The Volatility and Growth Trend of Field Production Mechanization Level in China - A Positive Analysis Based on Hodrick-Prescott Method - Efficient Decision Support Systems: Practice and Challenges from Current to Future - page 401

(1)
()



Here, is the K th analytical solution.

Step 5 Predicted results

xk

(0) (1) (1)

  

xk (  ) xk (  ) xk ( ) (12)

(0)



In the equation (12), is the K th predicted result by reduction.

Then, it is to get the following predicted data sequences

xk ()

(0) (0) (0) (0) (0)

    

x  { x ( ), x ( ),  , x ( n ), x ( n  )} (13)

Step 6 checking the error size, analyzing the accuracy of predicted results
GM(1,1) group includes four, five, six data and multiple data GM(1,1). In the prediction
process, researchers could choose corresponding GM (1, 1) model according to the actual
situation of the forecasted object (Deng, 1989).

3.2 Results and discussions
According to the formula (8) ~ (13), we established the GM(1,1) prediction model of
mechanical harvesting level and mechanical sowing level based on data four, data five and
data six. The main purpose is to choose the best GM (1, 1) forecasting model as the final
prediction model through comparison. Among them, the GM(1,1) of data four uses the
trend value of mechanical harvesting level and mechanical sowing level from 2002 to 2005,
the GM (1, 1) model of data five uses the trend value of mechanical harvesting level and
mechanical sowing level from 2001 to 2005 and the GM(1,1) of data six uses the trend value
of mechanical harvesting level and mechanical sowing level from 2000 to 2005. The actual
trend value of mechanical harvesting level and mechanical sowing level from 2006 to
2008 is used as inspection samples. Calculation results are showed in table 8 and
table 9.
Table 8 shows that the average prediction error of the trend value of mechanical sowing
based on data four dimension is 2.56%, obviously less than that based on data five and data
six. This demonstrates that the result of GM (1, 1) model based on data four is more precise
and this model can be used to forecast the trend value of mechanical sowing. The calculation
equation is:

xk (  1) 551.5702 exp(0.0488 ) k  524.8002 (14)

Using the equation (14), we can predict the trend value of Chinese mechanical sowing level
from 2009 to 2015 and the results are shown in table 10.
Similarly, table 9 shows that GM(1,1) based on data four is the most precise one, thus we
choose this model as the final model. The calculating equation is:

xk (   1) 235.5554 exp(0.0800 ) k  216.9254 (15)

Using the equation (15), the trend value of mechanical harvesting level from 2009 to 2015
can be predicted, the results are also shown in table 10.