https://github.com/siddeshsambasivam/a-comparative-study-on-sciml-for-kd
This repository is the official implementation of "A Comparative Study on Machine Learning Algorithms for Knowledge Discovery."
https://github.com/siddeshsambasivam/a-comparative-study-on-sciml-for-kd
knowledge-discovery machine-learning sparse-regression symbolic-regression
Last synced: about 2 months ago
JSON representation
This repository is the official implementation of "A Comparative Study on Machine Learning Algorithms for Knowledge Discovery."
- Host: GitHub
- URL: https://github.com/siddeshsambasivam/a-comparative-study-on-sciml-for-kd
- Owner: SiddeshSambasivam
- License: mit
- Created: 2022-07-15T02:33:30.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2022-10-18T03:20:39.000Z (over 2 years ago)
- Last Synced: 2025-02-05T16:48:57.953Z (3 months ago)
- Topics: knowledge-discovery, machine-learning, sparse-regression, symbolic-regression
- Language: Jupyter Notebook
- Homepage:
- Size: 13.1 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# A Comparative Study on Machine Learning Algorithms for Knowledge Discovery
This repository is the official implementation of **"A Comparative Study on Machine Learning Algorithms for Knowledge Discovery."**
> :rocket: **Update:** The paper has been accepted for 17th International Conference on Control, Automation, Robotics and Vision (ICARCV 2022).
#### **Table of Contents**
[A. Requirements](#a-requirements)
[B. Dataset](#b-datasets)
[C. Baselines](#c-baselines)
[D. Benchmark](#d-benchmark)
[E. Results](#e-results)
[F. Citation](#f-citation)
**Overview:** The paper aims to summarize key research works in symbolic regression and perform a comparative study to understand the _strengths_ and _limitations_ of each method. Finally, we **highlight the challenges** in the current methods and future research directions in the application of machine learning in knowledge discovery.
### **A. Requirements**
- Install [miniconda](https://docs.conda.io/en/main/miniconda.html) to manage experiments' dependencies.
To install requirements:
```bash
conda env create -f environment.yml
```
### **B. Datasets**
1. **_Feynman-03:_** All equations with up to `3 input variables` from the AI-Feynman dataset were sampled. The resulting dataset contained 52 equations.
2. **_Nguyen-12:_** The datasets consisted of 12 equations with a maximum of `2 input variables`. It is important to highlight that few equations contain terms such as `x^6` and `x^5` which were included to test the methods’ ability to understand high-frequency terms.
### **C. Baselines**
1. **_Genetic programming (GPL)_**: A python library called gplearn was used to perform genetic programming (GP).
2. **_Deep symbolic regression (DSR)_**: An auto-regressive approach based on reinforcement learning search to optimize the symbolic search space. The opensource implementation of (Petersen et. el.) with default parameters was used for the benchmark tasks.
3. **_AIFeynman (AIF)_**: A heuristic based search approach that uses recurring patterns in symbolic formulas describing natural phenomena. A custom wrapper was written around the open-source python package, AIFeynman, for the evaluation.
4. **_Neural Symbolic Regression that Scales (NeSymRes)_**: A symbolic language modelling approach that was pretrained on a large distribution of millions of equations. The pretrained model on 100 million equations was used for the benchmark.
### **D. Benchmark**
The baseline models can be benchmarked using the following command and arguments:
- Models: `gpl`, `dsr`, `aif`, `nesymres`
- Datasets: `feynman`, `nguyen`
```bash
make - noise= num_points=
```
For example, to run the benchmark for the `gpl` model on the `feynman` dataset with `noise=0.1` and `num_points=1000`, run:
```bash
make feynman-gpl noise=0.1 num_points=1000
```
### **E. Results**
Figure 1: Effect of noise on accuracy in Feynman-03 dataset
Figure 2: Effect of noise on accuracy in Nguyen-12 dataset
Other results can be found in the **results and discussion** section of the paper.
### **F. Citation**
If you find this work useful in your research, please consider citing:
```
@inproceedings{,
title={A Comparative Study on Machine Learning Algorithms for Knowledge Discovery},
author={Siddesh Sambasivam Suseela, Yang Feng, Kezhi Mao},
booktitle={17th International Conference on Control, Automation, Robotics and Vision (ICARCV 2022)},
year={2022},
organization={Nanyang Technological University}
}
```
### **G. Contact**
For any questions, please contact Siddesh Sambasivam Suseela ([email protected])