【摘要】 讨论S型曲线Y=k[1+bexp(-aX)]的最佳拟合问题,提出一种扫描回归的新方法,并利用实际数据说明,与普通回归相比,该扫描回归方法具有相关系数较大、精确度较高的优点更多还原

【Abstract】 In this paper, the question of best Logistic regression curve, Y=k/[1+b exp (-aX)] , is discussed. The paper gives a new method of regression by scanning. Applied to practical data, this method is proved to have higher accuracy with greater correlative coefficient than ordinary ways of regression.更多还原