【作者】 苗开超；

【导师】 胡学钢； 刘月成；

【作者基本信息】 合肥工业大学 ， 软件工程， 2009， 硕士

【摘要】 近年来,随着科学技术的飞速发展,经济和社会都取得了极大的进步,与此同时,在各个领域产生了大量的数据,其中也包括农产品价格数据。激增的数据背后隐藏着许多重要的信息。人们不再满足于数据库的查询功能,希望能够对其进行更高层次的分析,以便能从数据中提取信息或者知识为决策服务。农产品价格受市场影响的程度特别大,特别是受农产品的供求关系影响较大,同时价格本身又受自然条件、社会和经济条件的影响,特别是国际市场的影响。从价格本身来看,受供求、季节等发生波动,受外界各种影响比较多,这就造成了价格预测的困难。但从长期看,农产品价格随着时间的推移仍然呈现一定规律性。指数平滑一种数学上比较成熟的随机时间序列预测方法,已经在卫生、经济等领域得到广泛应用,它是目前公认的用于一个国家或地区经济预测中比较先进的适用的科学的时间序列分析模型之一。那么这种方法是否也适用于农产品的价格走势的预测呢?目前,这方面的研究的较少,本文就此课题进行了研究。本文研究时间序列数据挖掘,通过对过去历史行为的客观记录分析,揭示其内在的规律,进而完成预测未来行为等决策性工作。通过研究时间序列分析法中的指数平滑法,预测农产品的价格趋势。论文基于指数平滑法,通过历史上一段时间某地西红柿价格变化情况,建立西红柿的价格指数平滑模型,用该模型拟合西红柿的历史走势,根据模型预测该农产品未来几年的价格变化情况。试验结果表明,本方法能够有效地对西红柿价格曲线进行拟合,预测。最后,为了实现农产品价格预测的自动化处理,设计并实现了一个基于Web数据挖掘的农产品价格搜集,查询,检索,预测系统。更多还原

【Abstract】 In recent years, with the rapid development of science and technology, economy and society have made great progress, meanwhile a large mount of date such as agricultural prices have been produced in various fields. It has come to the point that many important information are behind the dramatically increasing of date. Now simply query with database is no longer satisfying our demand. Additionally we hope to be able to carry out data analysis over a higher level in order to extract useful information or knowledge from the large mount of data or knowledge for decision-making.Prices of agricultural products are affected by the market in a high extent, especially the relationship between supply and demand of agricultural products, while the price itself especially in the international markets is subject to natural conditions, social and economic conditions. Therefore forecasting agriculture products prices shall be a difficult but significant job. Though price is influenced by the fluctuations of supply and demand, seasonal, region etc, it still has certain regularities when related to a long period.Exponential smooth is a forecasting method using random time series which is quite mature in mathematical field. It has been widely used in health, economic and other fields, it is now recognized as one of the advanced madels which are applicative in economic forecasting for a country or a region. Whether this approach is also adaptive for forecasting agricultural products prices or not is ambiguous, so this theis is object to this subject.We research time series data mining through the objectively analyze records of past acts in history, reveal its inherent la regular pattern, further more forecast future behavior which is available for decision-making. Our thesis is in view of time series analysis based on exponential smoothing method in order to forecast price trends of agricultural products.This thesis is based on the exponential smooth model, we establish a tomatoes exponential smoothing model which according to its price fluctuation during a long time, fitting the model with the historical trend of tomatoes. And then we can obtain the price of tomatoes in following years according to this model. The results show out that this model can effectively carry out fitting and forecasting the price curve of tomatoes.Finally, in order to achieve automatic processing in forecast prices of agricultural products, a Web-based data mining to collect the prices of agricultural products is designed and implemented, then this system can be realized query, retrieval, prediction.更多还原