maeforecast.dfm2: Economic Forecasting with High-Dimensional Data (Dynamic...

In google-trends-v1/gtm: Functions for Economic Nowcasting with Google Trends Data

Description

Like its counterpart maeforecast.dfm, this function makes out-of-sample forecasts based on a dynamic factor model. The difference is in the way dynamic factors are extracted. The maeforecast.dfm2 function first implements a clustering process to the covariate time series based on the partitional method, and one dynamic factor is then extracted within each cluster either based on the two-step method proposed by Doz, Gianone & Reichlin (2011) or by aggregation.

Usage

maeforecast.dfm2(data, w_size, window="recursive", y.index=1,
            factor.num=3, method="two-step", clustor.type="partitional",
            h=0, t.select, t.update=F)

Arguments

data	a data frame or a matrix; the first column should contain the time series variable for which the forecasts are to be made. Other columns should contain the covariates.
w_size	numeric, indicating the index where the forecasting should begin. If the first point forecast should be made at the 73th observation, for example, w_size should be set to be 72.
window	character, indicating the forecasting scheme to be applied. Options include "recursive", "rolling", and "fixed".
y.index	numeric, indicating the column position of the time series for which the forecasts are made (Y). Defualt is 1.
t.select	number of covariates to be included. If omitted, every covariate will be included. Otherwise, a regression between the dependant variable, its lag and each covariate will be run and a statistical test will be applied for the significance of the covariate's coefficient. The covariates will then be ranked based on their test statistics, and t.select number of them with the highest test statistics will be included in the machine learning algorithms. Note that the forecasting horizon h is considered.
t.update	logical, indicating wheter the preselection process should be repeated in evert iteration, if t.select is specified. Under the defualt FALSE, the preselection process will be implemented only with for the first window.
h	forecasting horizon. Default is 0.
factor.num	numeric, indicating the number of dynamic factors to be extracted from the covariates in the Dynamic Factor Model. Default is 3.
method	character, indicating which method should be used to extract dynamic factors. If "aggregation", the covariate time series are first clustered based on partitional method, and a simple aggreagtion is applied to each cluster. If "two-step", one factor is extracted within each cluster based on Doz, Giannone & Reichlin (2011).
clustyor.type	the type of clustering method to be applied. Options include "partitional", "hierarchical", "tadpole", and "fuzzy". Default is "partitional".

Value

Forecasts	data matrix, containing the point forecasts, realized values, forecast errors, signs of the forecasts and realized values, and success in predicting the signs.
MSE	numeric, mean squred error of the point forecasts.
SRatio	numeric, success ratio of the point forecasts. Success is claimed when the point forecasts and realized values have the same sign.
Data	the data as used in the model.
Model	some specifics about the model used.

Author(s)

Zehua Wu

See Also

maeforecast.dfm

google-trends-v1/gtm documentation built on June 5, 2019, 5:13 p.m.