摘要
本文用时序分析法对“深圳证券交易所”自１９９３年１２月２８日至１９９４年１月１０日共１０个交易日的股票指数进行分析，推导出股指动态数学模型：Ｘｔ－２３７．３＝０．７１７（Ｘ（ｔ－１）－２３７．３）－０．５１（Ｘ（ｔ－２）－２３７．３）＋αｔ（αｔ白噪声序列）并用此模型对后三个交易日分别作了一步、二步、三步预报，得到预报的相对误差在０．３２％～０．４３％的满意结果。
n this article, we analyse, by using the method of sequential analysis, the share indexes of ten dates of exchange from December 28, 1993 to January 10, 1994 in Shenzhen Stock Exchange, and deduce a mathematical model of the trends of share indexes:X_t-237.3=0.717(X_(t-1)-237.3)-0. 51(X_(t-2)-237.3) + a_t(a_t here is the white noisy sequence).Then we use the model to forecast the latter three exchange dates with three steps respectively ,and obtain a satisfactory result in which the relative errors are between 0.32% and 0.43%.
