A.I.-based Diamond price estimation

Scientific project

Diamonds are a physical asset traded on specialised markets. Natural diamonds are formed under conditions of very high temperature and pressure in the earth’s mantle and are extracted in mines after being lifted by magma (eruption). The characteristics of rough diamond crystals vary greatly depending on the environment (chemical composition, temperature, pressure) in which they are produced. Thus, two crystals mined side by side will have different characteristics that can strongly influence its price.

In this context, we think it would be interesting to set up automatic diamond price estimation methods based on A.I. algorithms. These linear or non-linear models can capture fine interactions and may allow to understand the non-explicit formula followed by gemologists who estimate the price of diamonds. We will of course base ourselves at least on the 4Cs and other characteristics available in the database provided. The first descriptive study will allow us to understand the correlations between the explanatory variables and the parameters that make up the price or variations in the price of a diamond.

 

The objectives of the project are to:

  • Understand the determinants of diamond prices;
  • Understand the evolution of diamond price drivers;
  • Develop a machine learning algorithm to automatically estimate diamond prices from various characteristics.

 

Scientific officers

Jean-Michel Beacco
Jean-Michel Beacco
See CV
Louis Boulanger
Louis Boulanger
See CV

Economic Partner