| With the rapid development of modern industry, the market competition is increasingly fierce. Cost has become an important factor affecting the development of the industry, so cost management is very critical. As a traditional pillar industry of the national economy in our country, the textile industry has made a great contribution to the social development. In 2012, in order to implement "The Twelfth Five-Year Plan for National Economic and Social Development" and "Industrial Restructuring and Upgrading Planning (2011-2015)", ministry of industry and information technology formulated and issued the "Twelfth Five-Year Development Plan of Textile Industry" that pointed out the textile industry faced huge challenges in the international and domestic. Many factors affected the development of the textile industry, including change in exchange rates, international trade barriers, rising prices of the industry raw materials, increased labor costs, technical reform, environment, energy and so on. And these factors are directly or indirectly affect the industry’s costs. The textile industry is an industry with small profit margins, so cost management becomes very important to it. Therefore, it’s necessary to research on cost and establish cost early warning system of textile industry.Preparedness ensures success, un-preparedness spells failure. Establishing an effective cost early warning system can help the industry analysis cost, control cost, enhance economic benefits and promote sustainable development of the textile industry. So this article will be based on textile industry to study cost early warning.Firstly, the paper describes the background and significance of the study of cost early warning on textile industry and summarizes the status of related research at home and abroad. Meanwhile, this article also introduces related theory on cost early warning, the definition of data mining, BP neural networks, support vector machine and analog complexing algorithm.Secondly, according to related literature research on cost early warning and the status of the textile industry, this article analyzes the influencing factors of cost from the perspective of qualitative. And combining with the quantitative analysis of gray relation method, the paper establishes the index system of cost early warning of textile industry. Meanwhile, the paper classifies index of alert warning degree, and quantized the early-warning interval. Then this paper uses stepwise regression method to screen index, eliminate insignificant variables and reduce the input dimension of the model. And building the BP-SVM-AC early-warning model based on data mining methods, the paper conducts to study on cost early warning in two phases. The First phase, the paper analyzes the degree and strength of alert of cost based on the improved BP-SVM model. The Second phase, the paper predicts the alert of cost with AC algorithm.Finally, the article does empirical research by means of numerical software Matlab R2012a. Through the analysis of the results of BP-SVM-AC early-warning model, it is proved that the model has good warning effect. So the BP-SVM-AC early warning model has certain scientific and practical characteristics, which could be use as analysis of the alert of cost on textile industry in the future, help the relevant administrative departments of the textile industry grasp the development trend of the alert of cost in advance. It’ll help them take measures to deal with the alert in time and also provide a certain reference value to industry enterprises to research cost early warning. |
