softmax.fun: Adjusted softmax transformation function

In xqnwang/fuma: Forecast uncertainty based on model averaging (FUMA)

Description

In mathematics, softmax function maps a K-dimensional arbitrary real vector to another K-dimensional probability vector. Based on the properties of softmax function, the fitted values obtained from the pre-trained algorithm for each time series are transformed into probabilities. According to the adjusted softmax function, the method with a smaller score value in the candidate methods for each time series has a larger probability than the other methods. Consequently, the appropriate methods for each time series can be picked based on the obtained probabilities.

Usage

softmax.fun(msis)

Arguments

x	a dataframe with n rows and F columns, where n is the number of observations and F is the number of candidate methods.

xqnwang/fuma documentation built on May 29, 2021, 6:38 a.m.