Abstract: in many practical problems, the training data is different, usually some training data may be more important than other data. Each group of training data depict different function is usually through each training data gives an incredible weights, based on the fuzzy data, this paper has deduced the classic again in the regression method, and the relative cases of fuzzy regression method, and analyzed their statistical properties. In fuzzy statistic analysis of the traditional methods are incredibly weight by setting the fuzzy membership functions, such as to determine the time sequence function etc. Considering the DEA method in the analysis of the relative effectiveness evaluation decision unit has not underestimate the superiority of the DEA method, then we introduce fuzzy statistic analysis, and usually use the fuzzy membership functions of the incredible weight determined comparing method, fuzzy statistic analysis method in DEA model is superior. Keywords: DEA analysis method, Least-squares regression analysis, Fuzzy statistic,

Procter & gamble co., LTD. Is a U.S. consumer commodity producers, is currently the world's largest commodity one of companies. Headquartered in the United States in Cincinnati, Ohio, there are thousands of employees around the world. Procter & gamble's products involved in all aspects, such as: shampoo, hair care, cosmetics, baby care products, women's health supplies, medicine and food, etc. Procter & gamble's products in the world have a considerable reputation, operating more than 300 brands, sold to many countries, and has set up factories in more than 80 countries and companies. In p&g's marketing strategy undoubtedly become the magic weapon for winning the multi-brand strategy. From procter & gamble's brand strategy research, this paper explore its brand in the process of the development of the fine, find out the problems existing in the brand strategy ? 宝洁公司 Procter & gamble company 全球最大日用品公司 The world's largest commodity company 多品牌策略 Multi-brand strategy