Since its founding in 1998, Fenwei has committed to studying the correlation of coal price and supply-demand influencing factors.
From the supply-demand model in 2008 to the present big data forecast model, the system, through upgrades in calculation methods,
has evolved from judging the up or down-ward trend to predict the extent of price rise or drop, with constantly growing accuracy.
By studying correlation between coal price and influential factors, the forecast model helps enterprises improve
their pricing mechanism and enhance their ability to understand and predict the market.
Basic data
productionoperating
ratestocksrail platformsportsimport&
exportconsumption areas...
Modeling method
Building multiple regression model for market forecast, after simulating basic data through correlation analysis, z-score normalization dimensionless method, data dimension reduction, the principal component analysis and regression analysis.
Achievement
Fenwei finds the regression modeling of price forecast for next 42 days has the most optimal results, All model tests (F-test, T-test) met standards.
Economics test After five years of fitting data verification, all price inflection points could be precisely captured.
Thermal coal prediction model Only one trending error occurred in twelve predictive periods, forecast accuracy stood at 91.66% and the mean value of the absolute value of the integral error stood at 15.24 yuan/t.
It contains data influential in prices, scientifically selected by Fenwei, including production, stocks, operating rate and dispatch data of production bases, rail platforms, ports as well as consumption areas.
It includes the forecast results, and analysis on key data and their changes, the comparison to similar results in the past, etc.
Fenwei can also provide customized models to meet enterprises' specific requirements, with enterprise self-owned data to be added as variables for modeling.
For more details : +86 351 7219 322
low-sulfur primary coking coal at main production areas
Coking coal prediction model Only two trending errors occurred in twelve predictive periods, forecast accuracy stood at 83.33%% and the mean value of the absolute value of the integral error stood at 21.56 yuan/t.
It contains data influential in prices, scientifically selected by Fenwei, including production, stocks, operating rate and dispatch data of production bases, rail platforms, ports as well as consumption areas.
It includes the forecast results, and analysis on key data and their changes, the comparison to similar results in the past, etc.
Fenwei can also provide customized models to meet enterprises' specific requirements, with enterprise self-owned data to be added as variables for modeling.
For more details : +86 351 7219 322
Quasi Grade I met coke at Rizhao, Shandong.
Coke prediction model Three trending errors occurred in twelve predictive periods, forecast accuracy stood at 75% and the mean value of the absolute value of the integral error stood at 45 yuan/t.
It contains data influential in prices, scientifically selected by Fenwei, including production, stocks, operating rate and dispatch data of production bases, rail platforms, ports as well as consumption areas.
It includes the forecast results, and analysis on key data and their changes, the comparison to similar results in the past, etc.
Fenwei can also provide customized models to meet enterprises' specific requirements, with enterprise self-owned data to be added as variables for modeling.
For more details : +86 351 7219 322
Shanxi Fensheng Information Services Co., Ltd.
Phone：+86-351-7219322 （Business Hours 08:30-18:00 GMT+8）
Fax：+86-351-4728543 Email：inquiry@fwenergy.com
中文
