R quantmod buildModel()不适合模型

Question

当使用 quantmod 拟合模型时，我无法避免错误，buildModel 似乎不起作用，因此使用 tradeModel 会出现错误:

getSymbols('GS',src='yahoo')

q.model = specifyModel(Next(OpCl(GS)) ~ Lag(OpHi(GS),0:3))

buildModel(q.model, method='lm',training.per=c('2007-08-01','2007-09-30'))
> q.model # it seems that model is still 'None fitted' ?

quantmod object:        Build date:   

Model Specified: 
     Next(OpCl(GS)) ~ Lag(OpHi(GS), 0:3) 

Model Target:  Next.OpCl.GS              Product:  GS 
Model Inputs:   

Fitted Model: 

        None Fitted

tradeModel(q.model,plot.model=TRUE,trade.dates=c("2008-01-01","2008-12-31"))

Error in UseMethod("predict") : 
  no applicable method for 'predict' applied to an object of class "NULL"

Answer 1

好的，这是一个“平均”文档。对于 getSymbols() 函数似乎接管了对变量的赋值，但对于 buildModel() 却没有，即使 ?buildModel 中的示例可能表明。简而言之:您需要将 buildModel() 的结果分配给一个变量，例如q.模型。我希望你能对此大笑 ;)

library(quantmod)
getSymbols(Symbols = 'GS', src = 'yahoo')
q.model <- specifyModel(Next(OpCl(GS)) ~ Lag(OpHi(GS), 0:3))
q.model <- buildModel(q.model, method = 'lm', training.per = c('2007-08-01','2007-09-30'))
q.model