CN-121997188-A - Big data-based data behavior prevention and control and anomaly detection system

CN121997188ACN 121997188 ACN121997188 ACN 121997188ACN-121997188-A

Abstract

The invention discloses a data behavior prevention and control and abnormality detection system based on big data, and relates to the technical field of data security. The system comprises a data acquisition module, a characteristic construction module, a calculation module, a threshold judgment module and a model construction module, wherein the data acquisition module is used for acquiring user behavior data, the characteristic construction module is used for constructing emotion characteristics, space-time track characteristics and abnormal data characteristics according to the preprocessed user behavior data, the calculation module is used for calculating comprehensive scores according to the emotion characteristics, the space-time track characteristics and the abnormal data characteristics respectively, the threshold judgment module is used for carrying out threshold judgment on the comprehensive emotion scores, the comprehensive space-time scores and the comprehensive generated data scores respectively and generating a first abnormal signal, a second abnormal signal and a third abnormal signal when threshold conditions are exceeded, and the model construction module is used for constructing an abnormal detection model and generating a fourth abnormal signal when abnormality is detected. Through multidimensional feature fusion, real-time anomaly detection and a deep learning technology based on big data, the accuracy of anomaly behavior identification is improved, and the safety and reliability of the system are enhanced.

Inventors

TENG SHUHUA
CHEN CHENG
ZHANG RUI
ZHONG YUANYUAN
ZENG ZHUO
LI YAO
GAO FENG
ZHANG FENG
BAO YIPING

Assignees

山东协和学院
湖南大数据交易所有限公司

Dates

Publication Date: 20260508
Application Date: 20251224

Claims (10)

1. Big data-based data behavior prevention and control and anomaly detection system, which is characterized by comprising: The data acquisition module is used for acquiring user behavior data, including user interaction data, space-time track data and generated data; The feature construction module is used for preprocessing the user behavior data and constructing emotion features, space-time track features and abnormal data features according to the preprocessed user behavior data; the calculation module is used for calculating according to the emotion characteristics to obtain comprehensive emotion scores, calculating according to the space-time track characteristics to obtain comprehensive space-time scores, and calculating according to the abnormal data characteristics to obtain comprehensive generated data scores; The threshold judgment module is used for carrying out threshold judgment on the comprehensive emotion scores, generating a first abnormal signal when the threshold condition is exceeded, carrying out threshold judgment on the comprehensive space-time scores, generating a second abnormal signal when the threshold condition is exceeded, carrying out threshold judgment on the comprehensive generated data scores, and generating a third abnormal signal when the threshold condition is exceeded; The model construction module is used for constructing an anomaly detection model according to the emotion characteristics, the space-time track characteristics and the anomaly data characteristics, carrying out anomaly detection according to the anomaly detection model, and generating a fourth anomaly signal when the anomaly is detected.
2. The big data based data behavior prevention and anomaly detection system of claim 1, wherein emotional characteristics include emotional synchronicity, emotional ambiguity, emotional volatility, emotional coordination and emotional deviation; And calculating to obtain comprehensive emotion scores according to the emotion synchronicity, the emotion ambiguity, the emotion fluctuation rate, the emotion coordination schedule and the emotion deviation degree.
3. The big data based data behavior prevention and control and anomaly detection system of claim 2, wherein the emotion synchronicity calculation logic is: , wherein, Is emotion synchronism at time t, Is the emotion state of the user at the moment i, The emotional state of the external environment at the moment i comprises a system, a group and social media, Is that And An emotion angle difference of (2), As an indicator function, Is the time window size; The calculation logic of the emotion ambiguity is as follows: , wherein, An emotion ambiguity of time t, And Is emotion change, Is a regulating factor, Is an exponential decay factor; The calculation logic of the emotion fluctuation rate is as follows: , wherein, A mood fluctuation rate at time t, Sensitivity coefficient for controlling emotion fluctuation, Is a threshold value, Is a regulatory factor; The calculation logic of emotion coordination schedule is: , wherein, Is emotion cooperative scheduling at time t, An emotional state of a kth member in the group at a time i, Is the number of group members, To adjust the influence parameters of emotion matching degree, To adjust affective fluctuation influencing parameters, Parameters for controlling influence of emotion change on coordination degree; the calculation logic of the emotion deviation degree is as follows: , wherein, A mood deviation degree at time t, Expected emotional state for the system and society at time i, Is the nonlinear amplification coefficient of the emotion deviation degree, To control sensitivity of affective differences to deviation; The calculation logic of the comprehensive emotion score is as follows: , wherein, 、、、 And Is the weight coefficient of the emotion parameter.
4. The big data based data behavior prevention and anomaly detection system of claim 1, wherein the spatiotemporal trajectory features include a spatiotemporal relevance increasing index, a spatiotemporal trajectory anomaly offset, a spatiotemporal density mutual entropy, a spatiotemporal nonlinear ringing oscillation value, and a spatiotemporal adaptive jump threshold; And calculating to obtain the comprehensive space-time score according to the space-time correlation increment index, the space-time track abnormal offset, the space-time density mutual entropy, the space-time nonlinear oscillation value and the space-time self-adaptive jump threshold value.
5. The big data based data behavior prevention and anomaly detection system of claim 4, wherein the calculation logic of the spatio-temporal correlation increment index is: , wherein, Increasing the index for the time-space relevance, Is the time point i, Is the spatial position of the time point i, To adjust the time and space varying speed index; the calculation logic of the abnormal deviation degree of the space-time track is as follows: , wherein, Is the abnormal deviation of the space-time track, Is the track start time, Is the initial space position of the track, And Is an exponential coefficient, Is an abnormal behavior regulating factor; The calculation logic of the time density entropy is as follows: , wherein, Is the time density entropy, For the behavior density of the user in the time i and the space region j, And Edge densities over time period i and spatial region j, respectively, And Discretized components of time and space, respectively; The calculation logic of the space-time nonlinear dynamic oscillation value is as follows: , wherein, Is a space-time nonlinear dynamic oscillation value, To control the frequency of the time oscillation, For the previous time point space position, Is the maximum possible space deviation, To control the oscillation sensitivity parameter; the calculation logic of the space-time self-adaptive jump threshold value is as follows: , wherein, Is a space-time self-adaptive jump threshold value, And Parameters for controlling the influence of jump amplitude, Attenuation factors for controlling time body hopping; the calculation path logic of the comprehensive space-time score is as follows: , wherein, 、、、 And Is a weight coefficient of the space-time parameter.
6. The big data based data behavior prevention and control and anomaly detection system of claim 1, wherein the anomaly data features include generating a data blur factor, generating a data fluctuation factor, script period anomaly strength, and automated behavior disorder; And calculating according to the generated data fuzzy factor, the generated data fluctuation factor, the abnormal strength of the script period and the automatic behavior disorder degree to obtain the comprehensive generated data score.
7. The big data based data behavior prevention and anomaly detection system of claim 6, wherein the computation logic to generate the data blur factor is: , wherein, To generate a data blurring factor, Is the ith feature, Is characterized by A second order gradient of (2), Is characterized by Information entropy of (2), To adjust the ith feature ambiguity sensitivity parameter, features Information entropy of (2) The calculation logic of (1) is: , wherein, Is characterized by Specific value of (3) Probability of (2), Is characterized by Specific value of (3) Logarithm of probability of (2); the calculation logic for generating the data fluctuation factor is: , wherein, To generate a data fluctuation factor, Is characterized by A variation in the time dimension, Is characterized by A variation in spatial dimension, Is characterized by The variation in semantic dimension, Is a normalized characteristic quantity, 、 And To adjust the dimensional change sensitivity constant; the calculation logic of the abnormal strength of the script period is as follows: , wherein, For the intensity of the script period anomaly, Is the characteristic of the behavior sequence at the time t, Is an equilibrium value of behaviors in the whole time range, Is in the range of time period, A periodic sensitivity constant for adjusting behavior variation; the calculation logic of the automatic behavior disorder degree is as follows: , wherein, Is an automatic behavior disorder degree, A state sequence of the ith behavior, Entropy value of behavior sequence, Is the duration of the behavior, For entropy to be a sensitivity constant of behavior randomness, the entropy value of behavior sequence The calculation logic of (1) is: , wherein, Is in state of Probability of occurrence, Is the total number of different states in the behavior sequence, Is in state of Logarithm of probability of occurrence; the calculation logic for comprehensively generating the data scores is as follows: , wherein, 、、 And Is a weight coefficient.
8. The big data-based data behavior prevention and control and anomaly detection system of claim 1, wherein the threshold value determination is performed on the composite emotion score, the first anomaly signal is generated when the composite emotion score is greater than a preset composite emotion score threshold value, the threshold value determination is performed on the composite space-time score, the second anomaly signal is generated when the composite space-time score is greater than the preset composite space-time score threshold value, the threshold value determination is performed on the composite generated data score, and the third anomaly signal is generated when the composite generated data score is greater than the preset composite generated data score threshold value.
9. The big data based data behavior prevention and anomaly detection system of claim 1, wherein the anomaly detection model construction logic comprises: acquiring historical user behavior data and corresponding labels, constructing emotion characteristics, space-time track characteristics and abnormal data characteristics according to the historical user behavior data, and setting the emotion characteristics, space-time track characteristics and abnormal data characteristics as a data set, wherein the labels comprise normal and abnormal; carrying out frequency association analysis on the emotion characteristics, the space-time track characteristics and the abnormal data characteristics, and setting the characteristics corresponding to the maximum value obtained through calculation as the construction characteristics of the decision tree; constructing a feature according to the decision tree, and constructing an anomaly detection model according to a decision tree algorithm; The data set is divided into a training set and a testing set, the training set is used for training the constructed abnormality detection model, the verification set is used for verifying the abnormality detection model after training, after verification, the user behavior data are input into the abnormality detection model, the abnormality analysis is carried out on the user behavior, and a fourth abnormality signal is generated after the abnormal behavior is detected.
10. The big data based data behavior prevention and anomaly detection system of claim 9, wherein the computational logic of the frequency correlation analysis is: , wherein, Is a frequency correlation statistic, To observe the frequency, Is the desired frequency.

Description

Big data-based data behavior prevention and control and anomaly detection system Technical Field The invention relates to the technical field of data security, in particular to a data behavior prevention and control and abnormality detection system based on big data. Background With the rapid development of big data technology, more and more fields begin to utilize big data analysis to make intelligent decisions and behavior predictions. Particularly in the fields of user behavior prevention and control and anomaly detection, how to efficiently identify the abnormal behavior and take measures in time is a key for improving the safety of a system and user experience. Conventional anomaly detection methods rely mainly on rule-based detection and simple statistical analysis, however, these methods often fail to accurately identify complex anomaly patterns in the face of complex, multidimensional and large-scale data. In particular, the limitations of the conventional methods are increasingly manifested when various features such as emotion, behavior trace, and automated behavior data of the user are involved. Disclosure of Invention Based on the above-mentioned drawbacks of the prior art, an object of the present invention is to provide a data behavior control and anomaly detection system based on big data, so as to solve the above-mentioned technical problems. In order to achieve the above purpose, the invention provides a data behavior prevention and control and abnormality detection system based on big data, comprising: The data acquisition module is used for acquiring user behavior data, including user interaction data, space-time track data and generated data; The feature construction module is used for preprocessing the user behavior data and constructing emotion features, space-time track features and abnormal data features according to the preprocessed user behavior data; the calculation module is used for calculating according to the emotion characteristics to obtain comprehensive emotion scores, calculating according to the space-time track characteristics to obtain comprehensive space-time scores, and calculating according to the abnormal data characteristics to obtain comprehensive generated data scores; The threshold judgment module is used for carrying out threshold judgment on the comprehensive emotion scores, generating a first abnormal signal when the threshold condition is exceeded, carrying out threshold judgment on the comprehensive space-time scores, generating a second abnormal signal when the threshold condition is exceeded, carrying out threshold judgment on the comprehensive generated data scores, and generating a third abnormal signal when the threshold condition is exceeded; The model construction module is used for constructing an anomaly detection model according to the emotion characteristics, the space-time track characteristics and the anomaly data characteristics, carrying out anomaly detection according to the anomaly detection model, and generating a fourth anomaly signal when the anomaly is detected. The invention is further configured such that the emotional characteristics include emotional synchronicity, emotional ambiguity, emotional fluctuation rate, emotional coordination and emotional deviation; And calculating to obtain comprehensive emotion scores according to the emotion synchronicity, the emotion ambiguity, the emotion fluctuation rate, the emotion coordination schedule and the emotion deviation degree. The invention further provides that the emotion synchronicity calculation logic is as follows: , wherein, Is emotion synchronism at time t,Is the emotion state of the user at the moment i,The emotional state of the external environment at the moment i comprises a system, a group and social media,Is thatAndAn emotion angle difference of (2),As an indicator function,Is the time window size; The calculation logic of the emotion ambiguity is as follows: , wherein, An emotion ambiguity of time t,AndIs emotion change,Is a regulating factor,Is an exponential decay factor; The calculation logic of the emotion fluctuation rate is as follows: , wherein, A mood fluctuation rate at time t,Sensitivity coefficient for controlling emotion fluctuation,Is a threshold value,Is a regulatory factor; The calculation logic of emotion coordination schedule is: , wherein, Is emotion cooperative scheduling at time t,An emotional state of a kth member in the group at a time i,Is the number of group members,To adjust the influence parameters of emotion matching degree,To adjust affective fluctuation influencing parameters,Parameters for controlling influence of emotion change on coordination degree; the calculation logic of the emotion deviation degree is as follows: , wherein, A mood deviation degree at time t,Expected emotional state for the system and society at time i,Is the nonlinear amplification coefficient of the emotion deviation degree,To control sensitivity of affective differ