混合推荐点餐模型的优化实现

针对我国传统中餐点餐服务中欠缺针对性点餐推荐以及菜品推荐覆盖面较低的问题,提出一种关联规则结合基于菜品属性的推荐算法的混合推荐点餐模型。通过历史关联菜品组合和菜品的关键属性计算菜品关联度与相似度;然后,根据得到的菜品综合评分生成推荐规则来优化传统的关联规则FPgrowth算法推荐;最后根据顾客已点菜品启发式地进行后续点餐推荐。采集了真实的中餐馆历史点餐数据对模型和算法进行了有效性验证,实验结果表明该模型在达到一定菜品推荐数量时,在菜品推荐准确度和覆盖率方面优于传统的单一关联规则推荐,适合较多顾客中餐聚餐的点餐推荐。

An Optimized Implementation of a Hybrid Recommended Order Model

Aiming at the lack of targeted dishes recommendation and the low coverage of dishes recommendation in traditional Chinese meals ordering service, a hybrid ordering recommendation model combining association rules and recommendation algorithm based on dishes attributes was proposed. The correlation degree and similarity of dishes were calculated through the combination of historically associated dishes and the key attributes of dishes. Then, recommendation rules were generated based on the comprehensive score of dishes obtained to optimize the traditional association rule FP growth algorithm recommendation. Finally, follow up dishes recommendation was carried out heuristically according to the ordered dishes of customers. The validity of the model and algorithm was proved by collecting the real historical order data of Chinese restaurants. The experimental results show that the model is better than the traditional single association rule recommendation in the aspects of dishes recommendation precision and coverage when the dishes ordered reach certain amount, it is suitable for the Chinese dishes order recommendation of more customers gathering.