Ordinal Data Modeling: d1=plotfit(g,b,cov) - File Exchange

No BSD License

Highlights from
Ordinal Data Modeling

gpdisc(call) This is the machine-generated representation of a MATLAB object
Mb=b_probg(data,m); B_PROBIT Produces a simulated sample from a probit regression model
[Mb,accept]=breg_bay(data,m,l... B_LOGIT Produces a simulated sample from a logistic regression model
[Mbeta,Ms2,stds,sample]=logit... fit of random effects model for conduct example of Chapter 3
[Zij,Z,S,Cats,B,Sr,accept]=sa... sampleReg returns a sample from the posterior on the
[av,gv,th_m,th_s,ath,aag]=l_i... item_r - fits a 2-parameter logistic item response model of the form
[av,gv,th_m,th_s,av_m,av_s2]=... item_r_h - fits a 2-parameter probit item response model of the form
[av,gv,th_m,th_s]=item_r(y,s_... item_r - fits a 2-parameter probit item response model of the form
[beta,s]=bay_reg(y,X,num) % BAY_REG Simulated sample from the posterior for a normal regression model.
[beta,var,fp,dev_df,pr,dr,adr... MLE Finds maximum likelihood estimates for a binary regression model
[beta,var,lint]=cmp(data,mKx) CMP Computation of posterior mode and log integral using Laplace method.
[bf,post1,post2]=mod_crit(mod... % MOD_CRIT Model criticism for discrete models.
[bv,th_m,th_s]=item_r1(y,s_b,... item_r - fits a 1-parameter probit item response model of the form
[lint2,mom,x,y1,f]=ad_quad2(l... AD_QUAD2 Summarizes a two-parameter posterior by adaptive quadrature.
[log_scores,sz]=llatent(Mb,da... LLATENT logistic scores plot for latent residuals in logistic
[m_sim,s_sim]=m_cont(data,num) % M_CONT Simulated sample from the posterior for normal data, vague prior.
[mode,var,log_int]=laplace(lo... LAPLACE summarizes a posterior density by the Laplace method
[p,mids]=irtobs(data,th_m,bin... % irtobs - computes observed proportion of correct
[p1,p2,prior]=pp_prior(dat,ty... % PP_PRIOR Construction of prior for two proportions using discrete models.
[p1,p2]=pp_exch(t,data,num) % PP_EXCH Posterior distribution for two proportions using an exchangeable prior.
[post,diff_dist]=pp_disc(p1,p... % PP_DISC Posterior distribution for two proportions using discrete models.
[pr,lo,hi]=irtpost(av,gv,th) % irtpost - computes 5th, 50th, 95th percentiles of
[set,eprob]=disc_int(dist,pro... % DISC_INT Computes a highest probability interval for a discrete distribution.
[spost,mom,x,f]=ad_quad1(logp... AD_QUAD1 Summarizes a one-parameter posterior by adaptive quadrature.
[summ_fitted,summ_resid]=lfit... LFITTED Summarizes posterior distribution for fitted probabilities
[v_th,arate]=metrop(logpost,t... METROP - simulates from a 1-parameter posterior using Metropolis algorithm
bf=c_table(y,a) % C_TABLE Bayes factor for testing independence in a contingency table.
d1=plotfit(g,b,cov)
fitted=lfitted2(Mb,cov,link) LFITTED2 Posterior distribution for fitted probabilities for logistic model.
norm_plot(data) Computes the standard normal quantile function of the vector x, 0
ordinalMLE(data,K,link) ordinalMLE computes the MLE for ordinal regression of N on X.
par=beta_sel(r,rplus) % BETA_SEL Selecting a beta prior using predictive statements.
par=normal_s(perc1,perc2) % NORMAL_S Finds normal distribution which matches two percentiles.
plotfitted(covariate,fitmatri... PLOTFITTED - produces errorbar plot of many distributions
post=bayes(prior,like,data) % BAYES Bayes' rule for a independent sequence of outcomes.
post=m_disc(m,prior,data) % M_DISC Posterior distribution for a normal mean with discrete models.
post=mod_cont(model,llike_fcn... % MOD_CONT Posterior distribution for continuous models.
post=mod_disc(model,prior,lli... % MOD_DISC Posterior distribution for discrete models.
post=p_disc(p,prior,data) % P_DISC Posterior distribution for a proportion with discrete models.
post=p_hist_p(mids,probs,data... % P_HIST_P Posterior distribution for a proportion using a histogram prior.
pred=p_beta_p(ab,n,s) % P_BETA_P Predictive distribution of number of successes in future binomial
pred=p_disc_p(p,probs,n,s) % P_DISC_P Predictive distribution of number of successes in future binomial
quan=o_post_pred(data,k,sampl... Phi computes the standard normal distribution function value at x
roc(N,D,TmT,sampSize,alpha,la... roc returns a sample from the posterior on the
sample=post_pred(Mb,data) POSTPRED Posterior predictive distribution of standard deviation(y*)
sampleMulti(N,sampSize,alpha,... sampleOrdProb returns a sample from the posterior on the
sampleOrdInform(N,X,mle,sampS... sampleOrdInform, patterned after sampleOrdProb,
sampleOrdProb(data,K,mle,samp... sampleOrdProb returns a sample from the posterior on the
summaries=plotpost(Mb,plotsum... PLOTPOSTS Summarizes posterior distribution for multivariate simulated sample
v_th=gibbs(logpost,th_0,m,sca... GIBBS- simulates from a posterior using Gibbs sampling
val=logpost1(T,data)
val=logpost2(xy,data)
val=logpost2(xy,data)
val=logpost3(xy,data)
val=phi(x)
val=rbeta(alpha,beta,n) rbeta(alpha,beta,n) generates a vector of n Beta(alpha,beta) variates
values=m_norm_t(m0,prob,t,dat... % M_NORM_T Performs a test that a normal mean is equal to a specific value.
values=p_beta_t(p0,prob,ab,da... % P_BETA_T Performs a test that a proportion is equal to a specific value.
values=pp_bet_t(prob,parH,par... % PP_BET_T Test of the equality of two proportions using beta priors.
y=ir_curve(theta,a,g)
Contents.m
contents.m
contents.m
contents.m
contents.m
data_analysis.m
example.m
example1.m
example1.m
example1.m
example1.m
example1.m
example1.m
example1.m
example1.m
example1.m
example1.m
example1.m
example2.m
example2.m
example2.m
example2.m
example2.m
example2.m
example2.m
example2.m
example2.m
example2.m
example2.m
example3.m
example3.m
example3.m
example3.m
example3.m
example4.m
example4.m
example4.m
example4.m
example5.m
example5.m
example6.m
example7.m
View all files

from Ordinal Data Modeling by Valen Johnson
Companion Software

d1=plotfit(g,b,cov)

function d1=plotfit(g,b,cov)

c=length(g);
g=[0 g];
n=length(cov);

probs=[]; k=length(g);
for i=k:-1:1
   probs=[probs;g(i)-b(1)-b(2)*cov];
end

probs=(exp(probs)./(1+exp(probs)))';
probs=[ones(n,1) probs zeros(n,1)];
d=-diff(probs')';
d1=d(:,5:-1:1);
bar(cov,d1,'stacked')
xlabel('SAT'),ylabel('PROBABILITY')
legend('F','D','C','B','A')
colormap('gray')

Highlights from Ordinal Data Modeling

Highlights from
Ordinal Data Modeling