-
Notifications
You must be signed in to change notification settings - Fork 13
/
Copy pathvc_pack.sthlp
85 lines (63 loc) · 3.29 KB
/
vc_pack.sthlp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
{smcl}
{* *! version 1.0.0 February 2020}{...}
{vieweralsosee "vc_bw" "help vc_bw"}{...}
{vieweralsosee "vc_bwalt" "help vc_bw"}{...}
{vieweralsosee "vc_reg" "help vc_reg"}{...}
{vieweralsosee "vc_preg" "help vc_preg"}{...}
{vieweralsosee "vc_bsreg" "help vc_bsreg"}{...}
{vieweralsosee "vc_graph" "help vc_graph"}{...}
{vieweralsosee "vc_predict" "help vc_predict"}{...}
{vieweralsosee "cv_regress" "help cv_regress"}{...}
{title:Title}
{phang}
{bf:vc_pack} {hline 2} Modules for the model selection, estimation, and visualization of Smooth Varying Coefficient Models (SVCM).
{marker description}{...}
{title:Description}
{pstd}
Non-parametric regressions are powerful statistical tools that can be used to model
relationships between dependent and independent variables with minimal assumptions on
the underlying functional forms.
{pstd}
Despite its potential benefits, these types of models have two weaknesses:
{pstd}
- The added flexibility creates a curse of dimensionality,
{pstd}
- And procedures available for model selection, like cross-validation, have a high computational
cost in samples with even moderate sizes.
{pstd}
An alternative to fully-nonparametric models are semiparametric models that combine the flexibility
of non-parametric regressions with the structure of standard models.
{pstd}
This package estimates a particular type of semiparametric models known as Smooth Varying Voefficient Models
(Hastie and Tibshirani 1993), based on kernel regression methods, assuming a single smoothing variable.
{pstd}
These commands aim to facilitate bandwidth selection, model estimation, implementation of specification tests, and create visualizations of the results.
{pstd}
In this package you will find the following commands:
{pstd}
{help vc_bw} and {help vc_bwalt} are commands used for the bandwidth selection, using two maximization methods for the crossvalidation procedure.
{pstd}
{help vc_reg}, {help vc_bsreg} and {help vc_preg} are commands used for the estimation of the SVCM, based on different strategies for the estimation of the Variance Covariance matrix of the coefficients.
{pstd}
{help vc_predict} is a commands used for the the estimation of predicted values, predicted errors, and leave-one-out errors.
It also provides basic summary statistics for the SVCM, and performs the approximate F test for model specification.
{pstd}
{help vc_test} is a commands used to implement the J-statistic specification test, using a wild bootstrap procedure.
{pstd}
{help vc_graph} is a commands used to obtain plots of the estimated smooth coefficients after the model has been estimated using
{help vc_reg}, {help vc_bsreg} or {help vc_preg}.
{pstd}
For details on the commands, please refer to Rios-Avila(2020).
{marker Author}{...}
{title:Author}
{pstd}Fernando Rios-Avila{break}
Levy Economics Institute of Bard College{break}
Blithewood-Bard College{break}
Annandale-on-Hudson, NY{break}
{marker references}{...}
{title:References}
{pstd}
Hastie, Trevor, and Robert Tibshirani. 1993. "Varying-Coefficient Models." Journal of the Royal Statistical Society. Series B (Methodological) 55 (4):757-796.
{pstd}
Rios-Avila, Fernando (2020) Smooth varying coefficient models in Stata. Working paper. {browse "https://drive.google.com/open?id=1dkd-NTsiZjzl8JGImegxfuOe4FZ1YsQ4":vc_reg paper}