【独家发布】Bayesian Cognitive Modeling

0关注
62粉丝

VIP

已卖：4194份资源

院士

67%

还不是VIP/贵宾

-

TA的文库 其他...

Bayesian NewOccidental

Spatial Data Analysis

东西方数据挖掘

0%

威望: 0 级
论坛币: 50288 个
通用积分: 83.6306
学术水平: 253 点
热心指数: 300 点
信用等级: 208 点
经验: 41518 点
帖子: 3256
精华: 14
在线时间: 766 小时
注册时间: 2006-5-4
最后登录: 2022-11-6

楼主

Lisrelchen 发表于 2014-11-30 00:06:34 |AI写论文

是否 +2 论坛币

k人参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

You can download

http://bayesmodels.com/
http://www.cambridge.org/us/academic/subjects/psychology/psychology-research-methods-and-statistics/bayesian-cognitive-modeling-practical-course

扫码加我拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏7 回帖

关键词：Cognitive Bayesian Modeling model Bayes

本帖被以下文库推荐

· Bayesian NewOccidental|主题: 578, 订阅: 78

沙发

Lisrelchen 发表于 2014-11-30 00:07:32

%% Inferring A Rate
clear;
sampler = 0; % Choose 0=WinBUGS, 1=JAGS
%% Data
k = 5;
n = 10;
%% Sampling
% MCMC Parameters
nchains = 2; % How Many Chains?
nburnin = 0; % How Many Burn-in Samples?
nsamples = 1e4; %How Many Recorded Samples?
nthin = 1; % How Often is a Sample Recorded?
doparallel = 0; % Parallel Option
% Assign Matlab Variables to the Observed Nodes
datastruct = struct('k',k,'n',n);
% Initialize Unobserved Variables
for i=1:nchains
S.theta = rand; % An Intial Value for the Success Rate
init0(i) = S;
end
if ~sampler
% Use WinBUGS to Sample
tic
[samples, stats] = matbugs(datastruct, ...
fullfile(pwd, 'Rate_1.txt'), ...
'init', init0, ...
'nChains', nchains, ...
'view', 1, 'nburnin', nburnin, 'nsamples', nsamples, ...
'thin', nthin, 'DICstatus', 0, 'refreshrate',100, ...
'monitorParams', {'theta'}, ...
'Bugdir', 'C:/Program Files/WinBUGS14');
toc
else
% Use JAGS to Sample
tic
fprintf( 'Running JAGS ...\n' );
[samples, stats] = matjags( ...
datastruct, ...
fullfile(pwd, 'Rate_1.txt'), ...
init0, ...
'doparallel' , doparallel, ...
'nchains', nchains,...
'nburnin', nburnin,...
'nsamples', nsamples, ...
'thin', nthin, ...
'monitorparams', {'theta'}, ...
'savejagsoutput' , 1 , ...
'verbosity' , 1 , ...
'cleanup' , 0 , ...
'workingdir' , 'tmpjags' );
toc
end;
%% Analysis
figure(1);clf;hold on;
eps = .01; binsc = eps/2:eps:1-eps/2; binse = 0:eps:1;
count = histc(reshape(samples.theta,1,[]),binse);
count = count(1:end-1);
count = count/sum(count)/eps;
ph = plot(binsc,count,'k-');
set(gca,'box','on','fontsize',14);
xlabel('Rate','fontsize',16);
ylabel('Posterior Density','fontsize',16);

复制代码

藤椅

fjrong

发表于 2014-11-30 00:09:24 来自手机

Lisrelchen 发表于 2014-11-30 00:06
You can download
the first two parts of the book
the associated code

谢谢分享

板凳

Lisrelchen 发表于 2014-11-30 00:09:42

%% Difference Between Two Rates
clear;
sampler = 0; % Choose 0=WinBUGS, 1=JAGS
%% Data (Observed Variables)
k1 = 0;
n1 = 5;
k2 = 5;
n2 = 10;
%% Sampling
% MCMC Parameters
nchains = 3; % How Many Chains?
nburnin = 0; % How Many Burn-in Samples?
nsamples = 5e4; %How Many Recorded Samples?
nthin = 1; % How Often is a Sample Recorded?
doparallel = 0; % Parallel Option
% Assign Matlab Variables to the Observed Nodes
datastruct = struct('k1',k1,'n1',n1,'k2',k2,'n2',n2);
%Initialize Unobserved Variables
for i=1:nchains
S.theta1 = 0.5; % An Intial Value for the Success Rate
S.theta2 = 0.5; % An Intial Value for the Success Rate
init0(i) = S;
end
if ~sampler
% Use WinBUGS to Sample
tic
[samples, stats] = matbugs(datastruct, ...
fullfile(pwd, 'Rate_2.txt'), ...
'init', init0, ...
'nChains', nchains, ...
'view', 1, 'nburnin', nburnin, 'nsamples', nsamples, ...
'thin', nthin, 'DICstatus', 0, 'refreshrate',100, ...
'monitorParams', {'theta1','theta2','delta'}, ...
'Bugdir', 'C:/Program Files/WinBUGS14');
toc
else
% Use JAGS to Sample
tic
fprintf( 'Running JAGS ...\n' );
[samples, stats] = matjags( ...
datastruct, ...
fullfile(pwd, 'Rate_2.txt'), ...
init0, ...
'doparallel' , doparallel, ...
'nchains', nchains,...
'nburnin', nburnin,...
'nsamples', nsamples, ...
'thin', nthin, ...
'monitorparams', {'theta1','theta2','delta'}, ...
'savejagsoutput' , 1 , ...
'verbosity' , 1 , ...
'cleanup' , 0 , ...
'workingdir' , 'tmpjags' );
toc
end;
%% Analysis
% Plot Posterior
figure(2);clf;hold on;
eps = .025; binsc = -1+eps/2:eps:1-eps/2; binse = -1:eps:1;
count = histc(reshape(samples.delta,1,[]),binse);
count = count(1:end-1);
count = count/sum(count)/eps;
ph = plot(binsc,count,'k-');
set(gca,'box','on','fontsize',14,'xtick',[-1:.2:1],'ytick',[1:ceil(max(get(gca,'ylim')))]);
xlabel('Difference in Rates','fontsize',16);
ylabel('Posterior Density','fontsize',16);
% Summaries of Posterior
% MEAN
disp(sprintf('Mean is %1.2f',stats.mean.delta));
% MODE
[~,ind] = max(count);
disp(sprintf('Mode is %1.2f',binsc(ind)));
% MEDIAN
disp(sprintf('Median is %1.2f',median(reshape(samples.delta,1,[]))));
% CREDIBLE INTERVAL
cred = 0.95;
b1 = (1-cred)/2;b2=1-b1;
val = sort(reshape(samples.delta,1,[]));
disp(sprintf('%d percent credible interval is [%1.2f, %1.2f]',cred*100,val(round(b1*nsamples*nchains)),val(round(b2*nsamples*nchains))));

复制代码

报纸

Lisrelchen 发表于 2014-11-30 00:12:52

%% Inferring A Common Rate
clear;
sampler = 0; % Choose 0=WinBUGS, 1=JAGS
%% Data
k1 = 5;
n1 = 10;
k2 = 7;
n2 = 10;
%% Sampling
% MCMC Parameters
nchains = 2; % How Many Chains?
nburnin = 0; % How Many Burn-in Samples?
nsamples = 1e4; %How Many Recorded Samples?
nthin = 1; % How Often is a Sample Recorded?
doparallel = 0; % Parallel Option
% Assign Matlab Variables to the Observed Nodes
datastruct = struct('k1',k1,'n1',n1,'k2',k2,'n2',n2);
% Initialize Unobserved Variables
for i=1:nchains
S.theta = 0.5; % An Intial Value for the Success Rate
init0(i) = S;
end
if ~sampler
% Use WinBUGS to Sample
tic
% Use WinBUGS to Sample
[samples, stats] = matbugs(datastruct, ...
fullfile(pwd, 'Rate_3.txt'), ...
'init', init0, ...
'nChains', nchains, ...
'view', 1, 'nburnin', nburnin, 'nsamples', nsamples, ...
'thin', nthin, 'DICstatus', 0, 'refreshrate',100, ...
'monitorParams', {'theta'}, ...
'Bugdir', 'C:/Program Files/WinBUGS14');
toc
else
% Use JAGS to Sample
tic
fprintf( 'Running JAGS ...\n' );
[samples, stats] = matjags( ...
datastruct, ...
fullfile(pwd, 'Rate_3.txt'), ...
init0, ...
'doparallel' , doparallel, ...
'nchains', nchains,...
'nburnin', nburnin,...
'nsamples', nsamples, ...
'thin', nthin, ...
'monitorparams', {'theta'}, ...
'savejagsoutput' , 1 , ...
'verbosity' , 1 , ...
'cleanup' , 0 , ...
'workingdir' , 'tmpjags' );
toc
end;
%% Analysis
figure(3);clf;hold on;
eps = .01; binsc = eps/2:eps:1-eps/2; binse = 0:eps:1;
count = histc(reshape(samples.theta,1,[]),binse);
count = count(1:end-1);
count = count/sum(count)/eps;
ph = plot(binsc,count,'k-');
set(gca,'box','on','fontsize',14,'xtick',[0:.2:1],'ytick',[1:ceil(max(get(gca,'ylim')))]);
xlabel('Rate','fontsize',16);
ylabel('Posterior Density','fontsize',16);

复制代码

地板

Lisrelchen 发表于 2014-11-30 00:14:41

%% Prior and Posterior Predictive
clear;
sampler = 0; % Choose 0=WinBUGS, 1=JAGS
%% Data
dataset = 1;
switch dataset
case 1, k = 1; n = 15; % Toy data
case 2, k = 24; n = 121; % Trouw nursing Home Data
end;
%% Sampling
% MCMC Parameters
nchains = 2; % How Many Chains?
nburnin = 1e2; % How Many Burn-in Samples?
nsamples = 1e4; %How Many Recorded Samples?
nthin = 1; % How Often is a Sample Recorded?
doparallel = 0; % Parallel Option
% Assign Matlab Variables to the Observed Nodes
datastruct = struct('k',k,'n',n);
% Initialize Unobserved Variables
for i=1:nchains
S.theta = 0.5; % Intial Value
init0(i) = S;
end
if ~sampler
% Use WinBUGS to Sample
tic
[samples, stats] = matbugs(datastruct, ...
fullfile(pwd, 'Rate_4.txt'), ...
'init', init0, ...
'nChains', nchains, ...
'view', 1, 'nburnin', nburnin, 'nsamples', nsamples, ...
'thin', nthin, 'DICstatus', 0, 'refreshrate',100, ...
'monitorParams', {'theta','thetaprior','postpredk','priorpredk'}, ...
'Bugdir', 'C:/Program Files/WinBUGS14');
toc
else
% Use JAGS to Sample
tic
fprintf( 'Running JAGS ...\n' );
[samples, stats] = matjags( ...
datastruct, ...
fullfile(pwd, 'Rate_4.txt'), ...
init0, ...
'doparallel' , doparallel, ...
'nchains', nchains,...
'nburnin', nburnin,...
'nsamples', nsamples, ...
'thin', nthin, ...
'monitorparams', {'theta','thetaprior','postpredk','priorpredk'}, ...
'savejagsoutput' , 1 , ...
'verbosity' , 1 , ...
'cleanup' , 0 , ...
'workingdir' , 'tmpjags' );
toc
end;
%% Analysis
figure(4);clf;
% Parameter Space
subplot(211);hold on;
eps=.015;binse=[0:eps:1];binsc=[eps/2:eps:1-eps/2];
count=histc(reshape(samples.thetaprior,1,[]),binse);
count=count(1:end-1);
count=count/sum(count)/eps;
ph=plot(binsc,count,'k--');
count=histc(reshape(samples.theta,1,[]),binse);
count=count(1:end-1);
count=count/sum(count)/eps;
ph=plot(binsc,count,'k-');
set(gca,'xlim',[0 1],'box','on','fontsize',14,'xtick',[0:.2:1]);
legend('Prior','Posterior');
set(gca,'box','on','fontsize',14);
xlabel('Rate','fontsize',16);
ylabel('Density','fontsize',16);
% Data Space
subplot(212);hold on;
kbins=[0:n];
count1=histc(samples.priorpredk,kbins);
count1=count1/sum(count1);
count2=histc(samples.postpredk,kbins);
count2=count2/sum(count2);
switch dataset
case 1,
ph=bar(kbins,count1,.6);set(ph,'facecolor','none','linewidth',1.5,'linestyle','--')
ph=bar(kbins,count2,.8);set(ph,'facecolor','none')
set(gca,'xlim',[-1 n+1],'box','on','fontsize',14,'xtick',[0:n]);
otherwise,
ph=bar(kbins,count1,.6);set(ph,'facecolor','none','linewidth',0.5,'linestyle',':')
ph=bar(kbins,count2);set(ph,'facecolor','none');
set(gca,'xlim',[-1 n+1],'box','on','fontsize',14,'xtick',[0 n],'ticklength',[0 0]);
end;
[lh oh] = legend('Prior','Posterior');
set(lh,'box','off');
xlabel('Success Count','fontsize',16);
ylabel('Mass','fontsize',16);

复制代码

7楼

Lisrelchen 发表于 2014-11-30 00:22:23

%% Prior and Posterior Predictive, Second Example
clear;
sampler = 0; % Choose 0=WinBUGS, 1=JAGS
%% Data
k1 = 0;
n1 = 10;
k2 = 10;
n2 = 10;
%% Sampling
% MCMC Parameters
nchains = 2; % How Many Chains?
nburnin = 1e2; % How Many Burn-in Samples?
nsamples = 1e4; %How Many Recorded Samples?
nthin = 1; % How Often is a Sample Recorded?
doparallel = 0; % Parallel Option
% Assign Matlab Variables to the Observed Nodes
datastruct = struct('k1',k1,'k2',k2,'n1',n1,'n2',n2);
% Initialize Unobserved Variables
for i=1:nchains
S.theta = 0.5; % Intial Value
init0(i) = S;
end
if ~sampler
% Use WinBUGS to Sample
tic
[samples, stats] = matbugs(datastruct, ...
fullfile(pwd, 'Rate_5.txt'), ...
'init', init0, ...
'nChains', nchains, ...
'view', 1, 'nburnin', nburnin, 'nsamples', nsamples, ...
'thin', nthin, 'DICstatus', 0, 'refreshrate',100, ...
'monitorParams', {'theta','postpredk1','postpredk2'}, ...
'Bugdir', 'C:/Program Files/WinBUGS14');
toc
else
% Use JAGS to Sample
tic
fprintf( 'Running JAGS ...\n' );
[samples, stats] = matjags( ...
datastruct, ...
fullfile(pwd, 'Rate_5.txt'), ...
init0, ...
'doparallel' , doparallel, ...
'nchains', nchains,...
'nburnin', nburnin,...
'nsamples', nsamples, ...
'thin', nthin, ...
'monitorparams', {'theta','postpredk1','postpredk2'}, ...
'savejagsoutput' , 1 , ...
'verbosity' , 1 , ...
'cleanup' , 0 , ...
'workingdir' , 'tmpjags' );
toc
end;
%% Analysis
figure(5);clf;
set(gcf,'units','norm','pos',[.2 .2 .6 .4],'paperpositionmode','auto');
% Parameter Space
subplot(121);hold on;
axis square;
eps=.015;bins=[0:eps:1];binsc=[eps/2:eps:1-eps/2];
count=histc(reshape(samples.theta,1,[]),bins);
count=count(1:end-1);
count=count/sum(count)/eps;
ph=plot(binsc,count,'k-');
set(gca,'xlim',[0 1],'box','on','fontsize',14,'xtick',[0:.2:1],'ytick',[1:ceil(max(get(gca,'ylim')))]);
set(gca,'box','on','fontsize',14);
xlabel('Rate','fontsize',16);
ylabel('Density','fontsize',16);
% Data Space
subplot(122);hold on;
axis equal;
axis([-1 n1+1 -1 n2+1]);
sc=70;
for i=0:n1
for j=0:n2
match=length(find(samples.postpredk1==i&samples.postpredk2==j))/nsamples/nchains;
if match>0
ph=plot(i,j,'ks');
set(ph,'markersize',sc*sqrt(match));
end;
end;
end;
ph=plot(k1,k2,'kx');
set(ph,'markersize',16,'linewidth',4);
set(gca,'box','on','fontsize',14,'xtick',[0:n1],'ytick',[0:n2]);
xlabel('Success Count 1','fontsize',16);
ylabel('Success Count 2','fontsize',16);

复制代码

8楼

oliyiyi 发表于 2014-11-30 09:52:08

谢谢分享

9楼

fjrong

发表于 2014-11-30 09:52:18 来自手机

Lisrelchen 发表于 2014-11-30 00:06
You can download
the first two parts of the book
the associated code

谢谢分享

10楼

meng山楂树 发表于 2014-12-1 07:16:12

[休闲其它] 【独家发布】Bayesian Cognitive Modeling [推广有奖]

经管之家送您一份

经管之家联合CDA

感谢您参与论坛问题回答

扫码加我拉你入群

相关帖子

本帖被以下文库推荐

浏览过的帖子

浏览过的版块

初级热心勋章

中级热心勋章

初级学术勋章

初级信用勋章

中级信用勋章

中级学术勋章

高级热心勋章

高级学术勋章

高级信用勋章

特级热心勋章

特级学术勋章

特级信用勋章

本版微信群

[休闲其它] 【独家发布】Bayesian Cognitive Modeling [推广有奖]

经管之家送您一份

经管之家联合CDA

感谢您参与论坛问题回答

扫码加我 拉你入群

相关帖子

本帖被以下文库推荐

浏览过的帖子

浏览过的版块

初级热心勋章

中级热心勋章

初级学术勋章

初级信用勋章

中级信用勋章

中级学术勋章

高级热心勋章

高级学术勋章

高级信用勋章

特级热心勋章

特级学术勋章

特级信用勋章

本版微信群

扫码加我拉你入群