Download presentation

Presentation is loading. Please wait.

Published byAndrea Sainsbury Modified over 2 years ago

1
A Conditionally Parametric Probit Model of Micro-Data Land Use in Chicago Daniel McMillen Maria Soppelsa

2
Overview Residential v. Commercial/Industrial Land Use in Chicago, 2010 A conditionally parametric (CPAR) approach produces smooth estimates over space Target points chosen using an adaptive decision tree approach (Loader, 1999) Interpolation from 182 target points to all 583,063 individual parcels in the data set

3
Estimation Procedures Case (1992). Special From for W McMillen (1992). EM Algorithm Pinkse and Slade (1998). GMM for spatial error model. LeSage (2000). Bayesian approach Klier and McMillen (2007). Linearized version of GMM probit/logit for spatial AR model.

4
GMM Probit

5
Linearized GMM Probit

6
CPAR Probit

7
Spatial AR v. LWR

8
Data Individual parcels in Chicago, 2010 Major Classes: 1.Vacant Land (33,139) 2.Residential, 6 units or fewer (728,541, 539,975 after geocoding) 3.Multi-Family Residential (11,529) 4.Non-Profit (316) 5.Commercial and Industrial (50,508, 43,088 after geocoding) 6.“Incentive Classes” (1,487)

9
Explanatory Variables Distance from parcel centroid to: 1.CBD 2.Lake Michigan 3.EL line 4.EL stop 5.Rail line 6.Major street 7.Park 8.Highway

10
Rogers Park

11
Descriptive Statistics VariableMeanStd. Dev.MinMax Residential Lot0.9260.2620.0001.000 Distance from CBD7.5183.4330.02217.006 Distance from Lake Michigan4.1162.7160.00512.321 Distance from EL Line1.3581.2770.0016.265 Distance from EL Stop1.2141.0810.0016.265 Distance from Rail Line0.4280.2940.0011.997 Distance from Major Street0.0800.0570.0000.508 Distance from Park0.2330.1530.0002.999 Distance from Highway1.4761.0270.0114.809

12
Probit Models, Probability Residential Standard ProbitCPAR Probit VariableCoef.Std. ErrorMeanStd. Dev. Intercept0.0610.0460.3511.008 Distance from CBD0.1320.0070.1010.266 Distance from Lake Michigan-0.0950.007-0.0860.308 Distance from EL Line0.0020.013-0.4231.168 Distance from EL Stop-0.0910.0130.5111.263 Distance from Rail Line0.6260.0140.6490.686 Distance from Major Street8.7480.07011.5706.427 Distance from Park-1.0990.020-0.8810.994 Distance from Highway0.2120.0070.0480.351 Log-likelihood-131518.9-120714.1 Pseudo-R 2 0.1440.215

13
Probability of Residential Land Use: Standard Probit

14
Probability of Residential Land Use: CPAR Probit, 10% Window Size

15
Difference, CPAR Probability – Standard Probit Probability

16
Kernel Density Estimates for CPAR Coefficients

17
LWR Estimates of CPAR Coefficients

18
Marginal Probabilities

26
Rogers Park

27
Rogers Park, n = 3,193 StandardGMMCPAR CoefStd. Err.CoefStd. Err.MeanStd. dev. Intercept49.97911.99942.97712.5920.0252.445 CBD-1.8040.462-1.5490.480 Lake Michigan-7.6211.672-6.5551.814-0.7265.314 EL Line-3.3240.651-2.9010.723-4.4499.934 EL Stop3.1270.6542.6980.7396.5939.706 Rail Line1.9060.3951.6590.4281.6754.059 Major Street7.1230.8375.9921.34615.9009.561 Park-1.7970.514-1.5940.525 Highway-7.2071.743-6.1971.809 Metra Stop0.0380.2160.0240.178 ρ0.1550.167 pseudo-R 2 0.084 0.343

28
Correlations, Predicted Probabilities StandardGMMCPAR Standard10.570.99 GMM0.571 CPAR0.990.571

29
Standard Probit Probabilities

30
CPAR Probit Probabilities

32
Standard Probit: Southwest

33
CPAR – Standard: Southwest

34
Standard Probit: Southeast

35
CPAR – Standard: Southeast

36
Standard Probit: Northwest

37
CPAR – Standard: Northwest

38
Standard Probit: Northeast

39
CPAR – Standard: Southeast

Similar presentations

OK

Bayesian inference Lee Harrison York Neuroimaging Centre 01 / 05 / 2009.

Bayesian inference Lee Harrison York Neuroimaging Centre 01 / 05 / 2009.

© 2017 SlidePlayer.com Inc.

All rights reserved.

Ads by Google