Efficient sparse subspace clustering by nearest neighbour filtering

Subspace identification has been used extensively because its ability to detail the internal subspace structure of data, which can be in a variety applications such as dimension reduction, anomaly detection and so on. However, many advanced algorithms are limited on their applicability large data sets due computation memory requirements with respect number input points. To overcome this problem, we propose simple method that screens out points by using k nearest neighbours recovery is performed reduced set. The proposed surprisingly significant reduction both computations requirements, yet possesses desirable probability lower bound for success context big data. Besides theoretical analysis, our experiments also show exceeds expectations outperforms existing similar algorithms.