【作者】 刘海；

【导师】 冯岑；

【作者基本信息】 苏州大学 ， 纺织工程， 2012， 硕士

【摘要】 国内外对丝绸行业的研究都很重视数据的采集，对产量、价格、进出口额等数据的采集是希望从这些数据中分析并获取市场信息，但是目前大多数的数据研究是基于已有数据上的定量分析，因此其在指导市场的判断上有滞后性，不具有前瞻性和预测作用。因此，在已知信息的基础上，建立合适的数学模型来预测和研究丝绸市场规律显得尤为重要。本课题以市场上运行的相关实际数据为研究背景，将灰色系统理论方法与灰色预测模型应到茧丝绸市场波动分析中。以相关实际数据为对象，分别建立GM (1,1)模型、改进的GM (1,1)模型和灰色马尔可夫模型，同时利用MATLAB工具，编程实现了数据的可视化处理，基于此从不同角度分析、研究和预测茧丝绸价格指数，并对结果进行分析、对比和评价。以干茧价格指数为研究对象，分别建立GM (1,1)模型和改进GM (1,1)模型进行预测、分析，显示两者均可用于价格指数的短期预测；同时，通过实证对比，验证了改进GM (1,1)模型在价格指数预测上的优势。基于改进GM (1,1)分别对茧丝价格指数和丝绸价格指数进行了预测研究，探讨了价格指数体系中各级指数预测精度的差异规律。针对茧丝绸价格指数数据波动性的特点，建立了灰色马尔可夫模型，并基于灰色马尔可夫模型对茧丝价格指数和丝绸价格指数进行了预测研究，验证了灰色马尔可夫模型应用的可行性。通过运用BP神经网络模型、改进GM (1,1)模型和灰色马尔可夫模型对丝绸面料价格指数进行预测研究，并对比分析了三种方法，显示灰色马尔可夫模型预测的预测效果较好。总而言之，这些方法均可用于茧丝绸价格指数的短期预测，其预测和研究对于实际的茧丝绸市场具有一定的现实指导意义。更多还原

【Abstract】 The study of silk industry at home and abroad all pays much attention to datacollection. People collect data of output, price and export to analyze and acquire marketinformation. But now most of the data study focuses on quantitative analysis based onthe existing data, making the study lagged behind in the guidance of the marketjudgment, not forward-looking or prediction. Therefore, establishing appropriatemathematical models based on the known information to forecast and study silk marketrule is very important.Using actual data operated in market as the research background, the paper appliesthe grey system theory method and grey model of prediction to the fluctuation analysisof Chrysalis silk market. It sets actual data related as the object; respectively buildGM (1,1)model, the improved GM (1,1)model and Grey-Markov model, while usingMATLAB tools to realize the data visualization processing through programming.Based on this analysis, it also studies, analyzes and predicts chrysalis silk price indexfrom different angles and analyzes, compares and evaluates the results. It uses drychrysalis price index as the research object, respectively establishes the GM (1,1)modeland the improved GM (1,1)model to forecast and analyze, displaying that both can beused for price index short term prediction. At the same time, it shows the advantages ofthe improved GM (1,1)model in prediction price index through empirical comparison.Based on the improved GM (1,1)model, it studies respectively of the prediction ofchrysalis silk price index and silk price index, discusses the difference rule of variousindex prediction accuracy in the price index system. In the light of fluctuationcharacteristic of chrysalis silk price index data, the paper establishes Grey-Markovmodel to forecast the chrysalis silk price index and silk price index, verifying the feasibility of the application of Grey-Markov model. At last, it studies the prediction ofsilk fabric price index using the BP neural network model, the improvedGM (1,1)model and Grey-Markov model and also compares the three methods, showingthat the Grey-Markov model does better in prediction. In a word, all of these methodscan be used in short term prediction of chrysalis silk price index. And the prediction andresearch for real silk market has certain realistic directive significance.更多还原