使用TreeBUGS的一般程序
![](data:image/png;base64,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)
在最简单的场景中,需要执行以下步骤:
- 以.eqn格式定义现有MPT模型文件的路径(参见multiTree;莫斯科,2010年)
- 使用单个频率定义数据集的路径(.csv文件:逗号分隔,行=人,列=标记的类别)
- 调用其中一个拟合函数
βMPT
或牵引MPT
(示例如下)
- 检查MCMC链的收敛性
- 总结并绘制结果
下面将更详细地解释这些步骤。请注意TreeBUGS需要JAGS软件的最新版本(https://mcmc-jags.sourceforge.io网站/).
1.步骤:EQN语法中的MPT模型文件
模型需要以标准.eqn文件格式传递(例如。,如multiTree;莫斯科,2010年)。例如,考虑简单的双高阈值模型(2HTM),每行定义一个包含树标签、类别标签和模型的处理路径方程:
#####标题:2HTM目标命中Do目标命中(1-Do)*g目标未命中(1-Do)*(1-g)诱饵FA(1-Dn)*g诱饵CR(1-Dn)*(1-g)勒尔CR Dn
请注意,类别标签(例如,命中、未命中…)必须以letter(不同于multiTree或HMMTree),并与的列名匹配.模型方程需要乘号*
和不应总结参数,例如a^2*(1-a)
.作为TreeBUGS的输入,模型文件(例如。,“2htm.txt”
)需要保存在当前工作目录中。否则必须指定文件的相对或绝对路径(例如。,型号/2htm.txt
). 检查TreeBUGS如何解释给定的.eqn文件,使用:
读取EQN(
文件= “文件路径.eqn”,#相对路径或绝对路径
限制= 列表(“Dn=完成”),#等式约束
param顺序= 真的
)#显示参数顺序
MPT参数的相等限制可以在中提供列表:
或通过硬盘上文本文件的路径(例如。,限制=“pathToFile.txt”
)包含等式的约束,每行一个:
Dn=执行克=0.5
2.步骤:具有单个频率的数据集
可以从中以逗号分隔的文本文件(.csv)加载数据以下格式:
Hit、Miss、FA、CR20, 10, 5, 2513, 7, 9, 21 15, 5, 6, 14.....
请注意,第一行包含类别标签,必须匹配.egn文件中的类别标签。其余行包含单个频率。与.eqn文件类似数据文件可以指定为数据输入.csv
如果是的话在当前工作目录中,或作为相对或绝对路径(例如。,“C:/models/data_ind.csv”
).
在R中使用TreeBUGS时,数据帧或矩阵可以提供与类别标签匹配的列名。
3.步骤:拟合层次MPT模型
等级制度Beta-MPT模型配有以下代码:
#加载程序包:
图书馆(树型无人值守地面传感器)
#拟合模型:
fit分级MPT<- βMPT(
eqnfile文件= “2htm.txt”,#.egn文件
数据= “data_ind.csv”,#个别数据
限制= 列表(“Dn=完成”),#参数限制(或文件路径)
###可选MCMC输入:
n.iter公司= 20000,#迭代次数
n.燃烧= 5000,#移除的烧入样本数
n.薄= 5,#去除样品的稀释率
n.链条= 三 #MCMC链的数量(并联运行)
)
类似地,可以通过更换βMPT
通过列车PT
.
4.步骤:检查MCMC链的收敛性
功能βMPT
和牵引MPT
返回a包含MCMC采样器原始样品的列表收敛检查。MCMC样品储存在装配模型$runjags$mcmc
作为mcmc.列表
对象(有关收敛诊断的概述,请参阅包)。TreeBUGS提供了一个方便的包装器来访问最重要的绘图功能:
#默认值:迹线图和密度
情节(fit分级MPT,#拟合模型
参数= “意思是” #要绘制哪个参数
)
#将进一步的参数传递给?绘图.mcmc.list
#自相关图:
情节(fitHierarchicalMPT,参数= “意思是”,类型= “acf”)
#Gelman-Rubin图:
情节(fit分级MPT,参数= “意思是”,类型= “盖尔曼”)
有关进一步的收敛统计信息和图表,请参阅和。请注意如果Markov-Chain Monte-Carlo(MCMC)采样器未收敛!
5.步骤:汇总并绘制结果
TreeBUGS生成MPT定制的参数估计和收敛统计:
- 关于参数后验分布的信息:平均值、标准差、,中位数、2.5%和97.5%分位数
- 汇聚:\(右上角)-统计学(所有参数应接近1.00,例如。,\(R<1.05))和有效数量考虑自相关时的样本数(应该很大)
要在拟合模型后获得摘要,只需使用:
以下函数允许绘制参数估计值,分布、拟合优度和原始频率:
绘图参数(fit分级MPT,#估计参数
包括个人= 真的 #是否绘制个人估计
)
plot分布(fit分级MPT)#估计的层次参数分布
绘图适配(fit分级MPT)#观测平均频率与预测平均频率
绘图适配(fit分级MPT,统计= “cov”)#观测与预测协方差
绘图频率(fit分级MPT)#每棵树的单个和平均原始频率
plotPriorPost(打印前立柱)(fit分级MPT)#前后对比(组级参数)
可以提取参数估计值(后验均值、中位数、SD)使用以下命令保存到文件:
#R中进一步使用的矩阵:
tt公司<- 获取参数(fit分级MPT,
参数= “θ”,
统计= “意思是”
)
tt公司
#将个人估算的完整摘要保存到文件中:
获取参数(fit分级MPT,
参数= “θ”,
统计= “摘要”,文件= “参数.csv”
)