defgf24uw
Dołączył: 01 Lis 2010
Posty: 959
Przeczytał: 0 tematów
Ostrzeżeń: 0/5
Skąd: England
|
 Wysłany: Śro 5:44, 12 Sty 2011 Temat postu: mbt calzature Prediction method based on a combina |
|
|
Prediction method based on a combination of settlement
BP neural network based forecasting model input. Measured in Table 1, from December 1997 to March 1999 the basis of actual settlement as output. Figure 2 shows the establishment of BP neural network based forecasting model: {= stallion, j and then use the combination forecasting model June 1999 and September 1999 were the basis for settlement of the predicted results in Table 3. The same time [3] predictions are also listed in Table 3, Table 2, Table 3 shows that BP neural network-based forecasting method, the prediction results than any single prediction method,[link widoczny dla zalogowanych], the prediction results are more close to the actual observations. Based on BP neural network forecasting model predicted curves, based on the measured value and the settlement of [3], the prediction curve shown in Figure 3 Figure 3 based on the measured and predicted subsidence curves for the accuracy of test results in Table 4, Table 4 shows,[link widoczny dla zalogowanych], Although the Grey Verhulst model, BP neural network model and BP neural network based forecasting model has the same prediction accuracy level. Able to meet project needs; but this created a group based on BP neural network subsidence observation data in Table 2 Table 3 Comparison of predicted time series observation period 12345678, month, year 12,19973,19986,[link widoczny dla zalogowanych],19989,[link widoczny dla zalogowanych],199812,19983,19996,[link widoczny dla zalogowanych],19999 / 1999} 4.983.13942.69491.63320.79o3O. 34o80.13960. o222 4.97993.46992.42831.57440.85480.4o230.16810.0589 different models of pre-test results / mm4.97883.46982.42941.56920.85950I39980.15830.0465 document 35.493. O82.481.450.69O. 3O0.120. Table 4 The accuracy of test results O5 3 business Huai handsome. Such as: a combination of foundation settlement prediction method combined prediction model 7 the highest prediction accuracy of the predicted results coincide with the actual value of the best and most reliable prediction model can be predicted on the basis of settlement. 5 Conclusion The example calculation of this article can be seen, based on artificial neural network forecasting model estimated on the basis of settlement has its special advantages, which effectively combines the data sequence is weakened Grey volatility characteristics and neural network-specific nonlinear adaptive information processing capacity, with a model is simple, do not need to determine the nonlinear function and so on. The model can be \. This deficiency is less sample data, poor stability of neural network model. If you can overcome the above shortcomings, will be on the basis of prediction of settlement with time is more intelligent, which leads to better prediction results.
More articles related to topics：
[link widoczny dla zalogowanych]
[link widoczny dla zalogowanych]
[link widoczny dla zalogowanych]
Post został pochwalony 0 razy
|
|
FAQ
Szukaj
Użytkownicy
Grupy
Galerie
Rejestracja
Zaloguj się, by sprawdzić wiadomości
Zaloguj
