Method and system for spot pricing via clustering based demand estimation
View Patent ↗A method and system can be used to determine an optimal price of a commodity on a spot market. Market state conditions can be generated using historical data, such as transactional and other data. The market state conditions may include nearly any number of attributes, and the forecast market state conditions may include a forecast price for the commodity during the next time period. The historical data and market state condition may be used in a clustering module that separates the data into clusters. The cluster with the market state condition is identified, and the data in that cluster is used to generate a price-demand curve for the next time period. The price for the commodity on the spot market during that time period can be determined consistent with the company's business rules.
1. A computer-implemented method of determining a spot price for a commodity on a spot market, comprising:
generating market states using historical data, wherein the historical data includes transactional data and non-transactional data, wherein the transactional data includes prices and quantities of the commodity and date sold in past transactions, wherein the non-transactional data includes non-transactional information or conditions that affect the spot price or demand of the commodity, wherein the market states include market stat attributes, and wherein the market state attributes include product- or service-based data, customer-based data, competitor-based data, seasonal variations, and special events;
calculating a forecast of the market states for a next pricing period, wherein the forecast includes a forecast price for the commodity on the spot market during the next pricing period;
generating a clustering index for each of the past transactions and each of the forecasted market states;
comparing clustering indices of the past transactions and the forecasted market states for the next pricing period; and
generating a price-demand curve for the commodity on the spot market for the next pricing period using records from the past transactions having clustering indices that are the same or comparable to the cluster index of the forecasted market states for the next pricing period.
2. The computer-implemented method of claim 1 , wherein the forecast of the markets states for a next pricing period comprises at least one of a maximum price for the commodity, a minimum price for the commodity, a company's price rank, or the nearest higher price of the commodity.
3. The computer-implemented method of claim 1 , wherein generating is performed without using data from any other cluster.
4. The computer-implemented method of claim 1 , further comprising determining the spot price for the commodity on the spot market for the next pricing period using the price-demand curve.
5. The computer-implemented method of claim 4 , wherein determining the spot price for the commodity on the spot market comprises determining the spot price consistent with maximizing profit, volume, and revenue.
6. The computer-implemented method of claim 1 , wherein the commodity is a product.
7. The computer-implemented method of claim 1 , wherein the commodity is a service.
8. A computer readable medium having code embodied therein, which when executed causes a computer to perform the steps comprising:
an instruction for generating market states using historical data, wherein the historical data includes transactional data and non-transactional data, wherein the transactional data includes prices and quantities of the commodity and date sold in past transactions, wherein the non-transactional data includes non-transactional information or conditions that affect the spot price or demand of the commodity, wherein the market states include market state attributes, and wherein the market state attributes include product- or service-based data, customer-based data, competitor-based data, seasonal variations, and special events;
an instruction for calculating a forecast of the market states for a next pricing period, wherein the forecast includes a forecast price for the commodity on the spot market during the next pricing period;
an instruction for generating a clustering index for each of the past transactions and each of the forecasted market states;
an instruction for comparing clustering indices of the past transactions and the forecasted market states for the next pricing period; and
an instruction for generating a price-demand curve for the commodity on the spot market for the next pricing period using records from the past transactions having clustering indices that are the same or comparable to the cluster index of the forecasted market states for the next pricing period.
9. The computer readable medium of claim 8 , wherein the forecast market state condition comprises at least one of a maximum price for the commodity, a minimum price for the commodity, a company's price rank, or the nearest higher price of the commodity.
10. The computer readable medium of claim 8 , wherein the instructions for generating is executed without using data from any other cluster.
11. The computer readable medium of claim 8 , wherein the code further comprises an instruction for determining a spot price for the commodity on the spot market for the next pricing period using the price-demand curve.
12. The computer readable medium of claim 8 , wherein the instruction for determining the spot price for the commodity comprises an instruction for determining the spot price consistent with maximizing profit, volume, or revenue.
13. The computer readable medium of claim 8 , wherein the commodity is a product.
14. The computer readable medium of claim 8 , wherein the commodity is a service.
15. A system for determining a spot price for a commodity on a spot market, comprising:
a database comprising historical data for the commodity, wherein the historical data includes transactional data and non-transactional data, wherein the transactional data includes prices and quantities of the commodity and date sold in past transactions, wherein the non-transactional data includes non-transactional information or conditions that affect the spot price or demand of the commodity, wherein the market states include market state attributes, and wherein the market state attributes include product- or service-based data, customer-based data, competitor-based data, seasonal variations, and special events;
a processor comprising:
a market state generation module that is adapted to generate market states using the historical data and a forecast of the market states for a next pricing period;
a clustering module that is adapted to generate a clustering index for each of the past transactions and each of the forecasted market states;
a demand curve generation module that is adapted to generate a price-demand curve for the commodity on the spot market for the next pricing period using records from the past transactions having clustering indices that are the same or comparable to the cluster index of the forecasted market states for the next pricing period.
16. The system of claim 15 , further comprising a price determination module that is adapted to use a demand curve from the demand curve generation module and a business rule to determine the spot price for the commodity on the spot market for the next pricing period.
17. The system of claim 15 , wherein:
the market state comprises a prediction of the spot price for the commodity during the next pricing period; and
the records used by the demand curve generation module comprise the prediction of the spot price for the commodity.