Overview

This repository provides R Source codes to reproduce numerical experiments in the following manuscript:

@article{okuno2024N3POM,
    year      = {2024},
    publisher = {Taylor \& Francis},
    volume    = {33},
    number    = {},
    issue     = {4},
    pages     = {1454-1463},
    author    = {Akifumi Okuno and Kazuharu Harada},
    title     = {An Interpretable Neural Network-based Nonproportional Odds Model for Ordinal Regression},
    journal   = {Journal of Computational and Graphical Statistics}
}

Main scripts

0-demonstration.R

You can train N3POM and plot the estimated coefficient functions (for one instance). Please replace the dataset if you want (default: real-estate). Our computer, equipped with an AMD Ryzen 7 5700X 8-Core Processor operating at 3.4GHz and 32 GB of RAM, took approximately 75 seconds to train the N3POM in this script.

1-synthetic.R

This script computes the synthetic dataset experiments.

2-real.R

This script computes the real dataset experiments.

Contact info.

Akifumi Okuno (okuno@ism.ac.jp)