量子聚类代码_量子聚类

量子聚类代码

Following up on my recent article titled, “Maximum Quantum Classification,” in which I started with the data points already mapped to qubits and grouped together, here is an algorithm for doing the initial grouping of the data points. This algorithm is inspired by centroid-based clustering, with which you plot however many data points you have, calculate a number of centroids (center points), and group the data points by their proximity to the centroids. The most popular version of centroid-based clustering is the k means algorithm. My immediately previous article and this article are therefore inspired by the k-nearest neighbors and k-means algorithms, respectively.

在我最近发表的题为“ 最大量子分类 ”的文章之后，我从已经映射到量子位并分组在一起的数据点开始，这是一种对数据点进行初始分组的算法。 该算法的灵感来自基于质心的聚类，您可以使用它绘制出许多数据点，计算出许多质心(中心点)，并根据它们与质心的接近程度对其进行分组。 基于质心的聚类的最流行版本是k means算法。 因此，我的前一篇文章和这篇文章分别受到k最近邻居和k均值算法的启发。

Disclaimer

免责声明

Up front, I hereby admit that I don’t believe this clustering algorithm is optimal. I’m sharing it because it works, but the approach showcases classical thinking and I would really like to reduce the number of qubits and operations required. Otherwise, and notwithstanding the inefficiency of it all, scaling up the circuit would look ridiculous with so much redundancy. Therefore, I hope to share an optimized reimagining in the near future.

首先，我特此承认我不认为该聚类算法是最佳的。 我分享它是因为它可行，但是这种方法展示了经典的思想，我真的很想减少所需的量子位和操作数。 否则，尽管这一切都是低效率的，但由于具有如此多的冗余性，按比例放大电路看起来是荒谬的。 因此，我希望在不久的将来分享经过优化的重新构想。

Data

数据

If you can imagine