A multi-scale mouse track feature extraction method, device and system

Abstract

The invention relates to a multi-scale mouse trajectory feature extraction method, device and system. The multi-scale mouse trajectory feature extraction method comprises the following steps: collecting a mouse trajectory, obtaining a mouse trajectory expression, and dividing the mouse trajectory expression into a training set and a test set; performing base scale feature extraction on the training set data firstly, selecting subdivision scale features corresponding to each base scale feature on the basis of the base scale feature extraction, and extracting the local trajectory features in themouse trajectory expression; selecting the advantageous features in the base scale features, the subdivision scale features and the local trajectory features by means of the stratified random forest;and constructing a parallel voting decision tree with the training set as a sample by means of the advantageous features, and performing classification and identification on the mouse trajectory in the test set.

Description

technical field [0001] The invention relates to a multi-scale mouse track feature extraction method, device and system. Background technique [0002] Mouse track features are the necessary basis for mouse track recognition tasks, and extracting multi-scale mouse track features is the most important preparatory work before subsequent processing of mouse tracks. At present, the methods for feature extraction of the mouse track mainly include: the method of extracting the features of the mouse track according to several fixed parameters such as the speed, distance, and number of clicks of the mouse track; As a method for extracting features of the mouse track; using a hierarchical division method, defining flexible standards for different features for different mouse behaviors, and extracting features of the mouse track, etc. However, in practice, due to the different scales and importance of the mouse trajectory features, and the local features of the mouse trajectory have ob...

AI Technical Summary

Problems solved by technology

However, in practice, due to the different scales and importance of the mouse trajectory features, and the local features of the mouse trajectory have obvious significance, the artificially specified features lack objectivity, and it is difficult to find general mouse trajectory features. Requirements for Mouse Track Feature Extraction

At present, based on the relevant theories of data mining and machine learning, the method of extracting mouse trajectory features has not yet appeared by using the fusion of local trajectory features and global trajectory features, and the fusion of multi-scale features.

Images