测试-训练R中空间模型的分裂预测

Question

所以我需要为空间误差模型创建测试火车分割预测。

我们在80％的观测值中对模型进行估算，然后让模型预测20％，但是我无法对其余20％的预测进行预测

因此，我想寻求任何建议

示例：

beforeEach(() => {
  moxios.install()
  moxios.stubRequest(
    'http://jsonplaceholder.typicode.com/comments',
    {
      status: 200,
      response: [{ name: 'Fetched #1' }, { name: 'Fetched #2' }]
    }
  )
}) 

it('should fetch list of comments and display them', (done) => {
  const component = mount(
    <Root>
      <App />
    </Root>
  )


  component.find('#fetch-comments').simulate('click')

  moxios.wait(() => {
    component.update()
    expect(component.find('li').length).toEqual(2)

    done()
    component.unmount()
  }, 500)
});

返回错误，但是线性回归 library(spdep) data(oldcol) index_train <- sample(1:nrow(COL.OLD), round(0.80 * nrow(COL.OLD))) df_train <- COL.OLD[index_train, ] df_test <- COL.OLD[-index_train, ] # Creating W matrix for train subset CORD = cbind(COL.OLD$X, COL.OLD$Y) cns <- knearneigh(CORD[index_train, ], k=4, longlat=T) scnsn <- knn2nb(cns, row.names = NULL, sym = T) W <- nb2listw(scnsn) model <- lagsarlm(CRIME ~ INC + HOVAL, data=df_train, W) summary(model) # Creating network matrix for test subset cns <- knearneigh(CORD[-index_train, ], k=4, longlat=T) scnsn <- knn2nb(cns, row.names = NULL, sym = T) W <- nb2listw(scnsn) predict(model, newdata = df_test, listw = W) 的预测非常简单