论文部分内容阅读
目的:比较动态数列法、直线回归模型、对数模型、二次曲线模型、三次曲线模型、ARIMA模型在肿瘤专科医院门诊人次拟合效果的优劣,为医院行政部门提供合适的预测模型。方法:应用6种预测方法对肿瘤专科医院门诊人次进行预测并比较拟合值的绝对误差、平均相对误差绝对值和误差平方和。结果:对肿瘤专科医院门诊人次预测方面,ARIMA模型相对误差绝对值最小(11.29%),其次为三次曲线模型(15.28%)、二次曲线模型(15.43%)、直线回归模型(15.60%)、动态数列法(17.38%),相对误差绝对值最大的是对数模型,为20.17%。结论:ARIMA模型对肿瘤专科医院门诊人次发展变化规律的分析有较好的适应性和实用性,可以为肿瘤专科医院今后工作的发展规划提供一定的依据。
Objective: To compare the advantages and disadvantages of dynamic series method, linear regression model, logarithmic model, quadratic curve model, cubic curve model and ARIMA model in outpatient visits in oncology hospitals, and provide appropriate prediction models for hospital administrative departments. Methods: Six kinds of prediction methods were used to predict the outpatient visits of tumor specialist hospitals and compare the absolute value of the fitted value, the absolute value of the average relative error, and the sum of squared errors. RESULTS: The relative error of the ARIMA model was the smallest (11.29%), followed by the cubic curve model (15.28%), the quadratic curve model (15.43%), and the linear regression model (15.60%). In the dynamic array method (17.38%), the logarithmic model is the largest relative error absolute value, which is 20.17%. Conclusion: The ARIMA model has good adaptability and practicality to the analysis of the development and variation of outpatient visits in oncology hospitals, and it can provide a basis for the future development planning of oncology hospitals.