blog banner


e-commerce Diamonds jewelry

Machine learning for Diamond Pricing

Over the last several weeks TransPacific Software is working on various Linear and Neural Network models using Google's TensorFlow AI library to achieve fairly accurate diamond pricing tool. This tool can be used to find the fair price of the diamond in the current B2B market.

Complexity in Pricing model:

Sometimes to a consumer diamond prices appear illogical. For example, a 0.5 carat diamond has price double than the 1.25 Carat Diamond on the same website. The consumer is unaware of numerous factors affecting diamond pricing apart from carat and shape. The intricacies in diamond pricing lie in multiple factors including 4C (carat, cut, clarity and colour) as well as its optical properties like Table, Depth, Crown and Pavilion angle. Such complexities make "Diamond Pricing" an ideal case for Artificial Intelligence

Using the diamond trading data from various P2P trading exchanges like Rapnet , Polygon and IDEXonline ; multiple Machine Learning algorithms are tested for a time horizon of 7-8 months to finalise the Neural Network model which best predicts diamond pricing. The training data set is the size of about 1 million round diamonds

The proof of the pudding is in the eating

If you have a diamond with its attributes values like shape clarity colour etc try our Artificial Intelligence based Diamond Pricing module on The tool will compute the price of the diamond in today's wholesale diamond market. Most of the attributes are available with the Laboratory certificate provided with a diamond.

Best results for small diamonds between 0.2 carats to 2.5 carats

Important: AI computed price is the wholesale price in B2B market. A retailer will apply markups to the end user

Computational Details:(if you are a AI developer). AI Library: Google Tensor Flow / SciKit Model: Generalised regression neural network (GRNN) Training data: 150,000 Stratified Split 90/10 Accuracy: +- 8%

For More Information contact TransPacific Software [email protected]

Article précédent Article suivant

Google Reviews oogle Reviews