Detect and Track Latent Factors with Online Nonnegative Matrix Factorization
Detect and Track Latent Factors with Online Nonnegative Matrix Factorization
复制标题
DOI：
--
复制
发表时间：
2007-01
期刊：
影响因子：
--
通讯作者：
Bin Cao;Dou Shen;Jian-Tao Sun-;Xuanhui Wang;Qiang Yang;Zheng Chen
中科院分区：
文献类型：
--
作者：
Bin Cao;Dou Shen;Jian-Tao Sun-;Xuanhui Wang;Qiang Yang;Zheng Chen
从时态数据中检测和跟踪潜在因素是一项重要的任务。大多数现有的潜在主题检测算法，如非负矩阵分解（NMF）是针对静态数据设计的。这些算法无法捕捉随时间变化的数据流的动态特性。本文提出了一种在线NMF（ONMF）算法来检测潜在因素，并跟踪其演变的数据演变。通过利用已经检测到的潜在因素和新到达的数据，潜在因素被自动地和递增地更新以反映因素的变化。此外，通过对检测到的潜在因素施加正交性，我们不仅可以保证NMF的唯一解，而且还可以减轻部分数据问题，这可能会导致NMF失败时，数据是稀缺的或分布是不完整的。在合成数据和真实的数据上的实验验证了ONMF算法的有效性。
Detecting and tracking latent factors from temporal data is an important task. Most existing algorithms for latent topic detection such as Nonnegative Matrix Factorization (NMF) have been designed for static data. These algorithms are unable to capture the dynamic nature of temporally changing data streams. In this paper, we put forward an online NMF (ONMF) algorithm to detect latent factors and track their evolution while the data evolve. By leveraging the already detected latent factors and the newly arriving data, the latent factors are automatically and incrementally updated to reflect the change of factors. Furthermore, by imposing orthogonality on the detected latent factors, we can not only guarantee the unique solution of NMF but also alleviate the partial-data problem, which may cause NMF to fail when the data are scarce or the distribution is incomplete. Experiments on both synthesized data and real data validate the efficiency and effectiveness of our ONMF algorithm.