Method for detecting abnormal user behaviours mined on the basis of variable-length sequence mode
A technology of pattern mining and detection methods, applied in the field of data analysis, it can solve the problems of offline analysis but cannot accurately describe the complex behavior of users, and achieve the effect of improving accuracy and real-time performance.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0029] figure 1 It is a flow chart of an embodiment of the user abnormal behavior detection method based on variable-length sequence pattern mining provided by the present invention, as shown in figure 1 As shown, the method includes two stages, namely:
[0030] The first stage 1: the user's normal behavior training stage, this stage is mainly to use the user's historical behavior data modeling in the offline system to calculate the user's normal behavior pattern;
[0031] The second stage 2: user abnormal behavior detection stage, this stage is mainly to extract the user's current behavior pattern in the online system and match it with the normal behavior pattern in the database to see if the current behavior is abnormal.
[0032] Specifically, the normal user behavior training phase includes the following steps:
[0033] Step 11. Preprocessing the user’s normal behavior log data in the database to obtain multiple user normal behavior variable-length sequence streams; in th...
Embodiment 2
[0041] figure 2 The flow chart of Embodiment 2 of the abnormal user behavior detection method based on variable-length sequence pattern mining provided by the present invention, as shown in figure 2 As shown, on the basis of Embodiment 1, Embodiment 2 further includes:
[0042] Step 15, on the basis of constructing and generating the user's normal behavior pattern by each user's normal behavior variable-length sequence flow and the number of occurrences thereof, calculate the IDF (InverseDocumentFrequency) value of each user's normal behavior variable-length sequence flow, and according to the The IDF value updates the user's normal behavior pattern to obtain an optimized user's normal behavior pattern. The IDF value reflects the importance of a sequence. The higher the IDF value of a short sequence, the more important the sequence is to the user and the higher its recognition degree, that is, the current user can be distinguished from other users through this sequence; the...
Embodiment 3
[0044] image 3 The flow chart of Embodiment 3 of the abnormal user behavior detection method based on variable-length sequence pattern mining provided by the present invention, as shown in image 3 As shown, on the basis of the above-mentioned embodiments, this embodiment also includes:
[0045] Step 22: Calculate the IDF value of each variable-length sequence of user behavior to be detected, and if it is lower than the predetermined IDF threshold, delete the variable-length sequence of user behavior to omit the judgment of the variable-length sequence of user behavior. This step can screen many unnecessary tests and improve the detection efficiency. Furthermore, different predetermined IDF thresholds can be set correspondingly according to different sequence lengths in the user behavior variable-length sequence. When judging, when all the judgment values are greater than the IDF value of its corresponding length, it is judged as the normal behavior of the user. Further p...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com