韓文喜

摘要：在殘差修正的非等距灰色預(yù)測模型的基礎(chǔ)上，將傳統(tǒng)無偏灰色模型還原為非等距模型時間函數(shù)進(jìn)行原始數(shù)據(jù)列的預(yù)測。在殘差修正時用時間函數(shù)和組合灰色預(yù)測非等距模型進(jìn)行殘差修正，然后將兩種殘差預(yù)測值進(jìn)行加權(quán)得到最終殘差，得到最后修正預(yù)測值。將預(yù)測值與提供的數(shù)據(jù)列進(jìn)行比較，表明此方法可以得到較好的預(yù)測效果。

Abstract： Based on the non-equidistant gray prediction model with residual error， the traditional unbiased gray model is reduced to the non-equidistant model time function to predict the original data series. The residuals are corrected by using the time function and the combined gray prediction non-equidistant model， and then the residuals are weighted by the two residual prediction values to obtain the final modified prediction value. Comparing the predicted value with the data provided， it shows that this method can obtain better prediction effect.

關(guān)鍵詞：改進(jìn)灰色模型；非等距；殘差的修正和加權(quán)；沉降預(yù)測

Key words： improved gray model；non-equidistant；correction and weighting of residuals；settlement prediction

中圖分類號：TU4 文獻(xiàn)標(biāo)識碼：A 文章編號：1006-4311（2017）03-0136-02