Learn R Programming

⚠️There's a newer version (2.1.1) of this package.Take me there.

qrnn (version 2.0.1)

Quantile Regression Neural Network

Description

Fit quantile regression neural network models with optional left censoring, partial monotonicity constraints, generalized additive model constraints, and the ability to fit multiple non-crossing quantile functions following Cannon (2011) and Cannon (2017) .

Copy Link

Version

Install

install.packages('qrnn')

Monthly Downloads

784

Version

2.0.1

License

GPL-2

Maintainer

Alex Cannon

Last Published

December 11th, 2017

Functions in qrnn (2.0.1)

Derivative of the approximation to the ramp function

Modified generalized additive model plots for interpreting QRNN models

Huber norm function

Approximation to the ramp function

Exponential linear function

Radial basis function kernel

qrnn.initialize

Initialize a QRNN weight vector

Derivative of the exponential linear function

Evaluate quantiles from trained QRNN model

composite.stack

Reformat data matrices for composite quantile regression

Derivative of the Huber norm function

Identity function

Convert a factor to a matrix of dummy codes

Smooth approximation to the tilted absolute value cost function

Main function used to fit a QRNN model or ensemble of QRNN models

Daily precipitation data at Vancouver Int'l Airport (YVR)

Sigmoid function

Derivative of the sigmoid function

Derivative of the linear function

Fit and make predictions from QRNN models with two hidden layers

Monotone composite quantile regression neural network (MCQRNN)

Interpolated quantile distribution with exponential tails

A hybrid mean/median function for left censored variables

Quantile Regression Neural Network

Approximation to the rectified linear function

Derivative of the approximation to the rectified linear function

tilted.approx.prime

Derivative of the approximate tilted absolute value function

Tilted absolute value function

Approximation to the tilted absolute value function