| Summary: | 碩士 === 國立雲林科技大學 === 財務金融系碩士班 === 93 === This thesis uses the artificial neural network(ANN) to investigate the linearity and non-linearity of relation between unexpected order flow and price changes. We analyze one minute intraday data of Taiwan stock index futures from January. 1, 2003 to December 31, 2003. The empirical result show as follow: Firstly, the relation between order flow and price changes is non-linear, and the ANN does add the explanation power on the price changes. Secondly, the information content of order flow increase with its size. Thirdly, buyer initiated trades and seller initiated trades have different information content and impact on price changes. Finally, the number of contracts traded offers the almost same explanation as the size of contracts traded.
|
|---|
