arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.05084cs.LGcs.AIcs.ITmath.IT
收藏

基于形状算子局部估计的自适应 $k$-近邻分类器

Adaptive $k$-nearest neighbor classifier based on the local estimation of the shape operator

Federal University of São Carlos(圣卡洛斯联邦大学);Sony Computer Science Laboratories(索尼计算机科学实验室);Queen Mary University of London(伦敦大学玛丽皇后学院)

Alexandre Luís Magalhães Levada, Frank Nielsen, Michel Ferreira Cardia Haddad

AI总结:

本文提出 $kK$-NN,利用局部协方差和 Hessian 矩阵估计形状算子与 Gaussian 曲率,从而自适应确定邻域大小,并在多个真实数据集上以更少训练样本取得优于传统及自适应 $k$-NN 的平衡准确率。

摘要:

$k$-近邻($k$-NN)算法是非参数分类中最流行的方法之一。然而,一个相关局限在于邻居数量 $k$ 的设定。该参数直接影响分类器的若干性质,例如偏差-方差权衡、决策边界的平滑性、对噪声的鲁棒性以及类别不平衡处理。在本文中,我们提出一种新的自适应 $k$-近邻($kK$-NN)算法,该算法通过考察样本处的局部曲率来自适应地确定邻域大小。其基本原理是,低曲率点可以具有较大的邻域(在局部,切空间能够很好地近似底层数据形状),而高曲率点则应具有较小的邻域(在局部,切空间只是一种粗略近似)。我们通过依据局部协方差矩阵以及局部 Hessian 矩阵计算局部形状算子的近似值,来估计局部 Gaussian 曲率。在许多真实世界数据集上的结果表明,与成熟的 $k$-NN 方法以及另一种自适应 $k$-NN 算法相比,新的 $kK$-NN 算法产生了更优的平衡准确率。这一点在训练数据中的样本数量有限时尤为明显,表明在许多相关情形下,$kK$-NN 能够用更少的数据学习到更具判别性的函数。

英文摘要:

The $k$-nearest neighbor ($k$-NN) algorithm is one of the most popular methods for nonparametric classification. However, a relevant limitation concerns the definition of the number of neighbors $k$. This parameter exerts a direct impact on several properties of the classifier, such as the bias-variance tradeoff, smoothness of decision boundaries, robustness to noise, and class imbalance handling. In the present paper, we introduce a new adaptive $k$-nearest neighbours ($kK$-NN) algorithm that explores the local curvature at a sample to adaptively defining the neighborhood size. The rationale is that points with low curvature could have larger neighborhoods (locally, the tangent space approximates well the underlying data shape), whereas points with high curvature could have smaller neighborhoods (locally, the tangent space is a loose approximation). We estimate the local Gaussian curvature by computing an approximation to the local shape operator in terms of the local covariance matrix as well as the local Hessian matrix. Results on many real-world datasets indicate that the new $kK$-NN algorithm yields superior balanced accuracy compared to the established $k$-NN method and also another adaptive $k$-NN algorithm. This is particularly evident when the number of samples in the training data is limited, suggesting that the $kK$-NN is capable of learning more discriminant functions with less data considering many relevant cases.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑