Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Metal Price Spike Prediction via a Neurosymbolic Ensemble Approach

Nathaniel Lee, Noel Ngu, Harshdeep Singh Sahdev, Pramod Motaganahall, Al Mehdi Saadat Chowdhury, Bowen Xi, Paulo Shakarian

TL;DR

This work introduces a neurosymbolic ensemble framework that integrates multiple neural models with symbolic error detection and correction rules to enhance predictive accuracy by correcting individual model errors and offering interpretability through rule-based explanations.

Abstract

Predicting price spikes in critical metals such as Cobalt, Copper, Magnesium, and Nickel is crucial for mitigating economic risks associated with global trends like the energy transition and reshoring of manufacturing. While traditional models have focused on regression-based approaches, our work introduces a neurosymbolic ensemble framework that integrates multiple neural models with symbolic error detection and correction rules. This framework is designed to enhance predictive accuracy by correcting individual model errors and offering interpretability through rule-based explanations. We show that our method provides up to 6.42% improvement in precision, 29.41% increase in recall at 13.24% increase in F1 over the best performing neural models. Further, our method, as it is based on logical rules, has the benefit of affording an explanation as to which combination of neural models directly contribute to a given prediction.

Metal Price Spike Prediction via a Neurosymbolic Ensemble Approach

TL;DR

This work introduces a neurosymbolic ensemble framework that integrates multiple neural models with symbolic error detection and correction rules to enhance predictive accuracy by correcting individual model errors and offering interpretability through rule-based explanations.

Abstract

Predicting price spikes in critical metals such as Cobalt, Copper, Magnesium, and Nickel is crucial for mitigating economic risks associated with global trends like the energy transition and reshoring of manufacturing. While traditional models have focused on regression-based approaches, our work introduces a neurosymbolic ensemble framework that integrates multiple neural models with symbolic error detection and correction rules. This framework is designed to enhance predictive accuracy by correcting individual model errors and offering interpretability through rule-based explanations. We show that our method provides up to 6.42% improvement in precision, 29.41% increase in recall at 13.24% increase in F1 over the best performing neural models. Further, our method, as it is based on logical rules, has the benefit of affording an explanation as to which combination of neural models directly contribute to a given prediction.

Paper Structure

This paper contains 16 sections, 2 equations, 1 figure, 3 tables, 3 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. EDCR For Metal Price Prediction
  4. EDCR Rule Learning Algorithms (recap of ref_bowen_edcr)
  5. EDCR Applied to MPSC
  6. Modified EDCR Rule Learner for MPSC
  7. Experiments
  8. Experimental Settings
  9. Dataset and Setup
  10. Machine Learning Models
  11. Hardware and Implementation
  12. Model Evaluation
  13. Ablation Studies
  14. Conclusion
  15. Acknowledgment
  16. ...and 1 more sections

Figures (1)

  • Figure 1: Ablation analysis of EDCR on ATTN (1) and model (left) and CNN (5) model (right) performance metrics for predicting Magnesium and Cobalt price spikes respectively. The bar charts demonstrate varying Precision and Recall tradeoffs when different models are used as rules for the base model.