Assessing similarities between gold ores, concentrates, and tailings with case-based reasoning

Research output: Contribution to journalArticleScientificpeer-review

Researchers

Research units

  • University of West London
  • VTT Technical Research Centre of Finland

Abstract

Searching for information has always been a challenge for a researcher. Decades ago it was a matter of gaining access to the information, but in this day and age, the problem is more about how to find the most relevant information amongst the ever-growing amount of scientific literature readily available. This research article presents an information retrieval tool that is designed to compare and rank gold raw materials, such as ores, concentrates, and tailings, according to the user's research interests. The tool was constructed based on interviews with metallurgists specialized in gold and by utilizing a knowledge modeling technique called case-based reasoning. The tool can compare a gold raw material to existing material descriptions based on six attributes: Material type, Ore type, Mineralogy, Chemical assay, Particle size, and Gold content, resulting in a list of scientific articles that utilized similar raw materials. Different features in the software used enable the user to take their preferences into account and modify the effect of each attribute on the end result, i.e., the articles retrieved. The tool's function was demonstrated with four test queries that ranked the entire case base of 84 gold raw material descriptions in relation to the test queries. While the tool's default settings were able to retrieve similar materials from the case base, the best results are obtained when the user can modify the basis of retrieval in accordance with their particular research interests.

Details

Original languageEnglish
Article number106113
JournalMinerals Engineering
Volume146
Publication statusPublished - 15 Jan 2020
MoE publication typeA1 Journal article-refereed

    Research areas

  • Gold raw materials, Information retrieval, Knowledge modeling, Set similarity

ID: 39178243