大区域内药品销售预测方法研究与仿真

Large Area Sales Forecast Method and Drug Simulation

研究药品未来销售量准确预测问题。在一些大区域中,药品销售网点分布范围比较广泛,销售额差异较大,会造成以此为基础建立的地域参数和单个销售参数维数过高。传统的Hilbert药品销售预测算法需要对相关参数进行内积运算,造成数据量过大,导致药品实际销售结果与预测结果在时间上存在滞后,降低了药品销售预测的准确率。提出了一种遗传算法优化支持向量机的药品销售预测方式。利用滞后补偿支持向量机,获取药品销售预测支持函数;建立药品销售预测滞后补偿规则,从而准确预测出药品销售相关参数。实验证明,利用该方法对大区域内药品销售预测方式进行预测,提高了药品销售预测的准确率,取得了令人满意的效果。

Drug sales predict accurately the future research. In some large area, drug sales outlets distribution is widely used. The traditional Hilbert drug sales forecast algorithms of concerned parameters are inner product opera- tion, causing excessive amount of data and leading to actual sales result lagging predicted results and the drug sales forecast accuracy is reduced. The paper put forward a drug sales forecast method using genetic algorithm to optimize the support vector machine（SVM）. Lag compensation support vector machine was used to get the drug sales forecast support functions. The drug sales forecast lag compensation rules were established to predict drug sales related param- eters. The experiment proved that using the method to large area drug sales forecasts can improve the drug sales fore- cast accuracy, and satisfactory results have been obtained.