LOF outlier detection method based on k-d tree

Abstract

The invention discloses an LOF (Local Outlier Factor) outlier detection method based on a k-d tree, and mainly solves the problem that the current LOF outlier detection method is high in time and space complexity during outlier detection. The technical scheme is as follows: the k-d tree is used to store data objects in the k-dimensional space for fast retrieval; the k-d tree is constructed by segmenting the k-dimensional data space with a hyperplane perpendicular to a coordinate axis; each node of the k-d tree corresponds to one k-dimensional hyper-rectangular region; and all outlier detectionis performed on the k-d tree structure. The LOF outlier detection method based on a k-d tree provided by the invention overcomes the problem that the LOF outlier detection method in the prior art hashigh time and space complexity and poor practicability when processing real-time large-scale high-dimensional data objects, and improves the efficiency and practicability of the calculation process while ensuring the accuracy of the outlier detection under a large-scale data set.

Description

technical field [0001] The invention belongs to the technical field of computers, and further relates to a method for detecting outliers based on a local outlier factor LOF (Local Outlier Factor) of a k-d tree in the technical field of data mining. The k-d tree is a tree-shaped data structure that stores data objects in the k-dimensional space for fast retrieval. It splits the k-dimensional data structure into a k-nearest neighbor search space partition tree to build a hierarchical structure of data objects. The invention can be used to mine outliers in a data set composed of network data streams, reduce time and space complexity during outlier mining while ensuring mining accuracy, and realize efficient mining of outliers. Background technique [0002] Outlier mining technology is an important research direction of data mining technology. By mining outliers, we can find potential useful information in the data set. Among them are deeper, potential and valuable information....

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to address the above-mentioned deficiencies in the prior art, and propose a LOF outlier detection method based on k-d tree, to solve the current LOF outlier detection method to large-scale data set detection efficiency is not high and exist in the calculation process The problem of excessive computational overhead

The invention overcomes the problem that the LOF outlier detection method in the prior art has high time and space complexity when processing real-time large-scale high-dimensional data objects, resulting in poor practicability, and ensures outlier detection under large-scale data sets At the same time, the efficiency and practicality of the calculation process are improved

Images