Modern recommendation systems leverage historical behavior information to generate precise recommendation results for users. However, when the data scale of users and items is large, it is difficult to generate recommendation results in time. Tang proposed a quantum-inspired recommendation algorithm, which could solve the recommendation problem in constant time complexity. However, Tang’s approach is based on a set of assumptions which rely heavily on some empirical parameters. The time complexity for calculating parameters is high. Thus, this approach cannot be directly applied in industrial applications. In this paper, we propose a method, namely, Quantum-inspired Recommendation system with threshold Proportion Interception (QRPI), which is based on the quantum-inspired recommendation system and more suitable for industrial environments. Compared with the existing widely used recommendation algorithms, we show through numerical experiments that our solution can achieve almost the same performance with better efficiency.

中文翻译：

---

具有阈值比例截取的量子启发推荐系统

现代推荐系统利用历史行为信息为用户生成精确的推荐结果。然而，当用户和物品的数据规模较大时，很难及时生成推荐结果。Tang提出了一种受量子启发的推荐算法，可以解决恒定时间复杂度的推荐问题。然而，唐的方法是基于一组严重依赖于一些经验参数的假设。计算参数的时间复杂度很高。因此，这种方法不能直接应用于工业应用。在本文中，我们提出了一种方法，即问灵感来源R有门槛的推荐系统磷比例一世拦截（QRPI），基于量子启发的推荐系统，更适用于工业环境。与现有广泛使用的推荐算法相比，我们通过数值实验表明，我们的解决方案可以以更好的效率实现几乎相同的性能。