中文摘要：

传统的股票预测方法通常采用前一段时间的股票价格预测后一段时间的股票价格．本文提出了一种仅用前一天的股票数据预测后一段时间的股票数据的新方法，先对股市收集到的原始数据进行错位排序，然后采用在线序列ELM算法进行预测．预测结果显示，这样不仅预测精度高，而且速度也很快．

英文摘要：

Traditional stock prediction method usually uses the stock price of some time ago to forecast after a period of time~s. This paper puts forward a new method which uses the stock data of the day before to forecast the following period of timeJs. Apply dislocation ranking method to the original data, after which use online sequential extreme learning machine algorithm to forecast. Forecasting results shows that the forecast precision and speed is high.