CN118245734B - POS machine data intelligent processing method based on 5G technology - Google Patents
POS machine data intelligent processing method based on 5G technology Download PDFInfo
- Publication number
- CN118245734B CN118245734B CN202410650488.6A CN202410650488A CN118245734B CN 118245734 B CN118245734 B CN 118245734B CN 202410650488 A CN202410650488 A CN 202410650488A CN 118245734 B CN118245734 B CN 118245734B
- Authority
- CN
- China
- Prior art keywords
- preset
- transaction
- characteristic
- index
- density
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 15
- 238000000034 method Methods 0.000 claims abstract description 46
- 230000002159 abnormal effect Effects 0.000 claims abstract description 24
- 238000001514 detection method Methods 0.000 claims abstract description 21
- 239000013598 vector Substances 0.000 claims description 36
- 230000005856 abnormality Effects 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 9
- 239000000463 material Substances 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 abstract description 2
- 238000000513 principal component analysis Methods 0.000 description 16
- 230000000694 effects Effects 0.000 description 7
- 230000001419 dependent effect Effects 0.000 description 4
- 230000007423 decrease Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 231100000279 safety data Toxicity 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Technology Law (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of transaction data anomaly monitoring, in particular to a POS machine data intelligent processing method based on a 5G technology. According to the method, firstly, distinguishing characteristic values of a preset transaction index are obtained according to density distribution conditions of all transaction data points in the dimension of the preset transaction index and centralized distribution characteristic values of the preset transaction index; acquiring relevant characteristic values of two preset transaction indexes according to the distribution condition of all transaction data points in the dimensions of the two preset transaction indexes and the difference of distinguishing characteristic values between the two preset transaction indexes; and further adjusting covariance coefficients of two preset transaction indexes to obtain an abnormal detection result of the transaction data to be analyzed. According to the invention, the covariance coefficient is adjusted based on the importance and the correlation of the preset transaction index, so that the adjusted covariance coefficient can reflect the inherent importance and the correlation of the preset transaction index, and the accuracy of the abnormal detection result of the transaction data to be analyzed is ensured.
Description
Technical Field
The invention relates to the technical field of transaction data anomaly monitoring, in particular to a POS machine data intelligent processing method based on a 5G technology.
Background
The POS machine is used as a key device for processing sales transactions, and transaction information recorded in POS machine data is crucial to merchants. With the popularization and development of 5G technology, POS machines meet new development opportunities. The 5G technology not only improves the speed and stability of data transmission, but also enhances the ability of data processing and analysis, enabling the POS to process large amounts of data more efficiently. Because POS machine data relates to sensitive information such as transaction amount, customer information and the like, once attacked or tampered, the POS machine data can cause loss of merchants and influence the trust degree of customers.
In the prior art, the PCA (PRINCIPAL COMPONENT ANALYSIS ) algorithm is utilized to perform anomaly detection on transaction data to be analyzed of the POS machine, and anomaly scores of all data points are obtained, so that abnormal data points deviating from a normal mode in the transaction data to be analyzed of the POS machine are analyzed. However, when the PCA algorithm is used for carrying out anomaly detection on the transaction data to be analyzed of the POS machine in the prior art, the numerical values of the transaction data points have larger differences under the dimensions of different types of preset transaction indexes, so that the covariance coefficient is difficult to accurately reflect the inherent importance and correlation of the preset transaction indexes, and the anomaly detection result of the POS machine data is inaccurate.
Disclosure of Invention
In order to solve the technical problems that when the PCA algorithm is utilized to carry out abnormal detection on transaction data to be analyzed of the POS machine in the prior art, the numerical values of transaction data points have larger differences under the dimensions of different types of preset transaction indexes, so that the covariance coefficient is difficult to accurately reflect the inherent importance and relativity of the preset transaction indexes, and the abnormal detection result of the POS machine data is inaccurate, the invention aims to provide the intelligent processing method for the POS machine data based on the 5G technology, and the adopted technical scheme is as follows:
a POS machine data intelligent processing method based on a 5G technology comprises the following steps:
Acquiring transaction data to be analyzed; the transaction data to be analyzed comprises all transaction data points; each transaction data point corresponds to the dimension of all preset transaction indexes;
In transaction data to be analyzed, acquiring a centralized distribution characteristic value of a preset transaction index according to the main characteristic distribution condition of all transaction data points in the dimension of the preset transaction index; acquiring a distinguishing characteristic value of a preset transaction index according to the density distribution condition of all transaction data points in the dimension of the preset transaction index and the centralized distribution characteristic value of the preset transaction index; acquiring related characteristic values of two preset transaction indexes according to the distribution condition of all the transaction data points in the dimensions of the two preset transaction indexes and the difference of the distinguishing characteristic values between the two preset transaction indexes;
Obtaining covariance coefficients of two preset transaction indexes; according to the distinguishing characteristic values and the related characteristic values of the two preset transaction indexes, the covariance coefficients of the two preset transaction indexes are adjusted, and adjusted covariance coefficients of the two preset transaction indexes are obtained; and acquiring an abnormality detection result of the transaction data to be analyzed according to the adjusted covariance coefficients of every two preset transaction indexes.
Further, the method for acquiring the centralized distribution characteristic value comprises the following steps:
Under the dimension of a preset trade index, taking the section of the preset trade index where all trade data points with the value of the preset trade index being larger than a preset first quantile threshold and smaller than a preset second quantile threshold are located as the main characteristic section of the preset trade index; taking the interval of all transaction data points in the preset transaction index as a reference characteristic interval of the preset transaction index;
calculating the ratio of the length of the main characteristic interval to the length of the reference characteristic interval, and carrying out negative correlation normalization on the comparison value to obtain the centralized distribution characteristic value of the preset transaction index.
Further, the method for acquiring the centralized distribution characteristic value comprises the following steps:
obtaining the centralized distribution characteristic value according to a centralized distribution characteristic value formula, wherein the centralized distribution characteristic value formula comprises: ; wherein, The concentrated distribution characteristic value of the a-th preset transaction index; The main characteristic interval of the a-th preset transaction index; the reference characteristic interval is the a-th preset transaction index; The length of the main characteristic interval of the a-th preset transaction index; the length of the reference characteristic interval of the a-th preset transaction index; is an exponential function based on e.
Further, the method for acquiring the distinguishing characteristic value comprises the following steps:
in the reference characteristic interval of the preset transaction index, an interval except the main characteristic interval is used as the remaining characteristic interval of the preset transaction index;
Taking the main characteristic interval and the residual characteristic interval as density characteristic intervals to be analyzed of the preset transaction indexes; obtaining a density maximum value, a density minimum value and a mean value density of the density characteristic interval to be analyzed;
Acquiring an attribute density vector of the density characteristic interval to be analyzed according to the density maximum value, the density minimum value and the average value density of the density characteristic interval to be analyzed;
And under the dimension of a preset transaction index, according to the difference between the attribute density vector of a main characteristic interval and the attribute density vector of the residual characteristic interval, the centralized distribution characteristic value of the preset transaction index is obtained, and the difference between the average value density of the main characteristic interval and the average value density of the residual characteristic interval is obtained.
Further, the method for acquiring the distinguishing characteristic value comprises the following steps:
acquiring the distinguishing characteristic value according to a distinguishing characteristic value formula, wherein the acquiring formula of the distinguishing characteristic value comprises the following steps:
; wherein, A distinguishing characteristic value of the preset transaction index of the a-th type; the concentrated distribution characteristic value of the a-th preset transaction index; The main characteristic interval of the a-th preset transaction index; the remaining characteristic interval of the a-th preset transaction index; Is the main characteristic interval Is a vector of attribute densities; For the remaining characteristic interval Is a vector of attribute densities; Is the main characteristic interval Is a mean density of (1); For the remaining characteristic interval Is a mean density of (1); Is a modulus of the vector; Is an absolute value symbol; Is a normalization function.
Further, the method for obtaining the attribute density vector comprises the following steps:
acquiring the attribute density vector according to an attribute density vector formula, wherein the acquisition formula of the attribute density vector comprises the following steps:
; wherein, For the density characteristic interval to be analyzedIs a vector of attribute densities; A density characteristic interval to be analyzed which is the a-th preset transaction index; For the density characteristic interval to be analyzed Is a maximum value of the density of the material; For the density characteristic interval to be analyzed Is a minimum of the density of (1); For the density characteristic interval to be analyzed Is a mean value density of (1).
Further, the method for acquiring the relevant characteristic value comprises the following steps:
under the dimension of two preset trade indexes, taking the trade data points which are simultaneously positioned in the main characteristic intervals of the two preset trade indexes as main repeated data points of the two preset trade indexes; taking transaction data points which are simultaneously positioned in the residual characteristic intervals of the two preset transaction indexes as residual repeated data points of the two preset transaction indexes;
acquiring the relevant characteristic value according to a relevant characteristic value formula, wherein the acquiring formula of the relevant characteristic value comprises the following steps:
; wherein, The related characteristic values of the m-th preset transaction index and the n-th preset transaction index are obtained; A number of primary repeat data points for an mth of said preset transaction indicators and an nth of said preset transaction indicators; the main characteristic interval of the m-th preset transaction index; Is the main characteristic interval A total number of all transaction data points; the main characteristic interval of the nth preset transaction index; Is the main characteristic interval A total number of all transaction data points; a number of said remaining duplicate data points for an mth said preset transaction indicator and an nth said preset transaction indicator; the remaining characteristic interval of the m-th preset transaction index; For the remaining characteristic interval A total number of all transaction data points; The remaining characteristic interval of the n-th preset transaction index; For the remaining characteristic interval A total number of all transaction data points; The m-th distinguishing characteristic value of the preset transaction index; the distinguishing characteristic value of the nth preset transaction index; Is a normalization function.
Further, the method for acquiring the adjusted covariance coefficient comprises the following steps:
Acquiring adjusted covariance coefficients of the two preset trade indexes according to the distinguishing characteristic values of the two preset trade indexes, the correlation characteristic values of the two preset trade indexes and the covariance coefficients of the two preset trade indexes; the distinguishing characteristic value, the correlation characteristic value, the covariance coefficient and the adjusted covariance coefficient are in positive correlation.
Further, the method for obtaining the covariance coefficients of the two preset transaction indexes comprises the following steps:
based on PCA algorithm, covariance coefficients of two preset transaction indexes are obtained.
Further, the method for obtaining the abnormality detection result includes:
Based on a PCA algorithm, acquiring an abnormal score of each transaction data point in the transaction data to be analyzed according to the adjusted covariance coefficients of each two preset transaction indexes; and acquiring an abnormality detection result of the transaction data to be analyzed according to the abnormality score of each transaction data point.
The invention has the following beneficial effects:
According to the invention, through the importance of the preset trade indexes and the correlation between the preset trade indexes, the covariance coefficients of the two preset trade indexes are further adjusted, and the covariance coefficients capable of reflecting the inherent importance and the correlation of the preset trade indexes are determined. In order to analyze the importance of the preset trade index, the main data distribution concentration condition under the dimension of the preset trade index is reflected through the concentrated distribution characteristic value of the preset trade index. Considering that the density of the main data is similar to the density difference of the residual data, the more likely that the residual data is special conditions to cause the occurrence of edge data, the less likely that the residual data is edge data generated by abnormal data, and the less likely that the preset transaction index has better distinguishing capability on the abnormal data. Considering that the larger the concentrated distribution characteristic value of the preset transaction index is, the more concentrated the main data distribution is represented under the dimension of the preset transaction index, the better the preset transaction index has the distinguishing capability on the data at the edge, and the distinguishing characteristic value of the preset transaction index is obtained; the larger the distinguishing characteristic value is, the better distinguishing effect of the preset trading index on the trading data points is represented, and the more important the preset trading index is. In order to analyze the correlation between the preset trade indexes, the smaller the difference between the data density characteristics and the distribution characteristics of the two preset trade indexes is, the stronger the correlation between the two preset trade indexes is considered; taking into consideration that the more the main data overlapping parts and the more the residual data overlapping parts of the two preset transaction indexes are, the stronger the correlation of the two preset transaction indexes is represented, and acquiring the correlation characteristic values of the two preset transaction indexes; the larger the correlation characteristic value is, the stronger the correlation between the two preset transaction indexes is represented. The covariance coefficient is adjusted based on the importance and the correlation of the preset transaction indexes, so that the adjusted covariance coefficient can reflect the inherent importance and the correlation of the preset transaction indexes, the problem of the difference of the data sizes of different preset transaction indexes can be more effectively processed, and the accuracy of the abnormal detection result of the transaction data to be analyzed is ensured.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a method for intelligent processing of POS machine data based on 5G technology according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a feature space coordinate system according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of specific implementation, structure, characteristics and effects of the 5G technology-based POS machine data intelligent processing method according to the invention in combination with the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The invention provides a specific scheme of a POS machine data intelligent processing method based on a 5G technology, which is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a 5G technology-based intelligent POS machine data processing method according to an embodiment of the invention is shown, where the method includes the following steps:
Step S1: acquiring transaction data to be analyzed; the transaction data to be analyzed comprises all transaction data points; each transaction data point corresponds to a dimension of all preset transaction indicators.
The method comprises the steps of obtaining a transaction record stored in any POS machine of a product vendor from a database of the POS machines of the product vendor, wherein the transaction record represents an actual transaction order of a customer for transaction at the product vendor and settlement at a corresponding POS machine, each transaction record comprises all preset transaction indexes, and in one embodiment, the preset transaction indexes comprise transaction place codes, transaction time, average price of transaction goods, average discount of the transaction goods and the like. In order to analyze abnormal transaction records, the distribution of all transaction records in each preset transaction index needs to be studied, and multidimensional transaction data to be analyzed needs to be constructed first.
Preferably, in one embodiment of the present invention, the method for acquiring transaction data to be analyzed includes:
Constructing a characteristic space coordinate system, wherein each axis of the characteristic space coordinate system represents the dimension of each preset transaction index; in the feature space coordinate system, taking all preset transaction indexes corresponding to each transaction record as a transaction data point and taking all the transaction data points as transaction data to be analyzed. Wherein each transaction data point represents a transaction record, and the transaction data to be analyzed reflects all transaction records stored by a POS machine of a product vendor. Fig. 2 is a schematic diagram of a feature space coordinate system according to an embodiment of the present invention, wherein three axes of the feature space coordinate system correspond to a transaction location code, a transaction time and a transaction commodity average unit price, respectively, and fig. 2 is only a schematic diagram for helping understanding the feature space coordinate system, and corresponds to all preset transaction indexes as follows: trade site coding, trade time, and trade commodity average unit price.
It should be noted that, in order to facilitate the operation, all index data involved in the operation in the embodiment of the present invention is subjected to data preprocessing, so as to cancel the dimension effect. The specific means for removing the dimension influence is a technical means well known to those skilled in the art, and is not limited herein.
Step S2: in transaction data to be analyzed, acquiring a centralized distribution characteristic value of a preset transaction index according to the main characteristic distribution condition of all transaction data points in the dimension of the preset transaction index; acquiring a distinguishing characteristic value of a preset transaction index according to the density distribution condition of all transaction data points in the dimension of the preset transaction index and the centralized distribution characteristic value of the preset transaction index; and acquiring related characteristic values of the two preset transaction indexes according to the distribution condition of all the transaction data points in the dimensions of the two preset transaction indexes and the difference of the distinguishing characteristic values between the two preset transaction indexes.
The core of the PCA algorithm is considered to be the principal component in the transaction data to be analyzed of the POS machine. The PCA algorithm is more prone to selecting preset transaction indexes which are important and highly relevant as main components, so that the influence of noise and redundant information is reduced, and the accuracy and reliability of an abnormal detection result of the transaction data to be analyzed of the POS machine are improved. Because the numerical values of the transaction data points have larger differences in the dimensions of different types of preset transaction indexes, the preset transaction indexes with larger numerical values may occupy larger weights in the covariance coefficients, thereby influencing the determination of the principal component by the PCA. Principal components analyzed by the PCA algorithm are biased towards preset trade indexes with larger values, and preset trade indexes with smaller values are ignored. In the prior art, different kinds of transaction indexes are preprocessed, such as standardized or normalized, so that scale differences among different preset transaction indexes are eliminated, however, the covariance coefficient determined by the method is difficult to reflect the inherent importance and the correlation of the preset transaction indexes. According to the invention, through the importance of the preset trade indexes and the correlation between the preset trade indexes, the covariance coefficients of the two preset trade indexes are further adjusted, and the covariance coefficients capable of reflecting the inherent importance and the correlation of the preset trade indexes are determined.
In order to analyze the importance of the preset trade index, the main data distribution concentration condition under the dimension of the preset trade index is reflected through the concentrated distribution characteristic value of the preset trade index. Considering that the density of the main data is similar to the density difference of the residual data, the more likely that the residual data is special conditions to cause the occurrence of edge data, the less likely that the residual data is edge data generated by abnormal data, and the less likely that the preset transaction index has better distinguishing capability on the abnormal data. Considering that the larger the concentrated distribution characteristic value of the preset transaction index is, the more concentrated the main data distribution is represented under the dimension of the preset transaction index, the better the preset transaction index has the distinguishing capability on the data at the edge, and the distinguishing characteristic value of the preset transaction index is obtained; the larger the distinguishing characteristic value is, the better distinguishing effect of the preset trading index on the trading data points is represented, and the more important the preset trading index is.
In order to analyze the correlation between the preset trade indexes, the smaller the difference between the data density characteristics and the distribution characteristics of the two preset trade indexes is, the stronger the correlation between the two preset trade indexes is considered; taking into consideration that the more the main data overlapping parts and the more the residual data overlapping parts of the two preset transaction indexes are, the stronger the correlation of the two preset transaction indexes is represented, and acquiring the correlation characteristic values of the two preset transaction indexes; the larger the correlation characteristic value is, the stronger the correlation between the two preset transaction indexes is represented.
Preferably, in order to analyze the importance of the preset trade index, first, the main data distribution concentration condition of the trade data point in the dimension of the preset trade index is analyzed, and in one embodiment of the present invention, the method for acquiring the concentrated distribution characteristic value includes:
Under the dimension of a preset trade index, acquiring the corresponding quantile of each trade data point in the preset trade index; taking the section of the preset transaction index where all transaction data points with the value of the preset transaction index being larger than the preset first quantile threshold and smaller than the preset second quantile threshold are located as the main characteristic section of the preset transaction index; taking the interval of all the transaction data points in the preset transaction index as a reference characteristic interval of the preset transaction index; the reference characteristic interval reflects the interval in which all transaction data points are located in the dimension of the preset transaction index. In one embodiment of the present invention, the first quantile threshold value is preset to be 5% quantile, the second quantile threshold value is preset to be 95% quantile, and the implementation can be set by the implementation according to the implementation scenario. In other embodiments of the present invention, the first threshold quantile is preset to be 10% quantile, the second threshold quantile is preset to be 90% quantile, and the implementation person can set the implementation scenario by himself. It should be noted that, the method for obtaining the quantiles is a technical means well known to those skilled in the art, and not described herein, the quantiles of the transaction data points reflect the distribution situation of the transaction data points, for example, under the dimension of the preset transaction index, the corresponding 5% quantiles of the transaction data points means that the value of the preset transaction index having 5% of the transaction data points under the dimension of the preset transaction index is less than or equal to 5% of the quantiles, and the value of the preset transaction index having 95% of the transaction data points is greater than 5% of the quantiles.
Calculating the ratio of the length of the main characteristic interval to the length of the reference characteristic interval, and carrying out negative correlation normalization on the comparison value to obtain the centralized distribution characteristic value of the preset transaction index;
the negative correlation represents that the dependent variable decreases with increasing independent variable, and the dependent variable increases with decreasing independent variable, and may be a subtraction relationship, a division relationship, or the like, which is determined by practical application.
Obtaining the centralized distribution characteristic value according to a centralized distribution characteristic value formula, wherein the centralized distribution characteristic value formula comprises:
; wherein, The concentrated distribution characteristic value of the a-th preset transaction index; The main characteristic interval of the a-th preset transaction index; the reference characteristic interval is the a-th preset transaction index; The length of the main characteristic interval of the a-th preset transaction index; the length of the reference characteristic interval of the a-th preset transaction index; is an exponential function based on e.
In the formula, considering that the main data distribution is more concentrated under the dimension of the preset transaction index, the preset transaction index has better distinguishing capability on the data of the edge, for example, the position of any POS machine of a product vendor is fixed, the place number is determined, and if different transaction place numbers occur, the abnormal data are usually generated; the more concentrated the main data distribution in the place coding dimension, the less different places appear on the representation, and the preset transaction index has better distinguishing capability on the edge data.The dispersion degree of the main characteristic interval is reflected, and the smaller the value is, the more concentrated the main data distribution is represented under the dimension of the preset transaction index; The concentration degree of the main characteristic interval is reflected, the larger the value is, the more concentrated the main data distribution is in the dimension of the preset transaction index, the more concentrated the concentrated distribution characteristic value is reflected in the main data distribution concentration condition in the dimension of the preset transaction index, and the better the preset transaction index has the capability of distinguishing the edge data.
Preferably, in order to analyze the importance of the preset transaction index, the distinguishing effect of the preset transaction index on the transaction data points needs to be known, and in one embodiment of the present invention, the method for obtaining the distinguishing feature value includes:
Taking the section except the main characteristic section as the rest characteristic section of the preset transaction index in the reference characteristic section of the preset transaction index; the remaining feature interval reflects the interval in which the edge data point is located.
Taking the main characteristic interval and the residual characteristic interval as density characteristic intervals to be analyzed of the preset transaction indexes; obtaining a density maximum value, a density minimum value and a mean value density of the density characteristic interval to be analyzed; the specific acquisition means are as follows: in one embodiment of the present invention, the density value of each transaction data point may be obtained by using a K-nearest neighbor method, and in other embodiments of the present invention, the density value of each transaction data point may be obtained according to a kernel density estimation method, where the method for obtaining the density value is a well known technology of those skilled in the art, and is not limited herein. Taking the maximum density value as the density maximum value of the density characteristic interval to be analyzed in the density characteristic interval to be analyzed; taking the minimum density value as the density minimum value of the density characteristic interval to be analyzed; and taking the average value of the density values of all the transaction data points as the average value density of the density characteristic interval to be analyzed.
Acquiring the attribute density vector of the density characteristic interval to be analyzed according to the density maximum value, the density minimum value and the average value density of the density characteristic interval to be analyzed;
acquiring the attribute density vector according to an attribute density vector formula, wherein the acquisition formula of the attribute density vector comprises the following steps:
; wherein, For the density characteristic interval to be analyzedIs a vector of attribute densities; A density characteristic interval to be analyzed which is the a-th preset transaction index; For the density characteristic interval to be analyzed Is a maximum value of the density of the material; For the density characteristic interval to be analyzed Is a minimum of the density of (1); For the density characteristic interval to be analyzed Is a mean value density of (1).
In the method, in the process of the invention,Reflecting the difference condition of the density maximum value relative to the common density in the density characteristic interval to be analyzed,And reflecting the difference condition of the density minimum value relative to the common density in the density characteristic interval to be analyzed, wherein the attribute density vector comprehensively reflects the fluctuation condition of the density distribution in the density characteristic interval to be analyzed.
Under the dimension of a preset transaction index, according to the difference between the attribute density vector of a main characteristic interval and the attribute density vector of the residual characteristic interval, the centralized distribution characteristic value of the preset transaction index is obtained, and the difference between the average value density of the main characteristic interval and the average value density of the residual characteristic interval is obtained;
acquiring the distinguishing characteristic value according to a distinguishing characteristic value formula, wherein the acquiring formula of the distinguishing characteristic value comprises the following steps:
; wherein, A distinguishing characteristic value of the preset transaction index of the a-th type; the concentrated distribution characteristic value of the a-th preset transaction index; The main characteristic interval of the a-th preset transaction index; the remaining characteristic interval of the a-th preset transaction index; Is the main characteristic interval Is a vector of attribute densities; For the remaining characteristic interval Is a vector of attribute densities; Is the main characteristic interval Is a mean density of (1); For the remaining characteristic interval Is a mean density of (1); Is a modulus of the vector; Is an absolute value symbol; Is a normalization function.
In the formula, considering that the special situation and the normal situation have large data difference, the special situation can also generate edge data, and the preset transaction index is difficult to distinguish the abnormal data at the moment; for example, for the dimension of the average price of the commodity of the transaction data to be analyzed of the fire extinguisher manufacturer, the fixed commodity is sold by the fire extinguisher manufacturer, the average selling price of the sold commodity is similar, the commodity is discounted and sold under special conditions, so that fluctuation of the average price is caused, at the moment, the edge data are likely to be edge data generated by discounted transaction, the distribution condition of the edge data is similar to that of the main data, a large difference exists between the edge data and the main data, and the preset transaction index is difficult to distinguish the abnormal data.Reflecting the main characteristic intervalIs defined as the attribute density vector and the remaining feature intervalThe larger the difference between the attribute density vectors is, the more obvious the density distribution difference between main data and residual data is, and the better the preset transaction index has the capability of distinguishing the edge data is; Reflecting the main characteristic interval Is defined as the attribute density vector and the remaining feature intervalThe smaller the difference is, the more likely the residual data is special conditions to cause edge data, the less likely the residual data is edge data generated by abnormal data, and the less the preset transaction index has better distinguishing ability for the abnormal data; the centralized distribution characteristic value reflects the main data distribution centralized condition under the dimension of the preset transaction index, and the more centralized, the better the preset transaction index has the better distinguishing capability for the edge data; the distribution difference condition between the general densities of the main data and the residual data is reflected, and the more obvious the difference is, the greater the density distribution difference condition of the whole main data and the whole residual data is, and the better the preset transaction index has the capability of distinguishing the edge data is; the larger the distinguishing characteristic value is, the better the distinguishing effect of the preset trading index on the trading data points is, and the more important the preset trading index is.
Preferably, in order to analyze the correlation between two preset transaction indexes, in one embodiment of the present invention, the method for acquiring the correlation characteristic value includes:
under the dimension of two preset trade indexes, taking the trade data points which are simultaneously positioned in the main characteristic intervals of the two preset trade indexes as main repeated data points of the two preset trade indexes; taking transaction data points which are simultaneously positioned in the residual characteristic intervals of the two preset transaction indexes as residual repeated data points of the two preset transaction indexes;
acquiring the relevant characteristic value according to a relevant characteristic value formula, wherein the acquiring formula of the relevant characteristic value comprises the following steps:
; wherein, The related characteristic values of the m-th preset transaction index and the n-th preset transaction index are obtained; A number of primary repeat data points for an mth of said preset transaction indicators and an nth of said preset transaction indicators; the main characteristic interval of the m-th preset transaction index; Is the main characteristic interval A total number of all transaction data points; the main characteristic interval of the nth preset transaction index; Is the main characteristic interval A total number of all transaction data points; a number of said remaining duplicate data points for an mth said preset transaction indicator and an nth said preset transaction indicator; the remaining characteristic interval of the m-th preset transaction index; For the remaining characteristic interval A total number of all transaction data points; The remaining characteristic interval of the n-th preset transaction index; For the remaining characteristic interval A total number of all transaction data points; The m-th distinguishing characteristic value of the preset transaction index; the distinguishing characteristic value of the nth preset transaction index; is a normalization function. It should be noted that, the covariance coefficient in the PAC algorithm is calculated for every two different dimensions, and the relevant eigenvalue is adjusted for the covariance coefficient, and is also calculated for the dimensions of two different preset transaction indexes.
In the method, in the process of the invention,The larger the duty ratio of the main data overlapping part reflecting the two preset transaction indexes, the stronger the correlation representing the two preset transaction indexes, and the larger the correlation characteristic value; the ratio of the overlapping part of the residual data reflecting the two preset transaction indexes is larger, the stronger the correlation of the two preset transaction indexes is represented, and the larger the correlation characteristic value is; the distinguishing characteristic value can represent the data density characteristic and the distribution characteristic of the preset transaction index, The smaller the difference between the data density characteristic and the distribution characteristic representing the two preset trade indexes is, the stronger the correlation between the two preset trade indexes is, the larger the correlation characteristic value is, and the stronger the correlation characteristic value comprehensively reflects the correlation between the two preset trade indexes.
Step S3: obtaining covariance coefficients of two preset transaction indexes; according to the distinguishing characteristic values and the related characteristic values of the two preset transaction indexes, the covariance coefficients of the two preset transaction indexes are adjusted, and adjusted covariance coefficients of the two preset transaction indexes are obtained; and acquiring an abnormality detection result of the transaction data to be analyzed according to the adjusted covariance coefficients of every two preset transaction indexes.
The covariance coefficient is adjusted based on the importance and the correlation of the preset transaction indexes, so that the adjusted covariance coefficient can reflect the inherent importance and the correlation of the preset transaction indexes, the problem of the difference of the data sizes of different preset transaction indexes can be more effectively processed, and the accuracy of the abnormal detection result of the transaction data to be analyzed is ensured.
Specifically, based on a PCA algorithm, obtaining covariance coefficients of two preset transaction indexes; it should be noted that the PCA algorithm is a technical means well known to those skilled in the art, and only brief steps for obtaining covariance coefficients of two preset transaction indexes according to the transaction data to be analyzed are described herein:
Because the transaction data to be analyzed is multidimensional data comprising the dimensions of all the preset transaction indexes, the average value of each preset transaction index is calculated according to the dimensions of each preset transaction index, then the decentralization processing is carried out, and the covariance between the two preset transaction indexes is calculated, so that the covariance coefficient of the two preset transaction indexes is obtained.
Preferably, in order to obtain a covariance coefficient that more accurately reflects the intrinsic importance and relevance of the preset transaction index, in one embodiment of the present invention, the method for obtaining the adjusted covariance coefficient includes:
acquiring adjusted covariance coefficients of the two preset trade indexes according to distinguishing characteristic values of the two preset trade indexes, correlation characteristic values of the two preset trade indexes and covariance coefficients of the two preset trade indexes; the distinguishing characteristic value, the correlation characteristic value, the covariance coefficient and the adjusted covariance coefficient are in positive correlation.
The positive correlation relationship indicates that the dependent variable increases with the increase of the independent variable, and the dependent variable decreases with the decrease of the independent variable, and the specific relationship may be a multiplication relationship, an addition relationship, an idempotent of an exponential function, and is determined by practical application.
Acquiring the adjusted covariance coefficient according to an adjusted covariance coefficient formula, wherein the acquiring formula of the adjusted covariance coefficient comprises the following steps:
; wherein, The adjusted covariance coefficients of the mth preset trade index and the nth preset trade index are obtained; The related characteristic values of the m-th preset transaction index and the n-th preset transaction index are obtained; The m-th distinguishing characteristic value of the preset transaction index; the distinguishing characteristic value of the nth preset transaction index; And (5) the covariance coefficients of the m-th preset trading index and the n-th preset trading index.
In the method, in the process of the invention,Reflecting the importance of the mth preset trading index and the nth preset trading index, amplifying the corresponding covariance coefficient when the preset trading index is more important, and adjusting the covariance coefficient through the importance of the preset trading index can further ensure that the important preset trading index is properly weighted in subsequent analysis.Reflecting the correlation between the mth and nth preset trade indexes, and when the correlation between the two preset trade indexes is strong, amplifying the covariance coefficient between them to reflect the strong correlation. By increasing the values of the corresponding elements of the two preset transaction indexes in the covariance coefficient, the correlation between the two preset transaction indexes in PCA analysis can be ensured to be fully considered, the adjusted covariance coefficient can reflect the inherent importance and the correlation of the preset transaction indexes, and the accuracy of the abnormal detection result of the transaction data to be analyzed is ensured.
In other embodiments of the present invention, the adjusted covariance coefficient is obtained according to an adjusted covariance coefficient formula, and the adjusted covariance coefficient obtaining formula includes:
; wherein, The adjusted covariance coefficients of the mth preset trade index and the nth preset trade index are obtained; The related characteristic values of the m-th preset transaction index and the n-th preset transaction index are obtained; The m-th distinguishing characteristic value of the preset transaction index; the distinguishing characteristic value of the nth preset transaction index; And (5) the covariance coefficients of the m-th preset trading index and the n-th preset trading index.
It should be noted that the PCA algorithm is a technical means well known to those skilled in the art, and only a brief method for obtaining the abnormal score of each transaction data point in the transaction data to be analyzed based on the PCA algorithm according to the adjusted covariance coefficients of each two preset transaction indexes is briefly described herein:
The adjusted covariance coefficients of every two preset transaction indexes constructed by the invention replace the original covariance coefficients in the PCA algorithm, and the adjusted covariance coefficients are utilized to decompose the eigenvalues to obtain the main components and the eigenvalues corresponding to the main components. The principal component represents the main direction of change in the data. The first few most important principal components are selected according to the magnitude of the eigenvalues. The original transaction data is projected into a new space made up of the selected principal components. This is typically accomplished by calculating the dot product of the original transaction data with each principal component, resulting in the coordinates of each transaction data point in the new space. In the new space, an anomaly score is calculated from the distribution characteristics of the transaction data points.
Further, when the anomaly score is smaller than the first set parameter, determining the transaction data point as safety data; when the abnormality score is not smaller than the first setting parameter and smaller than the second setting parameter, determining that the transaction data point is low risk data; when the abnormality score is not less than the second setting parameter and is less than the third setting parameter, determining that the transaction data point is the risk data; and when the abnormality score is not smaller than the third setting parameter, judging the transaction data point as high risk data. In this embodiment, the value of the first setting parameter is set to 0.3, the value of the second setting parameter is set to 0.7, and the value of the third setting parameter is set to 0.9, so that the implementer can set according to the implementation scenario. And labeling each transaction data point, and acquiring an abnormality detection result. For risky data it is necessary to check the source and corresponding destination of the data.
In summary, the embodiment of the invention provides a POS machine data intelligent processing method based on a 5G technology, which comprises the steps of firstly, obtaining distinguishing characteristic values of a preset transaction index according to density distribution conditions of all transaction data points under the dimension of the preset transaction index and the centralized distribution characteristic values of the preset transaction index; acquiring related characteristic values of two preset transaction indexes according to the distribution condition of all the transaction data points in the dimensions of the two preset transaction indexes and the difference of the distinguishing characteristic values between the two preset transaction indexes; and further, adjusting covariance coefficients of the two preset transaction indexes, and acquiring an abnormal detection result of the transaction data to be analyzed according to the adjusted covariance coefficients of each two preset transaction indexes. According to the embodiment of the invention, the covariance coefficient is adjusted based on the importance and the correlation of the preset transaction index, so that the adjusted covariance coefficient can reflect the inherent importance and the correlation of the preset transaction index, and the accuracy of the abnormal detection result of the transaction data to be analyzed is ensured.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (10)
1. The POS machine data intelligent processing method based on the 5G technology is characterized by comprising the following steps of:
Acquiring transaction data to be analyzed; the transaction data to be analyzed comprises all transaction data points; each transaction data point corresponds to the dimension of all preset transaction indexes;
In transaction data to be analyzed, acquiring a centralized distribution characteristic value of a preset transaction index according to the main characteristic distribution condition of all transaction data points in the dimension of the preset transaction index; acquiring a distinguishing characteristic value of a preset transaction index according to the density distribution condition of all transaction data points in the dimension of the preset transaction index and the centralized distribution characteristic value of the preset transaction index; acquiring related characteristic values of two preset transaction indexes according to the distribution condition of all the transaction data points in the dimensions of the two preset transaction indexes and the difference of the distinguishing characteristic values between the two preset transaction indexes;
Obtaining covariance coefficients of two preset transaction indexes; according to the distinguishing characteristic values and the related characteristic values of the two preset transaction indexes, the covariance coefficients of the two preset transaction indexes are adjusted, and adjusted covariance coefficients of the two preset transaction indexes are obtained; and acquiring an abnormality detection result of the transaction data to be analyzed according to the adjusted covariance coefficients of every two preset transaction indexes.
2. The method for intelligently processing POS machine data based on 5G technology according to claim 1, wherein the method for obtaining the centralized distribution characteristic value comprises the following steps:
Under the dimension of a preset trade index, taking the section of the preset trade index where all trade data points with the value of the preset trade index being larger than a preset first quantile threshold and smaller than a preset second quantile threshold are located as the main characteristic section of the preset trade index; taking the interval of all transaction data points in the preset transaction index as a reference characteristic interval of the preset transaction index;
calculating the ratio of the length of the main characteristic interval to the length of the reference characteristic interval, and carrying out negative correlation normalization on the comparison value to obtain the centralized distribution characteristic value of the preset transaction index.
3. The method for intelligently processing POS machine data based on 5G technology according to claim 2, wherein the method for obtaining the centralized distribution characteristic value comprises the following steps:
obtaining the centralized distribution characteristic value according to a centralized distribution characteristic value formula, wherein the centralized distribution characteristic value formula comprises:
; wherein, The concentrated distribution characteristic value of the a-th preset transaction index; The main characteristic interval of the a-th preset transaction index; the reference characteristic interval is the a-th preset transaction index; The length of the main characteristic interval of the a-th preset transaction index; the length of the reference characteristic interval of the a-th preset transaction index; is an exponential function based on e.
4. The method for intelligently processing POS machine data based on the 5G technology according to claim 2, wherein the method for obtaining the distinguishing characteristic value comprises the following steps:
in the reference characteristic interval of the preset transaction index, an interval except the main characteristic interval is used as the remaining characteristic interval of the preset transaction index;
Taking the main characteristic interval and the residual characteristic interval as density characteristic intervals to be analyzed of the preset transaction indexes; obtaining a density maximum value, a density minimum value and a mean value density of the density characteristic interval to be analyzed;
Acquiring an attribute density vector of the density characteristic interval to be analyzed according to the density maximum value, the density minimum value and the average value density of the density characteristic interval to be analyzed;
And under the dimension of a preset transaction index, according to the difference between the attribute density vector of a main characteristic interval and the attribute density vector of the residual characteristic interval, the centralized distribution characteristic value of the preset transaction index is obtained, and the difference between the average value density of the main characteristic interval and the average value density of the residual characteristic interval is obtained.
5. The method for intelligently processing POS machine data based on the 5G technology according to claim 4, wherein the method for obtaining the distinguishing characteristic value comprises the steps of:
acquiring the distinguishing characteristic value according to a distinguishing characteristic value formula, wherein the acquiring formula of the distinguishing characteristic value comprises the following steps:
; wherein, A distinguishing characteristic value of the preset transaction index of the a-th type; the concentrated distribution characteristic value of the a-th preset transaction index; The main characteristic interval of the a-th preset transaction index; the remaining characteristic interval of the a-th preset transaction index; Is the main characteristic interval Is a vector of attribute densities;
For the remaining characteristic interval Is a vector of attribute densities;
Is the main characteristic interval Is a mean density of (1); For the remaining characteristic interval Is a mean density of (1); Is a modulus of the vector; Is an absolute value symbol; Is a normalization function.
6. The intelligent processing method for POS machine data based on 5G technology of claim 4, wherein the method for obtaining the attribute density vector comprises:
acquiring the attribute density vector according to an attribute density vector formula, wherein the acquisition formula of the attribute density vector comprises the following steps:
; wherein, For the density characteristic interval to be analyzedIs a vector of attribute densities; A density characteristic interval to be analyzed which is the a-th preset transaction index;
For the density characteristic interval to be analyzed Is a maximum value of the density of the material; For the density characteristic interval to be analyzed Is a minimum of the density of (1); For the density characteristic interval to be analyzed Is a mean value density of (1).
7. The intelligent processing method for POS machine data based on 5G technology of claim 4, wherein the method for obtaining the relevant characteristic value comprises:
under the dimension of two preset trade indexes, taking the trade data points which are simultaneously positioned in the main characteristic intervals of the two preset trade indexes as main repeated data points of the two preset trade indexes; taking transaction data points which are simultaneously positioned in the residual characteristic intervals of the two preset transaction indexes as residual repeated data points of the two preset transaction indexes;
acquiring the relevant characteristic value according to a relevant characteristic value formula, wherein the acquiring formula of the relevant characteristic value comprises the following steps:
Wherein, The related characteristic values of the m-th preset transaction index and the n-th preset transaction index are obtained; A number of primary repeat data points for an mth of said preset transaction indicators and an nth of said preset transaction indicators; the main characteristic interval of the m-th preset transaction index; Is the main characteristic interval A total number of all transaction data points; the main characteristic interval of the nth preset transaction index; Is the main characteristic interval A total number of all transaction data points; a number of said remaining duplicate data points for an mth said preset transaction indicator and an nth said preset transaction indicator; the remaining characteristic interval of the m-th preset transaction index; For the remaining characteristic interval A total number of all transaction data points; The remaining characteristic interval of the n-th preset transaction index; For the remaining characteristic interval A total number of all transaction data points; The m-th distinguishing characteristic value of the preset transaction index; the distinguishing characteristic value of the nth preset transaction index; Is a normalization function.
8. The 5G technology-based POS machine data intelligent processing method according to claim 1, wherein the adjusted covariance coefficient obtaining method includes:
Acquiring adjusted covariance coefficients of the two preset trade indexes according to the distinguishing characteristic values of the two preset trade indexes, the correlation characteristic values of the two preset trade indexes and the covariance coefficients of the two preset trade indexes; the distinguishing characteristic value, the correlation characteristic value, the covariance coefficient and the adjusted covariance coefficient are in positive correlation.
9. The method for intelligently processing POS machine data based on 5G technology according to claim 1, wherein the method for obtaining covariance coefficients of two preset transaction indexes comprises:
based on PCA algorithm, covariance coefficients of two preset transaction indexes are obtained.
10. The intelligent processing method for POS machine data based on 5G technology of claim 1, wherein the method for obtaining the anomaly detection result comprises:
Based on a PCA algorithm, acquiring an abnormal score of each transaction data point in the transaction data to be analyzed according to the adjusted covariance coefficients of each two preset transaction indexes; and acquiring an abnormality detection result of the transaction data to be analyzed according to the abnormality score of each transaction data point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410650488.6A CN118245734B (en) | 2024-05-24 | 2024-05-24 | POS machine data intelligent processing method based on 5G technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410650488.6A CN118245734B (en) | 2024-05-24 | 2024-05-24 | POS machine data intelligent processing method based on 5G technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118245734A CN118245734A (en) | 2024-06-25 |
| CN118245734B true CN118245734B (en) | 2024-08-09 |
Family
ID=91556741
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410650488.6A Active CN118245734B (en) | 2024-05-24 | 2024-05-24 | POS machine data intelligent processing method based on 5G technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118245734B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114022284A (en) * | 2021-11-10 | 2022-02-08 | 中国工商银行股份有限公司 | Abnormal transaction detection method and device, electronic equipment and storage medium |
| CN115238815A (en) * | 2022-08-10 | 2022-10-25 | 中国工商银行股份有限公司 | Abnormal transaction data acquisition method, device, equipment, medium and program product |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR102433598B1 (en) * | 2020-11-26 | 2022-08-18 | (주)심플랫폼 | A System and Method for Deriving Data Boundary |
| CN112632469A (en) * | 2020-12-22 | 2021-04-09 | 深圳壹账通智能科技有限公司 | Method and device for detecting abnormity of business transaction data and computer equipment |
| CN113961434A (en) * | 2021-09-29 | 2022-01-21 | 西安交通大学 | Method and system for monitoring abnormal behaviors of distributed block chain system users |
| CN117725451B (en) * | 2023-12-15 | 2024-06-21 | 北京微保科技有限责任公司 | Automatic checking method and system for multidimensional transaction data |
-
2024
- 2024-05-24 CN CN202410650488.6A patent/CN118245734B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114022284A (en) * | 2021-11-10 | 2022-02-08 | 中国工商银行股份有限公司 | Abnormal transaction detection method and device, electronic equipment and storage medium |
| CN115238815A (en) * | 2022-08-10 | 2022-10-25 | 中国工商银行股份有限公司 | Abnormal transaction data acquisition method, device, equipment, medium and program product |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118245734A (en) | 2024-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8639618B2 (en) | System and method of detecting and assessing multiple types of risks related to mortgage lending | |
| Sullivan et al. | Change-point detection of mean vector or covariance matrix shifts using multivariate individual observations | |
| Altman | Revisiting credit scoring models in a Basel 2 environment | |
| US6182070B1 (en) | System and method for discovering predictive association rules | |
| US20070203939A1 (en) | Alert Flags for Data Cleaning and Data Analysis | |
| US20030097320A1 (en) | Method and system of data analysis for the detection of fraudulent financial transactions | |
| CN110717824A (en) | Method and device for conducting and calculating risk of public and guest groups by bank based on knowledge graph | |
| CN113852603A (en) | Method and device for detecting abnormality of network traffic, electronic equipment and readable medium | |
| US20220101069A1 (en) | Machine learning outlier detection using weighted histogram-based outlier scoring (w-hbos) | |
| CN111612038A (en) | Abnormal user detection method and device, storage medium and electronic equipment | |
| US11574211B2 (en) | Log analysis system, log analysis method, log analysis program, and storage medium | |
| CN111340086A (en) | Method, system, medium and terminal for processing label-free data | |
| Buonaguidi et al. | Bayesian quickest detection of credit card fraud | |
| JP2023183187A (en) | Program, and information processing device | |
| CN118245734B (en) | POS machine data intelligent processing method based on 5G technology | |
| Yli-Olli et al. | Transformation analysis applied to long-term stability and structural invariance of financial ratio patterns: US vs. Finnish firms | |
| US11935054B2 (en) | Systems and methods for automatically generating fraud strategies | |
| Parpoula et al. | On change-point analysis-based distribution-free control charts with Phase I applications | |
| CN115456679A (en) | Analysis method, device and medium for network retail shop index system | |
| CN116308370A (en) | Training method of abnormal transaction recognition model, abnormal transaction recognition method and device | |
| US20240152133A1 (en) | Threshold acquisition apparatus, method and program for the same | |
| CN109976934A (en) | Equipment uses abnormality determination method, device and computer storage medium | |
| US12020256B2 (en) | Generation of divergence distributions for automated data analysis | |
| WO2022249266A1 (en) | Fraud detection system, fraud detection method, and program recording medium | |
| KR102637818B1 (en) | Server for predicting of borrowed-name account and method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |