weighted overage agiciy ortalama T) Lageage... Interpolation aes Interpolation is the eStimation of tne vawe of a fdnction by taking the weightect. average of neighbouring points, excimple. fod = fq) 4 £03) -£ (rand Og x0) error =. 92 fl%2) = € Ne ite red. estimated value wedine + x Lagronge interpolation is to fit a (nA) ocater Polynomial fo @ doi set with a points (X,Y) . This can be fitled with exact Collocation (@). au. data poets. : pede For a linear fit to two deta points (Yoo) t (td) fle 0%) ee 60%) eal wpe | Oren 9 fideo tft) =f ante @xexy 9 FR) = flo) 10 =f l%)+ For three data points fle fl%) (x0) (X= Xe) (Yo -K4) (¥orXz) Generate fb) ov Yo Yo % XM +f UK) enor of lagrange — polynomials a r , ate ie v von] ee i:0 Tava! example. Fi On exe 3 Cubis spline’: navda fookla veriliyae. O2zamon £ ‘ler 3, olereciaten olmaly Ux-to) (X> x2) ( oo Zhitta) . where Ly = 7 re Or) (x-xi) 20 (X;=¥)) ai (F< x0) ct polynemial to the foltotwing koneet we 5 6 7 3 olerat. dineliia x1) (X=Xo) (x=¥4) (hy te) ayy) ROSIIHASMHAMSESKRHSHAMHKTDITTATTLK HTK KE HHSHOVHATITAINAAAKft) 1 = a FO) = 6 LEAD RB) 6 tamed) (r-3) (= 10-2) 103) GON 2 (4-3) 4% (RAO) 3) Ceo) KA 2) (2-0) (2-17 (2-3) (3-0) (3-17 (3-2) F0q) = Steer 20-3) 48 (K-23) *(-2) xO OA) + Bx (x4) (xD) ¢ 2 z ‘ Rid = Lyd aos + ares 6 3 NOTE! Fitting an n'* orcler polynomial 40 data moy give exact resus Gf points, but ovbsicle the points ; the fonction starts to "wigge! ave 40 5 polynomial nature. Thus rif becomes unrealistic Jo fii a polynomial function for higher on valves. In [for instance excel) his fit Ipoks Smooth = CUBIC SPLINE IL. Cobie Spline 6 data. points a S cubic splines # Cubic “splines. .oré._inter polation fonctions “based daa | piecewise |polynemial. nierpolation In. other werdsi'8 ase . rultyple’' polynomials Spline : here 4 & Foe a points 2 (n-4) Cuble functions ore firtied. » Each spline Sy = z Ske Oekel! 20 Kt bee 1k toy 02 but for 5 cubic splines for. 6 oala points) we have onknoun coef ficiend Givens mtd points. splines An known Coefficients (nt) data pn # of Splines, s¢ of untnques =4a A) Se (%E) fh bee) Sela) oo, Qld] Salto gs Sa lint! Son’ Si Ua) = 2 5 Se ltgdeyg , : "Sa (¥a) -Yp =2n equations / kag Peis varsa, 2 koh sir boslanges.. ye ) bir biti nottatineta by b2elligin Odh Collocertan { 2) SO) = Sate) swine oe fo be mast ( fesye Pats 3 (n-A) equakons ) (first cerivetive is. contiqvoes) B) SEE) = Seay Ota) a Serene clerive x) cont! 2 seaotvese corvaivre. oun digs) 3 (A-4) equations esithiyorum % We have. Ain ..onknowns . ¢. 4n-2.. equations we need 2 more. _ eqns. If we .assume. ...../" O..\ Netural aptine..( The. second derivative ci G20 Jat tre begining 4. end 1s..20r0) 3 Semi ilar spline ‘nn creas yophirdyjorta, 4 ase mv btemeles almqyscak ame. tonteptiol sory geletek oe oidugry bilma gare. VeuuceneueceescaGcnsceeeteaeeceaneeuanaaveaacans7 LO cod ppcions Lineor System of — equations Tricliagonal matrix ( bende matrix) time consuming than full matrix. St sap x3 sop x* +eEx tae Se) =e lage Pe Mest (Kee? a py Oe ¥) age Uren) he She. wonere hy = Xu XE > interval width mee Sp" lat) 9 setond obrivative pe flv) _ mene he 6 cel 2 = Fe) rye BE qe aE ; 5 (Sq-dnt) ee # For this System 4 Gauss elimination ig much more easier xr less* Cubis. Splines are — Conkicuon piewete polynomials... (214.2014 TL. Least Squares. . Regression eee eee ee "4 medel vem Akt fO) a Sometimes collocation . results in unphysical models e Wecan fil fitleax +h a No Collocation is forced fixlsox? thxre a Ersoy =f le) ~y| 7 Fries acosy~be” & Minimize “error” Somehow caller - taal Error Norms {lohe! norm cb we minimize?) er = [yi-flacd] + for each dala point € Q Maximum Neem : E xCerea,-. en), Error= JEM, = max lei) @ Average. Norm: giteasest 2 ren 5 Mele 4 3 i @ foot. Nean Square (RMS) errr + tent = (Ely mI = (yi- gui) tee LOBHSRHAARAAAHRSSHAARKRHSOMAHAKAANRALKS AHEM OHOHOSCHAEADIVAVM*NARA’ ” y ’ » ’ , ” 4. Least Squares Fit. of a Straight Line! ep For @ data set provides with "0" data. points + We would tke 4o fit Flalearex We wovtd like 4o minimize HEI, wrt ae b, ~o How do vse minimive ele wrt 2 parameters? Ela) DElaw) -9 y DELO) 5 oa ab These @¥b MINIMIZE the error aotm b b ———— alll = (24 , ~ fou)! = lye - aby)? elk = Elare) = z [yi~a-bxi? BE =0 “BE LO 3a 3b 9 2 38 [FP ty-a-ei? =O (2) da ~ da J int 26 2 2 = fyi-atoyl (<0 (x) Sb Ob Ning 3 i a 2 (yy -a@~bxil? 2 lyi-a-bail® : Be. Sg_ivt =O then, Fr Ba 3a 2 iy-a bn SAE, tea ceatty et Eley anid (yé-a-bac}] =O saueaeT Fy aia ZZ yt e-bay) 1) 0 tdGotinvecet 4 . ¥ LE Wira- bil 20 eqn lz) dot 20. ( z ah = a ‘Yi -a bhi) z (yt-a-bail] no, : #722 [ yrea-en)e] 20 eqn) i > We have 2 equations O Z ly-a-on) =-0 4 Z (yi-a-brii ti = oe Rearrange O x @ a a a a @ fj yi = Fat Soy 4 Fy 20a + F oj we tet ee ta fee [ee system of 2 equations o) @te 2 ne a a Mego 2 DO 4 bey DZ xyie Z ya + ZT by? ref tad ot ret tet me a a zy (2 ; ze . ia a b Ziyi i Sener preetee | q a sete By ta aa =] oo [ dé EM zt? 2 ay ( Moly ey * Swavid toto bate jdevame vars PFSFSIMIRAARSSANRRMAMNANRRADE BORE mee H~ ~~ -() , . , > , J » » oY» ” pb , y » ’ ’ > » > 3 Then. fixl ater WEI, 2? = (94 Exampie Excovated Soily, aa ve 2 . Z 35 2S + 10 ’ 3,5 10 + . 5 20 Fev yearby, 924 Ly bi gd 5 Hee 4 , = yp 24015 i = 13 Zt 2475 Sriyes PRS 4 AS fT @ 40.5 Q=-F,13 ay. ansfey. [15984 b- 5.3 I + Ga28 ards].Alternative Technique. to find arb (Linear Algebra) ‘ xi yi oe — piu 9 data — points 1 : 7 ‘ ‘ Obits Yt tbr = ye ? a+ bin =n 7] q ur " Ye i YL olny =] + 1 Mm Inve ga fart axe lit ret — (2ejterd? a {2x0} Dk kénik sata ererlernertical, by REnik slato. dolay ! LP OAISAAAARAMESHOROLLAGATATIF CIOL PLO LT AATATAINGLE,