Design of robust algorithm for machine learning based on deep search of outliers

Design of robust algorithm for machine learning based on deep search of outliers is implemented in this study. The depth first search is a graph traversal method. The traversal process is essentially the process of finding the leading point of each vertex. It starts from a vertex of the graph, and the visited vertices are marked with seven visited marks. The statistical inference rules under this new theoretical system not only take into account the requirements for asymptotic performance, but also seek to obtain optimal results under the condition of existing limited information. By transforming the original problem into a dual problem, the computational complexity of the support vector machine no longer depends on the spatial dimension, but on the number of support vectors in the sample. For the robust system, the abnormal detection is considered for the application scenario. Through the Hadoop testing, the ideas are validated.