论文部分内容阅读
上述数据进行了一元回归分析。 表中1959年和1960年产量较其前后年份变化幅度很大,考虑到当时的历史情况,将其视为奇异点,不计在统计分析的总体之内。将余下的34组数据分别用: 直线 Q=A+BT 双曲线 1/Q=A+B/T 指数曲线 Q=Ae~(BT) 指数曲数 Q=Ae~(-B/T) 幂曲线 Q=AT~B 二次曲线 Q=A+BT~2 对数曲线 Q=A+Blog(T/t+1) S形曲线 Q=1/(A+Be~(-1)) 拟合(式中:A、B—系数;Q—年产量;T—年份;t—对数曲线中的辅助参数),得到最佳回归曲线为:
The above data were a one-way regression analysis. In the table, the output in 1959 and 1960 is much larger than the years before and after the change. Considering the historical situation at the time, it is regarded as a singular point and not counted in the overall statistical analysis. The remaining 34 sets of data were used: Straight line Q = A + BT hyperbolic 1 / Q = A + B / T exponential curve Q = Ae ~ (BT) exponential number of turns Q = Ae ~ (-B / T) Q = AT ~ B quadratic curve Q = A + BT ~ 2 logarithmic curve Q = A + Blog (T / t + 1) S-shaped curve Q = 1 / (A + Be -1) In the formula: A, B-coefficient; Q-annual output; T-year; t-logarithmic curve in the auxiliary parameters), the best regression curve: