Base Metals Price Forecasting from Alternate Data Sources using AI

25 May, 2020 by NExT

Project Name: Base Metals Price Forecasting from Alternate Data Sources using AI

Project Sponsor: AISG and 4Elements Pte Ltd

Grant Amount: S$230,800.00

Duration: 18 months

Description: The future trend of base metal prices, which is affected by a wide range of factors, such as supply and demand and macro-economic, is crucial to us. In this project, NUS and 4 Elements Capital Pte Ltd (4ECap) aims to leverage machine learning techniques to learn and discover relevant knowledge from large scale alternate data sources for making predictions of base metal prices. Notably, we build a neural network-based solution for predicting the future contract prices of base metals (Copper, Nickel, Tin, Aluminum, Lead, and Zinc) traded on the London Metal Exchange with the following data sources:

  1. Market data (trading prices and volumes) of the selected base metals
  2. Macro-economic data
  3. Supply and demand data
  4. Third-party estimations
  5. Relevant social media information

The project team develops a solution that can achieve averaged Directional Correctness per horizon and across commodities is above 60% for at least three horizons, and the minimum for all horizons is above 52.5%. As a result, we improve the prediction quality of future contract prices of base metals profitable for the base metal trading of 4ECap and other market participants. Besides, the prediction can be further commercialized as a service for traders of base metal commodities and their associated derivatives.

External Projects

Sea State Identification

25 May, 2020

Singapore Restaurant Food Image Annotation

25 May, 2020

Measuring Physical Profile and Use of PCN with Deep Learning and MM Data Analytics

25 May, 2020

Base Metals Price Forecasting from Alternate Data Sources using AI

25 May, 2020

Deep Learning for Recommendation System

25 May, 2020

Multilingual Multimodal Knowledge Graph (M3KG) and Its Application in Smart Nation

25 May, 2020

Visual Saliency and Reward Saliency Guide Decision Making

25 May, 2020

A GPU-based Engine for Dynamic Sparse Matrix Analytics

25 May, 2020

You Are What You Surf: Characterizing Preference and Decision-Making Styles with Digital Footprints

21 May, 2020

Back