty of factors used lad evaluato, whch leads to formato loss. Multdmesoal scalg (MDS) orgates from psychoaalyss, whch s used to descrbe multdmesoal data hgher dmesos by trasformg data hgher dmesos to geometry structure Lower dmesos. I the Lad evaluato model supported by MDS, the data s trasformed to a smlar space ad lad evaluato s completed accordg to the spatal clusterg based o the data s spatal smlarty,whch s a method drved by data ad ot depedetg o others pror assumpto.takg exproprato dvso Xuzhou cty as a example, t shows that the lad evaluato based o MDS meets the requremets of lad classfcato. Keywords:Lad Evaluato MDS Spatal Cluster model 1. Itroducto Lad evaluato volves factors of sol, clmate, vegetato, topographc ad hydrology, whch s a aalyss tegratg spatal formato. I aalyss of tegratg spatal formato, spatal data of dfferet types ad dfferet sources ca be take as attrbute of spatal cells, whch have dfferet spatal resoluto ad dfferet spatal scale. So lad evaluato ca be take as tegratg spatal formato of dfferet sources ad dfferet spatal scales specfc formato space. GIS-MCE s the ma method lad evaluato,but t s a lear method ad eglects Multdmesoal complexty of factors used lad evaluato, whch leads to formato loss. Multdmesoal scalg(mds) orgated from psychoaalyss, ad t s used to descrbe multdmesoal data hgher dmesos by trasformg data of hgher dmesos to geometry structure Lower Dmesos. Based o ths,lad evaluato supported by MDS s a method drved by data ad ot depedetg o others pror assumpto..the prcples of MDS Set the Attrbute measuremet matrx C(C=(d ) ) of obects has bee gve r-dmesoal space,sad C s the smlarty matrx for the obects. MDS usg C to obta p (p r)-dmesoal vector Z = (z (1), z (),..., z () ), ad obects ca be expressed by vector Z, sad that Z s a Mmetc Structure of C. Set Dˆ s Attrbute measuremet matrx Obtaed from the Z,ad sad Dˆ s Mmetc dstace matrx of C. MDS's target s both to make the obects to acheve p (p r)-dmesoal expresso, but also to make Dˆ as close as possble to C, ad the process acheves the classfcato of obects,where oly -matrx C s avalable. Youg- Household theorem approachs creatg p-dmesoal vector Z from the matrx C. For the matrx 1 Foudato:scece ad techology support proect of Nag Uversty of Iformato Scece ad Techology(Proect Number:S81083001, S8109008001)
C = (d ), set B = (b ), where 1 b = (-d + 1 d + 1 1 d - 1 1 1 1 d ) (1) Setλ 1 λ λ r s postve characterstc roots of B, e 1,e,e r s egevector correspodg to λ 1 λ λ r take Z = (e 1, e,..., e r ) = (x ) r, each row of Z matrx correspods to a mmetc spatal coordate of obect. Take Z dmeso p <r to realz the classfcato of obects low-dmesoal space. 3. Lad evaluato model based o MDS the process of aglczg lad formato for dfferet types of factors, there s bg dfferece data preprocessg, such as data qualty cotrol, data regstrato, space ut ad so forth, but through a certa mode after pretreatmet dfferet sources, dfferet scales, dfferet elemets of the lad factor the samplg data ca be tegrated a formato space, whch ca study the characterstcs of spatal formato. Set X ( = 1... N) s the space evaluato ut of lad, ad the spatal coordates s (x, y). lad evaluato factors subordate to X ( = 1... N) costtute a r-dmesoal sampled data formato space. C s the the spatal smlarty matrx of X ( = 1... N) based o sampled data of lad evaluato factors. Z s a Mmetc Structure of X ( = 1... N), ths mmetc Structure, X s spatal coordates s (u, v ). The process of lad evaluato based o MDS ca be expressed as: C=t { X } (=1.):t s the smlarty measure matrx costructor; Z=H { C }:H s costructor to form Mmetc Structure;, F =E{(u,v )}:s spatal clusterg for X Based o (u, v ) the, E s the aalyss process; F=Q{ F, }:F s cluster expresso of the lad evaluato space ut based o (x, y ), Q s the, trasformato process of F to F. 4. Realzg Model Takg exproprato dvso Xuzhou cty as a example, Realzg the Model of lad evaluato based o MDS as follows. 3.1 Establshg mult-dmesoal formato space Takg vllage-level admstratve areas as the basc spatal ut,ad Takg Aual output value of ut lad area (PV), affect degree of cetral cty (CD), accessblty to roads (RD), exteral trasport facltato degree (TD) status of lad use (LU), lad supply ad demad (LS), the local ecoomc developmet (EL) as the formato dmeso,establsh mult-dmesoal formato space: y (pv,cd,rd,td,lu,ls,el ) Where pv, cd, rd, td, lu, ls, el s the regular, stadardzed property value of aual output value per ut area (PV), ceter cty degree (CD), accessblty to roads (RD), exteral trasport facltes degrees (TD) lad-use codtos ( LU), lad supply ad demad (LS), the local ecoomc developmet (EL). 3. The smlarty structure of spatal cells Takg pv, cd, rd, td, lu, ls, el as dstace measure, establsh Eucldea dstace smlarty measure betwee spatal ut, : s ( pv pv ) ( cd cd ) ( rd rd ) ( td td ) ( lu lu ) ( ls ls ) ( el el takg s as a smlarty measure to establsh spatal attrbute smlarty matrx of Exproprato Dvso as a table1 (Local) )
Tab.1 Smlarty Matrx of Spatal Cells Attrbute Zhagxaolou Zhagzhuag Xa Lumalu Pagzhuag Shx Shdog Lhuag Chegzhuag Gusha Wagxzhuag Zhoutu Yagx Zhagxaolou 0.000 Zhagzhuag 0.040 0.000 Xa 0.100 0.137 0.000 Lumalu 0.33 0.47 0.181 0.000 Pagzhuag 0.493 0.467 0.59 0.4 0.000 Shx 1.515 1.477 1.591 1.535 1.17 0.000 Shdog 0.77 0.84 0.38 0.063 0.373 1.494 0.000 Lhuag 0.35 0.361 0.76 0.430 0.804 1.836 0.493 0.000 Chegzhuag 0.443 0.477 0.343 0.314 0.74 1.846 0.353 0.380 0.000 Gusha 0.779 0.774 0.755 0.577 0.456 1.383 0.516 1.001 0.737 0.000 Wagxzhuag 0.555 0.556 0.50 0.340 0.374 1.455 0.8 0.759 0.508 0.4 0.000 Zhoutu 0.67 0.307 0.180 0.90 0.693 1.770 0.35 0.159 0.41 0.848 0.606 0.000 Yagx Yagdog 0.381 0.39 0.34 0.148 0.48 1.554 0.113 0.547 0.316 0.456 0.14 0.39 0.179 Wutu 0.439 0.464 0.35 0.35 0.594 1.71 0.44 0.495 0.171 0.566 0.339 0.337 0.15 Lwo 1.056 1.061 1.005 0.87 0.819 1.698 0.779 1.04 0.86 0.368 0.510 1.045 0.859 Lulou 0.97 0.938 0.863 0.693 0.781 1.766 0.654 1.037 0.680 0.384 0.4 0.879 0.714 Taq 0.753 0.771 0.675 0.55 0.738 1.815 0.50 0.815 0.445 0.471 0.368 0.660 0.59 3.3 Coordates calculato of mmetc structure Usg ALSCAL MATLAB to calculat the space coordates (u, v ) mmetc structure for each spatal ut as show Table : Tab. Spatal Cells Coordate Commo Mmetc Structure Vllage Name Coordate Vllage Name Coordate Vllage Name Coordate qapatao -0.07 0.741 zhaowu -0.478-0.184 xada 1.071 0.043 gsha -0.3-0.134 zhagtu -0.3-0.6 huagsha 0.63-0.35 dahuagsha -0.418-0.09 daha -0.384-0.15 luotuosha 1.075-0.145 kelazhuag -0.54-0.066 lzhuag -0.564-0.039 shzsha 0.858 0.046 wagkele -0.73-0.144 mazhuag -0.517-0.044 xahetou 0.75-0.054 laggudu -0.85-0.65 dgzhuag 0.344 1.41 saguamao 0.77 0.011 xaohuagsha -0.765-0.94 shaghetou -0.097-0.384 houshawo 0.844-0.4 xzhua -0.739-0.59 dawagmao 0.108 0.555 asha 0.46 0.10 qawaga -0.866-0.083 chegzhuag -0.08 0.566 tasha 0.617 0.15 pol -0.784-0.138 lupu -0.509-0.54 chapeg 0.375 0.4 damao -0.777 0.41 chaa -0.9 0.908 dashatou 0.3 0.339 ara -0.589 0.365 lu -0.365-0.75 hasha 0.671 0.3 gusha -0.474 0.443 zhaoda -0.85 0.083 xgshaz 0.179 0.34 zhagzhuag -0.653-0.04 suda -0.58 0.91 shzhuag -0.06 0.106 houyao -0.66-0.161 patag -0.518-0.009 quyg 0.00 0.119 hou -0.575 0.303 aglou -0.557-0.197 dglou -0.03-0.104 qayao -0.56 0.169 tagfag 0.078 0.147 susha 0.399 0.193 hetao -0.604 0.483 cuzhuag 0.535 0.9 huohua 0.36 0.4 dogcu -0.58 0.05 duasha 0.19-0.108 duazhuag 0.878-0.464 houba -0.688-0.077 caosha -0.135-0.9 shagsha -0.354 0.487
qaba -0.855-0.001 fegzhuag -0.3-0.166 doghe -0.13 0.347 dahu -0.607-0.01 yaozhuag 0.58 0.157 xhe -0.019 0.01 qaohu -0.10-0.137 shl 0.505-0.091 bal 0.976-0.146 chagsha -0.4-0.3 wafag 0.944 0.111 tul 0.037 0.185 tushash -0.041-0.338 daz 0.641-0.074 *Coordate value os the relatve value of 1 / 1000 as the basc ut 3.4 Spatal clusterg mmetc structure Realzg spatal vsualzato of mmetc structure accordg to coordates (u, v ) as show Fgure 1ad realzg spatal cells aggregato accordg to spatal relatos mmetc structure as show Fgure. Fg.1 Mmetc Structure of Spatal Cells Fg. Parttog Mmetc Structure 3.5 Implemetato of exproprato dvso Trasform the classfcato of spatal cells mmetc structure to real space to acheve exproprato dvso, results of the dvso show Fgure 4. Fgure 5 s the map of actual exproprato dstrbuto, Xuzhou cty. Comparso of Fgures 4 ad 5, Comparso of Fgures 4 ad 5, both the spatal dstrbuto betwee the two s cosstet. Idcatg the lad evaluato method based o MDS meets the requremets of lad classfcato.
Fg.4 Lad Requsto Blocks Parttoed by Multdmesoal Scalg Fg.5 Actual Lad Requsto Blocks of Xuzhou cty 4.Summary Lad evaluato based o MDS does ot rely o pror assumptos, the basc approach s to trasform the data to a smlar space ad complete spatal clusterg accordg to the spatal smlarty of the data. The example of exproprato dvso Xuzhou cty shows that the lad evaluato based o MDS meets the requremets of lad classfcato. Refereces [1] XUE Feg-chag, BIAN Zheg-Fu. GIS combed wth MCE to Evaluate Lad Qualty,The Frst Iteratoal Coferece o Computer ad Computg Techology Applcatos Agrculture. Papers 365-368(007) [] Atoo Páez1,Darre M. Scott.Spatal statstcs for urba aalyss: A revew of techques wth examples,geojoural[j].005,61(1):53-67 [3] Patacch, Eleoora.Local aalyss of ecoomc dspartes Italy: a spatal statstcs approach.008, 17(1):85-11 [4] Coxtf,Coxmaa. Multdmesoal Scalg [M].Lodo:Chapma&Hall,1994. [5] Yougfw,Hamerrm. Theory Ad Applcatos Of Multdmesoal Scalg [M].Hllsdale:Erbaumassocates,1994. [6] Borg,Groee,P. Moder Multdmesoal Scalg:Theory Ad Applcatos [M].Newyork:Sprer1997. [7] Igwer Borg, Patrck J.F. Groee.: Moder Multdmesoal Scalg :Theory Ad Applcatos [M] New York : Sprger, 005.