论文部分内容阅读
本文作者对海门口岸所在的台州地区具有双峰型季节特征的疾病进行了回归分析预测,1990~1992年病毒性肝炎和肾综合症出血热的前瞻性预测结果:全区九县(市)的预测值与实际发病率的平均拟合率达98.14%,不同标准差范围的比例与理论分布也比较接近,效果满意。同时进一步讨论了各变量值进行对数转换,根据标准差对变量频数分配理论给出预报量和及时吸收新近样本资料,不断更新方程等是提高回归分析预测效果的必要手段。
The author of the Haimen port in Taizhou area with bimodal seasonal disease regression analysis predicted that from 1990 to 1992, viral hepatitis and renal syndrome hemorrhagic fever forward-looking prediction results: the region’s nine counties (cities) The average fitting rate between predicted value and actual incidence was 98.14%. The proportions of different standard deviations ranged from the theoretical distribution and the results were satisfactory. At the same time, we further discuss the logarithmic transformation of each variable value. According to the standard deviation, we give the forecast quantity to the variable frequency distribution theory and absorb the recent sample data in time. It is necessary to update the equation continuously to improve the prediction effect of regression analysis.