Partial Linear Quantile Regression and Bootstrap Confidence Bands
|
|
- Linda Park
- 5 years ago
- Views:
Transcription
1 Partial Linear Quantile Regressin and Btstrap Cnfidence Bands Wlfgang K. Härdle Ya acv Ritv Sng Sng Ladislaus vn Brtkiewicz Chair f Statistics C.A.S.E. - Center fr Applied Statistics and Ecnmics Humbldt-Universität zu Berlin Department f Statistics The Hebrew University f Jerusalem
2 Risky t see nly part(s) f the truth!
3 $ N (age) lg(salary) Years Matthew Effect Yu et al. (2003)
4 LIDAR (Light Detectin and Ranging) Range Lgrati Lgrati (received light frm 2 laser surces) Range (distance befre the light is reflected back)
5 Mtivatin 1-4 Quantile Regressin QR: cnditinal behavir f a respnse Y Median regressin = mean regressin (symmetric) Gradually develping int a cmprehensive strategy fr cmpleting the regressin predictin, Kenker & Hallck (2001) Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
6 Mtivatin 1-5 Example Financial Market & Ecnmetrics VaR (Value at Risk) tl t measure risk, Lauridsen (2000) Detect cnditinal heterscedasticity, Kenker & Bassett (1982) Labr Market Investigate differential effects, Buchinsky (1995) lg (Incme) = A(year, age, etc) +β B(educatin, gender, natinality, unin status, etc) + ε Inequality analysis, Juhn et al. (1993) Intersectin bunds, Chernzhukv et al. (2009) Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
7 Mtivatin 1-6 Quantile Regressin l(x) = F 1 Y x (p) p-quantile regressin curve l(x) = linear (parametric) frm, Kenker & Bassett (1978) l h (x) quantile-smther Hw t decide between functinal frms? (glbal variability f the estimate, peak r valley really a feature?) Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
8 Stchastic Integrals and Differential Equatins 5-2 Y X Mtivatin 1-7 This part f the curse prvides the tls fr the valuatin f Therem ptins. (Härdle and Sng (2009)) An apprximate We will define (1stchastic α) 100% prcesses cnfidence as slutins band ver f stchastic [0, 1] is differential equatins (SDE). l h (t) ± (nh) 1/2 {p(1 p)/ˆf X (t)} 1/2ˆf 1 {l h (t) t} Stchastic prcess {d n + inc(α)(2δ cntinuus lg n) time 1/2 } {λ(k)} 1/2, (1) A stchastic prcess in cntinuus time where c(α) = lg 2 lg lg(1 α) and ˆf is a cllectin X (t), ˆf f randm variables {X t ; t R + } with a cntinuus time{l variable h (t) t} are t. cnsistent estimates fr f X (t), f {l(t) t}. Emil Nrbert Julius Gumbel Wiener n BBI: Tl: Hungarian machine gun", x R 1 (KMT) SFE Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
9 Mtivatin Figure 1: The 0.9-quantile curve, the 0.9-quantile smther with h 0.9 = 1.25 and 95% cnfidence bands. QRBSGumble Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
10 Mtivatin 1-9 Challenges L( l h - l ) (lg n) -1 exp{-exp(-x)} Gumbel, Emil J. 43 Emil Julius Gumbel BBI (lg n) -1 L * ( l * h - l g ) Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
11 Mtivatin 1-10 Opprtunities Extend this t x R d and imprve band precisin? Hall (1991): btstrap imprvement (density) Hahn (1995): cnsistency f btstraping cdf Hrwitz (1998): btstrap (pintwise) fr median PLM: Green & Yandell (1985), Denby (1986), Speckman (1988) and Rbinsn (1988) Variable selectin fr QR: Liang and Li (2009) LASSO QR fr sparsity prblem, Bellni and Chernzhukv (2009) Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
12 Outline 1. Mtivatin 2. Btstrap Cnfidence Bands 3. Btstrap Cnfidence Bands in PLMs 4. Mnte Carl Study 5. Labur Market Applicatins
13 Btstrap cnfidence bands 2-1 Quantile Regressin {(X i, Y i )} n i=1 i.i.d. rv s, x J = (a, b) fr sme 0 < a < b < 1, y R Suppse Y i = l(x i ) + ε i, ε i F ( X i ) with F (0 X i ) = p. Bth l & F are smth. Estimatr l h ( ): the slutin f n i=1 K h(x X i )1{Y i < l h (x)} n i=1 n i=1 K < p K h(x X i )1{Y i l h (x)} h(x X i ) n i=1 K h(x X i ) S n : any slwly varying functin (e.g., S 2 n = S n is valid... ). Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
14 Btstrap cnfidence bands 2-2 Lcal rate f cnvergence f l h δ n = h 2 + (nh) 1/2 = O(n 2/5 ) with h n = O(n 1/5 ) Auxiliary estimate l g with larger bandwidth g n = h n n ζ (ζ: 4/45) n j=1 ˆF ( X i ) = K h(x j X i )1{Y j l h (X i ) } n j=1 K h(x j X i ) Hw gd is this apprximatin? Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
15 Btstrap cnfidence bands 2-3 ˆF Apprximatin Perfrmance arund 0 Lemma [Franke and Mwita (2003), p14] If assumptins (A1, A2, A4) hld, then fr any small enugh (psitive) ε 0, sup ˆF (t X i ) F (t X i ) = O p (S n δ n ε 1/2 + ε 2 ). (2) t <ε,i=1,...,n,x i J Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
16 Btstrap cnfidence bands 2-4 The Btstrap Cuple U 1,..., U n : i.i.d. unifrm [0, 1] rv s Btstrap sample Y i = l g (X i ) + ˆF 1 (U i X i ), i = 1,..., n Cuple with the true cnditinal distributin: Y # i = l(x i ) + F 1 (U i X i ), i = 1,..., n. Given X 1,..., X n : Y 1,..., Y n and Y # 1,..., Y n # distributed. are equally Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
17 Btstrap cnfidence bands 2-5 A Very Clse Cuple Y i = l g (X i ) + ˆF 1 (U i X i ), i = 1,..., n Y # i = l(x i ) + F 1 (U i X i ), i = 1,..., n. Values f Y # i and Yi are meaningful nly if U i p < S n δ n. By the inverse functin therem arund p, we have i: Y # max Y # i i l(x i ) <S nδ n l(x i ) Y i + l g (X i ) = O p {S n δ 3/2 n }. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
18 Btstrap cnfidence bands 2-6 Hw Clse? q hi (Y 1,..., Y n ) def = l h (X i ) fr data set {(X i, Y i )} n i=1 Assumptin A3 gives: max l g (X i ) l g (X j ) l(x i ) + l(x j ) = O p (δ n ) X i X j <ch lh and l # h : lcal btstrap quantile and its cupled sample analgue. Then lh (X i) l g (X i ) = q hi [{Yj l g (X j ) + l g (X j ) l g (X i )}) n j=1] l # h (X i) l(x i ) = q hi [{Y # j l(x j ) + l(x j ) l(x i )} n j=1] Thus max lh (X i) l g (X i ) l # i h (X i) + l(x i ) = O p (δ n ). Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
19 Btstrap cnfidence bands 2-7 Btstrapping Apprximatin Rate Therem If assumptins (A1) (A3) hld, then sup lh (x) l g (x) l # x J h (x) + l(x) = O p(δ n ) = O p (n 2/5 ). Btstrap imprves the rate f cnvergence Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
20 Btstrap cnfidence bands 2-8 Why Oversmthing? Take care f bias with tuning parameter: g Härdle and Marrn (1991), let b h (x) ˆb h,g (x) def = E l # h (x) l(x) def = E lh (x) l g (x) [ } 2 ] Investigate E {ˆbh,g (x) b h (x) X1,..., X n. Hw fast it cnverges t 0? Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
21 Btstrap cnfidence bands 2-9 Oversmthing Therem Under sme assumptins, fr any x J [ } 2 ] E {ˆbh,g (x) b h (x) X1,..., X n h 4 {O p (g 4 ) + O p (n 1 g 5 )} T minimize RHS, g = O(n 1/9 ), g h, where h = O(n 1/5 ) Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
22 Btstrap Cnfidence Bands in PLMs 3-1 The Multivariate Case x = (u, v) R d, v R: l(x) = u β + l(v) nte β als depends n p generally. Estimatin idea: ANOVA, apprximately linear frm (lcally) Partitin [0, 1] (fr v) in a n intervals I ni & regard l(v) as a cnstant item inside I ni. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
23 Btstrap Cnfidence Bands in PLMs 3-2 Tw Stage Estimatin Prcedure Linear quantile regressin inside each I ni + Weighted mean yields ˆβ: n a n ˆβ = arg min min ψ{y i β T U i l j 1 ( ) V i I ni } β l 1,...,l an i=1 j=1 Smth quantile estimate l h (v) frm (V i, Y i U i ˆβ) n i=1. Therem psitive definite matrices D, C, s.t. n( ˆβ β) L N{0, p(1 p)d 1 CD 1 } as n. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
24 Btstrap Cnfidence Bands in PLMs 3-3 Unifrm Cnsistency f ˆl h (v) Lemma Under assumptins (A7) & (A8), we have a.s. as n sup l h (v) l(v) C max{(nh/ lg n) 1/2, h α } (3) v J with cnstant C nt depending n n. If additinally α {lg( lg n) lg( nh)}/lg h, (3) can be further simplified t: sup v J l h (v) l(v) C{(nh/ lg n) 1/2 }. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
25 Btstrap Cnfidence Bands in PLMs 3-4 Multidimensinal Unifrm Cnfidence Bands Estimatin errr fr parametric part: O p (n 1/2 ). Btstrapping apprximatin errr fr nnparametric part: O p (n 2/5 ), dminating! Crllary Under the assumptins (A1) - (A8), an apprximate (1 α) 100% cnfidence band ver R d 1 [0, 1] is 1d u ˆβ + l h (v) ± [ˆf { l h (x) x} ˆf X (x)] α, where d α is based n the btstrap sample (defined later). Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
26 Mnte Carl study 4-1 Hw t Btstrap? 1) Simulate {(X i, Y i )} n i=1, n = 1000 w.r.t. f (x, y). f (x, y) = f y x (y sin x)1(x [0, 1]), (4) where f y x (x) is the pdf f N(0, x). 2) Cmpute l h (x) f Y 1,..., Y n and residuals ˆε i = Y i l h (X i ), i = 1,..., n. If we chse p = 0.9, then Φ 1 (p) = , l(x) = sin(x) x and the bandwidth is h = Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
27 Mnte Carl study 4-2 3) Cmpute the cnditinal edf: ˆF (t x) = n i=1 K h(x X i )1{ˆε i t} n i=1 K h(x X i ) with the quartic kernel K(u) = (1 u2 ) 2, ( u 1). 4) Generate rv ε i,b ˆF (t x), b = 1,..., B and cnstruct the btstrap sample Yi,b, i = 1,..., n, b = 1,..., B as fllws: with g = 0.2. Y i,b = l g (X i ) + ε i,b, Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
28 Mnte Carl study 4-3 5) Fr each btstrap sample {(X i, Yi,b )}n i=1, cmpute l h and the randm variable ] def d b = sup [ˆf {l h (x) x} ˆf X (x) lh (x) l g (x). (5) x J 6) Calculate the (1 α) quantile d α f d 1,..., d B. 7) Cnstruct the btstrap unifrm cnfidence band centered arund l h (x), i.e. l h (x) ± [ˆf {lh (x) x} ˆfX (x)] 1d α. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
29 Mnte Carl study % Quantile f Y Univariate X Figure 2: The real 0.9 quantile curve, 0.9 quantile estimate with crrespnding 95% unifrm cnfidence band frm asympttic thery and cnfidence band frm btstrapping. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
30 Mnte Carl study 4-5 Cnvergence Rate (n small) n Cv. Prb. Area (0.642) 0.58 (1.01) (0.742) 0.42 (0.58) (0.862) 0.31 (0.36) Table 1: Simulated cverage prbabilities & areas f nminal asympttic (btstrap) 95% cnfidence bands with 500 repetitin. T achieve same cv. prb., quantile regressin usually need mre bservatins than mean regressin Fr small n, btstrap s» asympttic s & nt sacrifice much n the band s width Use larger bandwidth n bth X & Y (1/ˆf {l h (x) x}) Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
31 Mnte Carl study 4-6 PLM QR Bivariate data {(U i, V i, Y i )} n i=1, n = 8000 with: y = 2u + v 2 + ε Φ(p) (6) s.t. β = 2 even fr different p, where u [0, 2], v [0, 1] and ε is the standard nrmal rv. The real 0.9-quantile curve l(x) = 2u + v 2. h = 0.2 & g = 0.7. Fr the fllwing specific set f randm variables, a n = 20, ˆβ = Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
32 Mnte Carl study 4-7 # f Partitins? Beta Estimates Different Quantiles Beta Estimates Different Quantiles Beta Estimates Different Quantiles Figure 3: ˆβ with respect t different p fr different # f bservatins, i.e. n = 1000, n = 8000, n = Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
33 Mnte Carl study 4-8 a n n = 1000 n = 8000 n = n 1/3 / n 1/3 / n 1/3 / n 1/ n 1/ n 1/ n 1/ Table 2: SSE f ˆβ with respect t a n fr different numbers f bservatins. Suggest a n = n 1/3 (cst / perfrmance) Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
34 Mnte Carl study 4-9 Nnparametric Part V Figure 4: Nnparametric part smthing, real 0.9 quantile curve with respect t v, 0.9 quantile smther with crrespnding 95% btstrap unifrm cnfidence band. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
35 Labr Market Applicatins 5-1 Labr Market Applicatin Hw incme depends n age w.r.t. different educatin levels? Relatin: lg (Wage) β Educatin + l(age) Administrative data frm the German Natinal Pensin Office West Germany, male, brn , sample 25-59, full-time, begin receiving a pensin in , uncensred n = bservatins Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
36 Labr Market Applicatins 5-2 German Pensin Office Data + Lnger perid + Large sample size + Precise measure f incme + Cmplete earning histry, true panel Rich peple nt recded (truncatin) Censring Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
37 Labr Market Applicatins 5-3 Educatin categries: lw educatin, apprenticeship and university Intuitin: E(y v, u = Lw educatin) < E(y v, u = Apprenticeship) < E(y v, u = University) Only part f the truth? Are the differences significant (unifrmly)? Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
38 Labr Market Applicatins 5-4 Bx Plt Lg Real Earnings Ages Figure 5: Bxplts fr lw educatin, apprenticeship & university grups crrespnding t different ages. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
39 Labr Market Applicatins lw educatin", 2 apprenticeship" and 3 university" /3 /2 = 25 partitins Quartic kernel, h = (after rescaling) Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
40 Labr Market Applicatins 5-6 β Estimates Beta Estimates Different Quantiles Figure 6: ˆβ crrespnding t 6, 13, 25 partitins. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
41 Labr Market Applicatins 5-7 Lwer End Lg Real Earnings Ages Figure 7: 95% unifrm cnfidence bands fr 0.20-quantile smthers with 3 different educatin levels lw educatin, apprenticeship & university. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
42 Labr Market Applicatins 5-8 Median Lg Real Earnings Ages Figure 8: 95% unifrm cnfidence bands fr 0.50-quantile smthers with 3 different educatin levels lw educatin, apprenticeship & university. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
43 Labr Market Applicatins 5-9 Upper End Lg Real Earnings Ages Figure 9: 95% unifrm cnfidence bands fr 0.80-quantile smthers with 3 different educatin levels lw educatin, apprenticeship & university Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
44 Labr Market Applicatins 5-10 Real effect f educatin fr incme? Only significant fr median & upper ends f the incme distributin Panel data, nt exactly i.i.d. (further research) Lngitudinal data, Yu, Sng and Härdle (2009) Time flies (technlgy level ), mre and mre high incme jbs require high educated peple. Time variatin f the ˆβ? Differential analysis t gender, labur unin status, natinalities etc? Inequality analysis t the (cnditinal) incme distributin?... Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
45 Quantnet - Open fr sharing
46 Labr Market Applicatins 5-12 Future Wrk VaR (99.9% quantile): U 1,..., U n imprtance sampling Key in prving btstrap rate: Mntne + Bunded influence! Extensin?... Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
47 Partial Linear Quantile Regressin and Btstrap Cnfidence Bands Wlfgang K. Härdle Ya acv Ritv Sng Sng Ladislaus vn Brtkiewicz Chair f Statistics C.A.S.E. - Center fr Applied Statistics and Ecnmics Humbldt-Universität zu Berlin Department f Statistics The Hebrew University f Jerusalem
48 Labr Market Applicatins 5-14 References A. Bellni and V. Chernzhukv Intersectin bunds: estimatin and inference CEMMAP Wrking Paper, 10/09, M. Buchinsky Quantile regressin, Bx-Cx transfrmatin mdel, and the U.S. wage structure, Jurnal f Ecnmetrics, 65: , V. Chernzhukv, S. Lee and A. M. Rsen L 1 -Penalized quantile regressin in high-dimensinal sparse mdels CEMMAP Wrking Paper, 19/09, Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
49 Labr Market Applicatins 5-15 References J. Franke and P. Mwita Nnparametric Estimates fr cnditinal quantiles f time series Reprt in Wirtschaftsmathematik 87, University f Kaiserslautern, P. Hall On cnvergence rates f suprema Prbab. Th. Rel. Fields, 89: , Jinyng Hahn Btstrapping Quantile Regressin Estimatrs Ecnmetric Thery, 11(1): , Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
50 Labr Market Applicatins 5-16 References W. Härdle and J. S. Marrn Btstrap simultaneus errr bars fr nnparametric regressin Ann. Statist., 19: W. Härdle and S. Sng Cnfidence bands in quantile regressin Ecnmetric Thery, frthcming, J. Hrwitz Btstrap methds fr median regressin mdels Ecnmetrica, 66: , Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
51 Labr Market Applicatins 5-17 References K. Jeng, W. Härdle and S. Sng A Cnsistent Nnparametric Test fr Causality in Quantile Revise and resubmit t Ecnmetric Thery, C. Juhn, K. M. Murphy and B. Pierce Wage Inequality and the Rise in Returns t Skill Jurnal f Plitical Ecnmy, 101(3): , R. Kenker and K. F. Hallck Quantile regressin Jurnal f Ecnmetric Perspectives, 15(4): , Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
52 Labr Market Applicatins 5-18 References J. Kmlós, P. Majr and G. Tusnády An Apprximatin f Partial Sums f Independent RV -s, and the Sample DF. I Z. Wahrscheinlichkeitstherie verw., 32: , K. Yu, Z. Lu and J. Stander Quantile regressin: applicatins and current research areas J. Ry. Statistical Sciety: Series D (The Statistician), 52: , Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
53 Appendix 6-1 Appendix - Assumptins λ i and C i : generic cnstants. A1. X 1,..., X n are an i.i.d. sample, and f X (x) λ 0. The quantile functin satisfies: l ( ) λ 1, l ( ) λ 2. A2. F (t x) have a density, f (t x) λ 3 > 0, cntinuus in x, and in t in the neighbrhd f 0. That is, fr sme A( ) and f 0 ( ) F (t x ) = p + f 0 (x)t + A(x)(x x) + R(t, x ; x), where sup t,x,x R(t,x ;x) t 2 + x x 2 <. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
54 Appendix 6-2 Nte that by Assumptin A1, l h (x) is the quantile f a discrete distributin. This distributin is equivalent t a sample size f O p (nh) frm a distributin with p-quantile whse biased is O p (h 2 ) relative t the true value. Let δ n be the lcal rate f cnvergence f the functin l h, essentially δ n = h 2 n + (nh n ) 1/2 = O(n 2/5 ), with h n = O(n 1/5 ). Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
55 Appendix 6-3 A3. The estimate l g, satisfies: sup l g (x) l (x) = Op(1), x J sup l g (x) l (x) = Op(δ n /h) x J (7) Nte that there is n S n term in (7) exactly because the bandwidth g n used t calculate l g is slightly larger than that used fr l h. As a result l g has a slightly wrse rate f cnvergence (as an estimatr f the quantile functin), but its derivatives cnverge faster. We assume: (A4). f X (x) is twice cntinuusly differentiable and f (t x) is unifrmly bunded in x and t by, say, λ 4. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
56 Appendix 6-4 (A7). The cnditinal densities f ( y), y R, are unifrmly lcal Lipschitz cntinuus f rder α (ull- α) n J, unifrmly in y R, with 0 < α 1, and (nh)/ lg n. (A8). inf v J ψ{y l(v) + ε}df (y v) q ε, fr ε δ1, where δ 1 and q are sme psitive cnstants, see als [?]. This assumptin is satisfied if there exists a cnstant q such that f {l(v) v} > q/p, x J. Partial Linear Quantile Regressin and Btstrap Cnfidence Bands
Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff
Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeff Reading: Chapter 2 STATS 202: Data mining and analysis September 27, 2017 1 / 20 Supervised vs. unsupervised learning In unsupervised
More informationLecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff
Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeff Reading: Chapter 2 STATS 202: Data mining and analysis September 27, 2017 1 / 20 Supervised vs. unsupervised learning In unsupervised
More informationDistributions, spatial statistics and a Bayesian perspective
Distributins, spatial statistics and a Bayesian perspective Dug Nychka Natinal Center fr Atmspheric Research Distributins and densities Cnditinal distributins and Bayes Thm Bivariate nrmal Spatial statistics
More informationMATCHING TECHNIQUES. Technical Track Session VI. Emanuela Galasso. The World Bank
MATCHING TECHNIQUES Technical Track Sessin VI Emanuela Galass The Wrld Bank These slides were develped by Christel Vermeersch and mdified by Emanuela Galass fr the purpse f this wrkshp When can we use
More informationCAUSAL INFERENCE. Technical Track Session I. Phillippe Leite. The World Bank
CAUSAL INFERENCE Technical Track Sessin I Phillippe Leite The Wrld Bank These slides were develped by Christel Vermeersch and mdified by Phillippe Leite fr the purpse f this wrkshp Plicy questins are causal
More informationWhat is Statistical Learning?
What is Statistical Learning? Sales 5 10 15 20 25 Sales 5 10 15 20 25 Sales 5 10 15 20 25 0 50 100 200 300 TV 0 10 20 30 40 50 Radi 0 20 40 60 80 100 Newspaper Shwn are Sales vs TV, Radi and Newspaper,
More informationSimple Linear Regression (single variable)
Simple Linear Regressin (single variable) Intrductin t Machine Learning Marek Petrik January 31, 2017 Sme f the figures in this presentatin are taken frm An Intrductin t Statistical Learning, with applicatins
More informationTying the straps tighter for generalized linear models
Tying the straps tighter for generalized linear models Ya'acov Ritov Weining Wang Wolfgang K. Härdle Ladislaus von Bortkiewicz Chair of Statistics C.A.S.E. - Center for Applied Statistics and Economics
More informationLocalized Model Selection for Regression
Lcalized Mdel Selectin fr Regressin Yuhng Yang Schl f Statistics University f Minnesta Church Street S.E. Minneaplis, MN 5555 May 7, 007 Abstract Research n mdel/prcedure selectin has fcused n selecting
More informationCHAPTER 24: INFERENCE IN REGRESSION. Chapter 24: Make inferences about the population from which the sample data came.
MATH 1342 Ch. 24 April 25 and 27, 2013 Page 1 f 5 CHAPTER 24: INFERENCE IN REGRESSION Chapters 4 and 5: Relatinships between tw quantitative variables. Be able t Make a graph (scatterplt) Summarize the
More informationBootstrap Method > # Purpose: understand how bootstrap method works > obs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(obs) >
Btstrap Methd > # Purpse: understand hw btstrap methd wrks > bs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(bs) > mean(bs) [1] 21.64625 > # estimate f lambda > lambda = 1/mean(bs);
More informationTying the straps for generalized linear models
Ya'acov Ritov Weining Wang Wolfgang K. Härdle Ladislaus von Bortkiewicz Chair of Statistics C.A.S.E. - Center for Applied Statistics and Economics Humboldt-Universität zu Berlin Department of Statistics
More informationInternal vs. external validity. External validity. This section is based on Stock and Watson s Chapter 9.
Sectin 7 Mdel Assessment This sectin is based n Stck and Watsn s Chapter 9. Internal vs. external validity Internal validity refers t whether the analysis is valid fr the ppulatin and sample being studied.
More informationMidwest Big Data Summer School: Machine Learning I: Introduction. Kris De Brabanter
Midwest Big Data Summer Schl: Machine Learning I: Intrductin Kris De Brabanter kbrabant@iastate.edu Iwa State University Department f Statistics Department f Cmputer Science June 24, 2016 1/24 Outline
More informationPartial Linear Quantile Regression and Bootstrap Confidence Bands
SFB 649 Discussion Paper 2010-002 Partial Linear Quantile Regression and Bootstrap Confidence Bands Wolfgang Karl Härdle* Ya acov Ritov** Song Song* * Humboldt-Universität zu Berlin, Germany ** Hebrew
More informationThe general linear model and Statistical Parametric Mapping I: Introduction to the GLM
The general linear mdel and Statistical Parametric Mapping I: Intrductin t the GLM Alexa Mrcm and Stefan Kiebel, Rik Hensn, Andrew Hlmes & J-B J Pline Overview Intrductin Essential cncepts Mdelling Design
More informationModelling of Clock Behaviour. Don Percival. Applied Physics Laboratory University of Washington Seattle, Washington, USA
Mdelling f Clck Behaviur Dn Percival Applied Physics Labratry University f Washingtn Seattle, Washingtn, USA verheads and paper fr talk available at http://faculty.washingtn.edu/dbp/talks.html 1 Overview
More informationCOMP 551 Applied Machine Learning Lecture 5: Generative models for linear classification
COMP 551 Applied Machine Learning Lecture 5: Generative mdels fr linear classificatin Instructr: Herke van Hf (herke.vanhf@mail.mcgill.ca) Slides mstly by: Jelle Pineau Class web page: www.cs.mcgill.ca/~hvanh2/cmp551
More informationResampling Methods. Chapter 5. Chapter 5 1 / 52
Resampling Methds Chapter 5 Chapter 5 1 / 52 1 51 Validatin set apprach 2 52 Crss validatin 3 53 Btstrap Chapter 5 2 / 52 Abut Resampling An imprtant statistical tl Pretending the data as ppulatin and
More informationINSTRUMENTAL VARIABLES
INSTRUMENTAL VARIABLES Technical Track Sessin IV Sergi Urzua University f Maryland Instrumental Variables and IE Tw main uses f IV in impact evaluatin: 1. Crrect fr difference between assignment f treatment
More informationMATCHING TECHNIQUES Technical Track Session VI Céline Ferré The World Bank
MATCHING TECHNIQUES Technical Track Sessin VI Céline Ferré The Wrld Bank When can we use matching? What if the assignment t the treatment is nt dne randmly r based n an eligibility index, but n the basis
More informationBayesian nonparametric modeling approaches for quantile regression
Bayesian nnparametric mdeling appraches fr quantile regressin Athanasis Kttas Department f Applied Mathematics and Statistics University f Califrnia, Santa Cruz Department f Statistics Athens University
More informationLyapunov Stability Stability of Equilibrium Points
Lyapunv Stability Stability f Equilibrium Pints 1. Stability f Equilibrium Pints - Definitins In this sectin we cnsider n-th rder nnlinear time varying cntinuus time (C) systems f the frm x = f ( t, x),
More information4th Indian Institute of Astrophysics - PennState Astrostatistics School July, 2013 Vainu Bappu Observatory, Kavalur. Correlation and Regression
4th Indian Institute f Astrphysics - PennState Astrstatistics Schl July, 2013 Vainu Bappu Observatry, Kavalur Crrelatin and Regressin Rahul Ry Indian Statistical Institute, Delhi. Crrelatin Cnsider a tw
More informationSmoothing, penalized least squares and splines
Smthing, penalized least squares and splines Duglas Nychka, www.image.ucar.edu/~nychka Lcally weighted averages Penalized least squares smthers Prperties f smthers Splines and Reprducing Kernels The interplatin
More informationA Matrix Representation of Panel Data
web Extensin 6 Appendix 6.A A Matrix Representatin f Panel Data Panel data mdels cme in tw brad varieties, distinct intercept DGPs and errr cmpnent DGPs. his appendix presents matrix algebra representatins
More informationResampling Methods. Cross-validation, Bootstrapping. Marek Petrik 2/21/2017
Resampling Methds Crss-validatin, Btstrapping Marek Petrik 2/21/2017 Sme f the figures in this presentatin are taken frm An Intrductin t Statistical Learning, with applicatins in R (Springer, 2013) with
More informationSURVIVAL ANALYSIS WITH SUPPORT VECTOR MACHINES
1 SURVIVAL ANALYSIS WITH SUPPORT VECTOR MACHINES Wlfgang HÄRDLE Ruslan MORO Center fr Applied Statistics and Ecnmics (CASE), Humbldt-Universität zu Berlin Mtivatin 2 Applicatins in Medicine estimatin f
More informationTree Structured Classifier
Tree Structured Classifier Reference: Classificatin and Regressin Trees by L. Breiman, J. H. Friedman, R. A. Olshen, and C. J. Stne, Chapman & Hall, 98. A Medical Eample (CART): Predict high risk patients
More informationThe Law of Total Probability, Bayes Rule, and Random Variables (Oh My!)
The Law f Ttal Prbability, Bayes Rule, and Randm Variables (Oh My!) Administrivia Hmewrk 2 is psted and is due tw Friday s frm nw If yu didn t start early last time, please d s this time. Gd Milestnes:
More informationContributions to the Theory of Robust Inference
Cntributins t the Thery f Rbust Inference by Matías Salibián-Barrera Licenciad en Matemáticas, Universidad de Buens Aires, Argentina, 1994 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
More informationIN a recent article, Geary [1972] discussed the merit of taking first differences
The Efficiency f Taking First Differences in Regressin Analysis: A Nte J. A. TILLMAN IN a recent article, Geary [1972] discussed the merit f taking first differences t deal with the prblems that trends
More informationMath 10 - Exam 1 Topics
Math 10 - Exam 1 Tpics Types and Levels f data Categrical, Discrete r Cntinuus Nminal, Ordinal, Interval r Rati Descriptive Statistics Stem and Leaf Graph Dt Plt (Interpret) Gruped Data Relative and Cumulative
More informationPSU GISPOPSCI June 2011 Ordinary Least Squares & Spatial Linear Regression in GeoDa
There are tw parts t this lab. The first is intended t demnstrate hw t request and interpret the spatial diagnstics f a standard OLS regressin mdel using GeDa. The diagnstics prvide infrmatin abut the
More informationKinetic Model Completeness
5.68J/10.652J Spring 2003 Lecture Ntes Tuesday April 15, 2003 Kinetic Mdel Cmpleteness We say a chemical kinetic mdel is cmplete fr a particular reactin cnditin when it cntains all the species and reactins
More informationPure adaptive search for finite global optimization*
Mathematical Prgramming 69 (1995) 443-448 Pure adaptive search fr finite glbal ptimizatin* Z.B. Zabinskya.*, G.R. Wd b, M.A. Steel c, W.P. Baritmpa c a Industrial Engineering Prgram, FU-20. University
More informationMeasurement Errors in Quantile Regression Models
Measurement Errrs in Quantile Regressin Mdels Sergi Firp Antni F. Galva Suyng Sng June 30, 2015 Abstract This paper develps estimatin and inference fr quantile regressin mdels with measurement errrs. We
More informationLocal Quantile Regression
Wolfgang K. Härdle Vladimir Spokoiny Weining Wang Ladislaus von Bortkiewicz Chair of Statistics C.A.S.E. - Center for Applied Statistics and Economics Humboldt-Universität zu Berlin http://ise.wiwi.hu-berlin.de
More informationEvaluating enterprise support: state of the art and future challenges. Dirk Czarnitzki KU Leuven, Belgium, and ZEW Mannheim, Germany
Evaluating enterprise supprt: state f the art and future challenges Dirk Czarnitzki KU Leuven, Belgium, and ZEW Mannheim, Germany Intrductin During the last decade, mircecnmetric ecnmetric cunterfactual
More informationINFERENCE FOR EXTREMAL CONDITIONAL QUANTILE MODELS (EXTREME VALUE INFERENCE FOR QUANTILE REGRESSION)
INFERENCE FOR EXTREMAL CONDITIONAL QUANTILE MODELS (EXTREME VALUE INFERENCE FOR QUANTILE REGRESSION) VICTOR CHERNOZHUKOV Abstract. Quantile regressin is a basic tl fr estimatin f cnditinal quantiles f
More informationOn Huntsberger Type Shrinkage Estimator for the Mean of Normal Distribution ABSTRACT INTRODUCTION
Malaysian Jurnal f Mathematical Sciences 4(): 7-4 () On Huntsberger Type Shrinkage Estimatr fr the Mean f Nrmal Distributin Department f Mathematical and Physical Sciences, University f Nizwa, Sultanate
More informationHypothesis Tests for One Population Mean
Hypthesis Tests fr One Ppulatin Mean Chapter 9 Ala Abdelbaki Objective Objective: T estimate the value f ne ppulatin mean Inferential statistics using statistics in rder t estimate parameters We will be
More information1b) =.215 1c).080/.215 =.372
Practice Exam 1 - Answers 1. / \.1/ \.9 (D+) (D-) / \ / \.8 / \.2.15/ \.85 (T+) (T-) (T+) (T-).080.020.135.765 1b).080 +.135 =.215 1c).080/.215 =.372 2. The data shwn in the scatter plt is the distance
More informationEstimation and Inference for Actual and Counterfactual Growth Incidence Curves
Discussin Paper Series IZA DP N. 10473 Estimatin and Inference fr Actual and Cunterfactual Grwth Incidence Curves Francisc H. G. Ferreira Sergi Firp Antni F. Galva January 2017 Discussin Paper Series IZA
More informationAP Statistics Notes Unit Two: The Normal Distributions
AP Statistics Ntes Unit Tw: The Nrmal Distributins Syllabus Objectives: 1.5 The student will summarize distributins f data measuring the psitin using quartiles, percentiles, and standardized scres (z-scres).
More informationCN700 Additive Models and Trees Chapter 9: Hastie et al. (2001)
CN700 Additive Mdels and Trees Chapter 9: Hastie et al. (2001) Madhusudana Shashanka Department f Cgnitive and Neural Systems Bstn University CN700 - Additive Mdels and Trees March 02, 2004 p.1/34 Overview
More informationLeast Squares Optimal Filtering with Multirate Observations
Prc. 36th Asilmar Cnf. n Signals, Systems, and Cmputers, Pacific Grve, CA, Nvember 2002 Least Squares Optimal Filtering with Multirate Observatins Charles W. herrien and Anthny H. Hawes Department f Electrical
More informationSome Theory Behind Algorithms for Stochastic Optimization
Sme Thery Behind Algrithms fr Stchastic Optimizatin Zelda Zabinsky University f Washingtn Industrial and Systems Engineering May 24, 2010 NSF Wrkshp n Simulatin Optimizatin Overview Prblem frmulatin Theretical
More informationFunctional Form and Nonlinearities
Sectin 6 Functinal Frm and Nnlinearities This is a gd place t remind urselves f Assumptin #0: That all bservatins fllw the same mdel. Levels f measurement and kinds f variables There are (at least) three
More informationx 1 Outline IAML: Logistic Regression Decision Boundaries Example Data
Outline IAML: Lgistic Regressin Charles Suttn and Victr Lavrenk Schl f Infrmatics Semester Lgistic functin Lgistic regressin Learning lgistic regressin Optimizatin The pwer f nn-linear basis functins Least-squares
More informationCOMP 551 Applied Machine Learning Lecture 4: Linear classification
COMP 551 Applied Machine Learning Lecture 4: Linear classificatin Instructr: Jelle Pineau (jpineau@cs.mcgill.ca) Class web page: www.cs.mcgill.ca/~jpineau/cmp551 Unless therwise nted, all material psted
More informationMaximum A Posteriori (MAP) CS 109 Lecture 22 May 16th, 2016
Maximum A Psteriri (MAP) CS 109 Lecture 22 May 16th, 2016 Previusly in CS109 Game f Estimatrs Maximum Likelihd Nn spiler: this didn t happen Side Plt argmax argmax f lg Mther f ptimizatins? Reviving an
More informationCS 109 Lecture 23 May 18th, 2016
CS 109 Lecture 23 May 18th, 2016 New Datasets Heart Ancestry Netflix Our Path Parameter Estimatin Machine Learning: Frmally Many different frms f Machine Learning We fcus n the prblem f predictin Want
More informationDepartment of Economics, University of California, Davis Ecn 200C Micro Theory Professor Giacomo Bonanno. Insurance Markets
Department f Ecnmics, University f alifrnia, Davis Ecn 200 Micr Thery Prfessr Giacm Bnann Insurance Markets nsider an individual wh has an initial wealth f. ith sme prbability p he faces a lss f x (0
More information[COLLEGE ALGEBRA EXAM I REVIEW TOPICS] ( u s e t h i s t o m a k e s u r e y o u a r e r e a d y )
(Abut the final) [COLLEGE ALGEBRA EXAM I REVIEW TOPICS] ( u s e t h i s t m a k e s u r e y u a r e r e a d y ) The department writes the final exam s I dn't really knw what's n it and I can't very well
More informationCONFIDENCE BANDS IN NONPARAMETRIC REGRESSION WITH LENGTH BIASED DATA
Ann. Inst. Statist. Math. Vl. 56, N. 3, 475-496 (2004) Q2004 The Institute f Statistical Mathematics CONFIDENCE BANDS IN NONPARAMETRIC REGRESSION WITH LENGTH BIASED DATA J. A. CRISTOBAL, J. L. OJEDA AND
More informationT Algorithmic methods for data mining. Slide set 6: dimensionality reduction
T-61.5060 Algrithmic methds fr data mining Slide set 6: dimensinality reductin reading assignment LRU bk: 11.1 11.3 PCA tutrial in mycurses (ptinal) ptinal: An Elementary Prf f a Therem f Jhnsn and Lindenstrauss,
More informationChapter 3: Cluster Analysis
Chapter 3: Cluster Analysis } 3.1 Basic Cncepts f Clustering 3.1.1 Cluster Analysis 3.1. Clustering Categries } 3. Partitining Methds 3..1 The principle 3.. K-Means Methd 3..3 K-Medids Methd 3..4 CLARA
More informationIntroduction to Regression
Intrductin t Regressin Administrivia Hmewrk 6 psted later tnight. Due Friday after Break. 2 Statistical Mdeling Thus far we ve talked abut Descriptive Statistics: This is the way my sample is Inferential
More informationYou need to be able to define the following terms and answer basic questions about them:
CS440/ECE448 Sectin Q Fall 2017 Midterm Review Yu need t be able t define the fllwing terms and answer basic questins abut them: Intr t AI, agents and envirnments Pssible definitins f AI, prs and cns f
More informationinitially lcated away frm the data set never win the cmpetitin, resulting in a nnptimal nal cdebk, [2] [3] [4] and [5]. Khnen's Self Organizing Featur
Cdewrd Distributin fr Frequency Sensitive Cmpetitive Learning with One Dimensinal Input Data Aristides S. Galanpuls and Stanley C. Ahalt Department f Electrical Engineering The Ohi State University Abstract
More informationk-nearest Neighbor How to choose k Average of k points more reliable when: Large k: noise in attributes +o o noise in class labels
Mtivating Example Memry-Based Learning Instance-Based Learning K-earest eighbr Inductive Assumptin Similar inputs map t similar utputs If nt true => learning is impssible If true => learning reduces t
More informationDifferentiation Applications 1: Related Rates
Differentiatin Applicatins 1: Related Rates 151 Differentiatin Applicatins 1: Related Rates Mdel 1: Sliding Ladder 10 ladder y 10 ladder 10 ladder A 10 ft ladder is leaning against a wall when the bttm
More informationThree-Step Censored Quantile Regression and Extramarital Affairs
Victr Chernzhukv and Han Hng Three-Step Censred Quantile Regressin and Extramarital Affairs This article suggests very simple three-step estimatrs fr censred quantile regressin mdels with a separatin restrictin
More informationLECTURE NOTES. Chapter 3: Classical Macroeconomics: Output and Employment. 1. The starting point
LECTURE NOTES Chapter 3: Classical Macrecnmics: Output and Emplyment 1. The starting pint The Keynesian revlutin was against classical ecnmics (rthdx ecnmics) Keynes refer t all ecnmists befre 1936 as
More informationFunctional Form Diagnostics for Cox s Proportional Hazards Model
Functinal Frm Diagnstics fr Cx s Prprtinal Hazards Mdel Larry F. León Department f Bistatistics, Harvard University, Bstn, MA 02115, U.S.A. and Chih-Ling Tsai Graduate Schl f Management, University f Califrnia,
More informationCS 477/677 Analysis of Algorithms Fall 2007 Dr. George Bebis Course Project Due Date: 11/29/2007
CS 477/677 Analysis f Algrithms Fall 2007 Dr. Gerge Bebis Curse Prject Due Date: 11/29/2007 Part1: Cmparisn f Srting Algrithms (70% f the prject grade) The bjective f the first part f the assignment is
More informationNonparametric Quantile Estimations For Dynamic Smooth Coefficient Models
Nnparametric Quantile Estimatins Fr Dynamic Smth Cefficient Mdels Zngwu Cai and Xiaping Xu This Versin: January 13, 2006 In this paper, quantile regressin methds are suggested fr a class f smth cefficient
More informationInstrumental Variables Quantile Regression for Panel Data with Measurement Errors
Instrumental Variables Quantile Regressin fr Panel Data with Measurement Errrs Antni Galva University f Illinis Gabriel Mntes-Rjas City University Lndn Tky, August 2009 Panel Data Panel data allws the
More informationSIZE BIAS IN LINE TRANSECT SAMPLING: A FIELD TEST. Mark C. Otto Statistics Research Division, Bureau of the Census Washington, D.C , U.S.A.
SIZE BIAS IN LINE TRANSECT SAMPLING: A FIELD TEST Mark C. Ott Statistics Research Divisin, Bureau f the Census Washingtn, D.C. 20233, U.S.A. and Kenneth H. Pllck Department f Statistics, Nrth Carlina State
More information, which yields. where z1. and z2
The Gaussian r Nrmal PDF, Page 1 The Gaussian r Nrmal Prbability Density Functin Authr: Jhn M Cimbala, Penn State University Latest revisin: 11 September 13 The Gaussian r Nrmal Prbability Density Functin
More informationCOMP 551 Applied Machine Learning Lecture 11: Support Vector Machines
COMP 551 Applied Machine Learning Lecture 11: Supprt Vectr Machines Instructr: (jpineau@cs.mcgill.ca) Class web page: www.cs.mcgill.ca/~jpineau/cmp551 Unless therwise nted, all material psted fr this curse
More informationQUANTILE AUTOREGRESSION
QUANTILE AUTOREGRESSION ROGER KOENKER AND ZHIJIE XIAO Abstract. We cnsider quantile autregressin (QAR) mdels in which the autregressive cefficients can be expressed as mntne functins f a single, scalar
More informationDiscussion on Regularized Regression for Categorical Data (Tutz and Gertheiss)
Discussin n Regularized Regressin fr Categrical Data (Tutz and Gertheiss) Peter Bühlmann, Ruben Dezeure Seminar fr Statistics, Department f Mathematics, ETH Zürich, Switzerland Address fr crrespndence:
More informationAP Statistics Practice Test Unit Three Exploring Relationships Between Variables. Name Period Date
AP Statistics Practice Test Unit Three Explring Relatinships Between Variables Name Perid Date True r False: 1. Crrelatin and regressin require explanatry and respnse variables. 1. 2. Every least squares
More informationInference in the Multiple-Regression
Sectin 5 Mdel Inference in the Multiple-Regressin Kinds f hypthesis tests in a multiple regressin There are several distinct kinds f hypthesis tests we can run in a multiple regressin. Suppse that amng
More informationComparing Several Means: ANOVA. Group Means and Grand Mean
STAT 511 ANOVA and Regressin 1 Cmparing Several Means: ANOVA Slide 1 Blue Lake snap beans were grwn in 12 pen-tp chambers which are subject t 4 treatments 3 each with O 3 and SO 2 present/absent. The ttal
More informationThe blessing of dimensionality for kernel methods
fr kernel methds Building classifiers in high dimensinal space Pierre Dupnt Pierre.Dupnt@ucluvain.be Classifiers define decisin surfaces in sme feature space where the data is either initially represented
More informationHomology groups of disks with holes
Hmlgy grups f disks with hles THEOREM. Let p 1,, p k } be a sequence f distinct pints in the interir unit disk D n where n 2, and suppse that fr all j the sets E j Int D n are clsed, pairwise disjint subdisks.
More informationSUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical model for microarray data analysis
SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical mdel fr micrarray data analysis David Rssell Department f Bistatistics M.D. Andersn Cancer Center, Hustn, TX 77030, USA rsselldavid@gmail.cm
More informationWe say that y is a linear function of x if. Chapter 13: The Correlation Coefficient and the Regression Line
Chapter 13: The Crrelatin Cefficient and the Regressin Line We begin with a sme useful facts abut straight lines. Recall the x, y crdinate system, as pictured belw. 3 2 1 y = 2.5 y = 0.5x 3 2 1 1 2 3 1
More informationPattern Recognition 2014 Support Vector Machines
Pattern Recgnitin 2014 Supprt Vectr Machines Ad Feelders Universiteit Utrecht Ad Feelders ( Universiteit Utrecht ) Pattern Recgnitin 1 / 55 Overview 1 Separable Case 2 Kernel Functins 3 Allwing Errrs (Sft
More informationFebruary 28, 2013 COMMENTS ON DIFFUSION, DIFFUSIVITY AND DERIVATION OF HYPERBOLIC EQUATIONS DESCRIBING THE DIFFUSION PHENOMENA
February 28, 2013 COMMENTS ON DIFFUSION, DIFFUSIVITY AND DERIVATION OF HYPERBOLIC EQUATIONS DESCRIBING THE DIFFUSION PHENOMENA Mental Experiment regarding 1D randm walk Cnsider a cntainer f gas in thermal
More informationGMM with Latent Variables
GMM with Latent Variables A. Rnald Gallant Penn State University Raffaella Giacmini University Cllege Lndn Giuseppe Ragusa Luiss University Cntributin The cntributin f GMM (Hansen and Singletn, 1982) was
More information7 TH GRADE MATH STANDARDS
ALGEBRA STANDARDS Gal 1: Students will use the language f algebra t explre, describe, represent, and analyze number expressins and relatins 7 TH GRADE MATH STANDARDS 7.M.1.1: (Cmprehensin) Select, use,
More informationAccreditation Information
Accreditatin Infrmatin The ISSP urges members wh have achieved significant success in the field t apply fr higher levels f membership in rder t enjy the fllwing benefits: - Bth Prfessinal members and Fellws
More informationEASTERN ARIZONA COLLEGE Introduction to Statistics
EASTERN ARIZONA COLLEGE Intrductin t Statistics Curse Design 2014-2015 Curse Infrmatin Divisin Scial Sciences Curse Number PSY 220 Title Intrductin t Statistics Credits 3 Develped by Adam Stinchcmbe Lecture/Lab
More informationLecture 13: Markov Chain Monte Carlo. Gibbs sampling
Lecture 13: Markv hain Mnte arl Gibbs sampling Gibbs sampling Markv chains 1 Recall: Apprximate inference using samples Main idea: we generate samples frm ur Bayes net, then cmpute prbabilities using (weighted)
More informationENSC Discrete Time Systems. Project Outline. Semester
ENSC 49 - iscrete Time Systems Prject Outline Semester 006-1. Objectives The gal f the prject is t design a channel fading simulatr. Upn successful cmpletin f the prject, yu will reinfrce yur understanding
More informationEdinburgh Research Explorer
Edinburgh Research Explrer Supprting User-Defined Functins n Uncertain Data Citatin fr published versin: Tran, TTL, Dia, Y, Suttn, CA & Liu, A 23, 'Supprting User-Defined Functins n Uncertain Data' Prceedings
More informationQuantile Regression for Dynamic Panel Data
Quantile Regressin fr Dynamic Panel Data Antni F. Galva University f Illinis at Urbana-Champaign June 03, 2008 Abstract This paper studies estimatin and inference in a quantile regressin dynamic panel
More informationON-LINE PROCEDURE FOR TERMINATING AN ACCELERATED DEGRADATION TEST
Statistica Sinica 8(1998), 207-220 ON-LINE PROCEDURE FOR TERMINATING AN ACCELERATED DEGRADATION TEST Hng-Fwu Yu and Sheng-Tsaing Tseng Natinal Taiwan University f Science and Technlgy and Natinal Tsing-Hua
More informationMATHEMATICS SYLLABUS SECONDARY 5th YEAR
Eurpean Schls Office f the Secretary-General Pedaggical Develpment Unit Ref. : 011-01-D-8-en- Orig. : EN MATHEMATICS SYLLABUS SECONDARY 5th YEAR 6 perid/week curse APPROVED BY THE JOINT TEACHING COMMITTEE
More informationAEC 874 (2007) Field Data Collection & Analysis in Developing Countries. VII. Data Analysis & Project Documentation
AEC 874 (2007) Field Data Cllectin & Analysis in Develping Cuntries VII. Data Analysis & Prject Dcumentatin Richard H. Bernsten Agricultural Ecnmics Michigan State University 1 A. Things t Cnsider in Planning
More informationModelling Record Times in Sport with Extreme Value Methods
utput 2017/5/9 5:13 page 1 #1 Malaysian Jurnal f Mathematical Sciences *(*): ** ** (201*) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES Jurnal hmepage: http://einspem.upm.edu.my/jurnal Mdelling Recrd Times
More information1 The limitations of Hartree Fock approximation
Chapter: Pst-Hartree Fck Methds - I The limitatins f Hartree Fck apprximatin The n electrn single determinant Hartree Fck wave functin is the variatinal best amng all pssible n electrn single determinants
More informationA Scalable Recurrent Neural Network Framework for Model-free
A Scalable Recurrent Neural Netwrk Framewrk fr Mdel-free POMDPs April 3, 2007 Zhenzhen Liu, Itamar Elhanany Machine Intelligence Lab Department f Electrical and Cmputer Engineering The University f Tennessee
More informationSAMPLING DYNAMICAL SYSTEMS
SAMPLING DYNAMICAL SYSTEMS Melvin J. Hinich Applied Research Labratries The University f Texas at Austin Austin, TX 78713-8029, USA (512) 835-3278 (Vice) 835-3259 (Fax) hinich@mail.la.utexas.edu ABSTRACT
More informationModule 3: Gaussian Process Parameter Estimation, Prediction Uncertainty, and Diagnostics
Mdule 3: Gaussian Prcess Parameter Estimatin, Predictin Uncertainty, and Diagnstics Jerme Sacks and William J Welch Natinal Institute f Statistical Sciences and University f British Clumbia Adapted frm
More informationElements of Machine Intelligence - I
ECE-175A Elements f Machine Intelligence - I Ken Kreutz-Delgad Nun Vascncels ECE Department, UCSD Winter 2011 The curse The curse will cver basic, but imprtant, aspects f machine learning and pattern recgnitin
More information