Detailed Description
The embodiment of the specification provides a data derivation method, a data derivation device and data derivation equipment.
In order to make those skilled in the art better understand the technical solutions in the present specification, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only a part of the embodiments of the present specification, and not all of the embodiments. All other embodiments obtained by a person skilled in the art based on the embodiments in the present specification without any inventive step should fall within the scope of protection of the present specification.
Example one
As shown in fig. 1, an execution subject of the method may be a terminal device or a server, where the terminal device may be a device such as a personal computer, or may also be a mobile terminal device such as a mobile phone and a tablet computer, and the terminal device may be a terminal device used by a user. The server may be an independent server, or a server cluster composed of a plurality of servers, and the server may be a background server of a certain service, or a background server of a certain website (such as an online shopping website or a payment application), and the like. The method can be used for deriving more related data based on basic data and the like, and in practical application, the method can be applied to various scenes, for example, in a wind control system, in the process of creating fraud prevention and control rules, more training data can be acquired through the embodiment of the specification, or the method can also be applied to risk detection and the like.
In order to improve the data processing efficiency, in this embodiment, the execution main body is taken as an example for description, and for the case of the terminal device, processing may be performed according to the following related contents, which is not described herein again. The method may specifically comprise the steps of:
in step S102, gridding processing is performed on the target data to generate one or more grids covering the target data.
The target data can be any data, such as data related to network transaction performed in the process of network shopping; the method comprises the following steps that related data of mutual transfer among different users are carried out through a network payment application program; the target data may include a plurality of pieces of data, which may be collected within a predetermined period of time, for example, data generated within 30 days before the current time of a shopping website, which may include data related to network transactions generated within 30 days before the current time of the shopping website, and data related to user login or registration of the shopping website. The gridding process may be a process of dividing the target data into a plurality of parts to form a grid, and the gridding process may be implemented by a plurality of dividing methods, such as average division, or division according to a predetermined division rule, specifically, division according to a geographic area (such as a city or a set geographic range) generated by the data, or division according to different users or related information (such as age or gender) of the user, which is not limited in the embodiments of the present specification. The mesh may be obtained by dividing the target data into a plurality of divided parts, each of which may be a mesh, and the mesh may be formed by one or more pieces of target data.
In implementation, the target data may be obtained through a variety of ways, for example, taking obtaining data of a certain shopping website within a certain time period as the target data, and recording all operations and behaviors performed by a user in the shopping website under the condition that the user agrees, where the operations and behaviors may include related data in the processes of registering the shopping website, logging in the shopping website, selecting a commodity from the shopping website and placing an order, paying, and the like, and the recorded related data of the operations or behaviors may be determined as the target data, or the server may purchase required related data from the user, or guide the user to enter required related data in a reward manner, and the server may use the obtained data as the target data.
The gridding processing of the data is an efficient management mode and a processing mode for the data, and grid variables have been widely applied To various domains and have a very good effect, for example, in a risk prevention and control system, the grid variables can be applied To real-time prevention and control of fraudulent use (for example, a frag To Gross Fraud (FTG) variable which describes a risk form in a certain grid in the past period is called To perform strong and weak control on the risk Fraud), and can also be used for risk detection (for example, a risk trend is sensed by monitoring the transaction of certain index data in the certain grid so as To quickly and timely make a risk response), and the like.
However, in different scenarios, the grid dimension and granularity in the grid processing are often divided differently, which is a difficulty in the grid variable construction process, for example, for a wallet system mainly based on transactions between accounts, a seller account may be a main grid dimension, and in an e-commerce system mainly based on card transactions, a grid variable described by the grid dimension of a seller (or a seller account) is less differentiated. Meanwhile, in the current practical application, there are often problems that the grid generation is cumbersome and selective, the grid generation process often needs a technician to manually input SQL (Structured Query Language) codes, and a large amount of repetitive work is often accompanied in the process of cross dimension (i.e. 2-dimensional) and extended dimension (3-dimensional or higher), so that the current grid generation often uses empirical risk types that have been deposited on corresponding services, and for unknown or existing potential risks, and currently unknown dimensions are often ignored without being analyzed. Therefore, the embodiment of the present specification provides an implementation scheme, which may specifically include the following:
after the target data is obtained in the above manner, the target data may be subjected to gridding processing, for example, the target data may be averagely divided into a plurality of portions according to the total data amount of the target data, the target data of each portion may form a grid, specifically, the total data amount of the target data is 2GB, the target data may be averagely divided into 100 portions, the total data amount of each portion is 20.48MB, that is, each grid includes 20.48MB of data, and the like. Alternatively, the target data may be randomly divided to obtain a plurality of corresponding grids. It should be noted that the gridding processing on the target data may be controlled and implemented by a preset code script, or may be implemented by other modes that can be automatically executed, which is not limited in the embodiment of the present specification. In addition, in one or more grids obtained by the gridding processing, the data in all the grids are combined together to form the target data.
In step S104, dimension information is generated from information contained in each mesh, the dimension information including single-dimension information and multi-dimension information.
The single-dimensional information may be information of a certain aspect contained in the data, such as age, gender, location of the user, and dimension information of time, location and the like in the transaction, specifically, information of transaction time, payment method, commodity category of the transaction may be generated in the process that the user performs online shopping and performs the transaction with the merchant, and any one of the information of the transaction time, the payment method, and the commodity category of the transaction may form the single-dimensional information, i.e., the time dimension, the payment method dimension, the category dimension and the like. The multidimensional information may be information included in the data, in which two or more aspects are aggregated, for example, the single-dimensional information includes a time dimension, a payment method dimension, and a category dimension, and then the corresponding multidimensional information may include a time-payment method dimension, a payment method-category dimension, a time-category dimension, and a time-payment-category dimension, and the like.
In implementation, after the target data is subjected to gridding processing, each obtained grid includes at least one piece of data, each piece of data includes information related to operation and behavior of a user in a corresponding website or a certain service, each piece of data in the grid can be analyzed, information in a plurality of different aspects, such as time aspect or category aspect, can be extracted or split from each piece of data, dimension information corresponding to each aspect can be determined based on features included in the information in each aspect, and the dimension information can be generated. For example, a certain grid includes a transaction event, the transaction information included in the transaction event may include a BIN Identification Number (e.g., the first 6 digits of the card Number of the Bank card), a category, and transaction time, and if the dimension information to be generated includes single-dimensional information and two-dimensional information, 6 kinds of dimension information, including a BIN code, a category, a time period, a BIN code-time period, a category-time period, and a BIN code-category, may be generated, where the BIN code, the category, and the time period are three kinds of single-dimensional information, and the BIN code-time period, the category-time period, and the BIN code-category are three kinds of two-dimensional information.
It should be noted that, when generating the dimension information, the embodiments of the present specification are not limited to only the single-dimension information that can be directly extracted from the grid information, and also need to further derive one or more pieces of two-dimension information, three-dimension information, or even higher dimension information based on the single-dimension information, so that the dimension information is more extensive, and a basis is provided for a user to solve corresponding problems from more dimensions.
In step S106, a basic index corresponding to the target data is determined according to the dimension information, and the time period and the event type included in each grid.
The time period included in the grid may also be referred to as a time window of the grid, and the time period may be a time period from the earliest time to the latest time included in the grid, for example, if the data included in the grid is data from 30 days before the current time, the time period is 30 days before the current time. The event type may include multiple types, and may be specifically set according to an actual situation, for example, a login type, a registration type, a transaction type, a transfer type, and the like, which is not limited in this embodiment of the specification. The basic index may be a type of information summarized by a calculation process such as direct or simple statistical analysis, such as failure rate or case rate.
It should be noted that, in different scenarios, there is a certain difference in the selection of indexes corresponding to data, for example, in domestic transactions in the country, case rate needs to be selected as an index, and in international transactions, due to the influence of case report back period and other reasons, case rate is selected as an index, and thus actual requirements cannot be met, and in this case, rejection rate or failure rate is usually selected as an index. Also, the relevant variables of the current grid (i.e., grid variables) mainly take the form of proportional variables, such as the case rate in a predetermined period of time before the current time as mentioned above, or the usage rate of new users, etc.
In the implementation, the single-dimensional information and the multi-dimensional information may be generated through the processing of the step S104, and the single-dimensional information and the multi-dimensional information may be respectively used to perform statistical analysis by combining the time window of the grid where the corresponding single-dimensional information or the multi-dimensional information is located and the event type corresponding to the corresponding data in the grid, to determine the corresponding index, and may use the index as the basic index, and may accumulate each basic index, for example, the amount of money, the amount of cases, the amount of audit events, the amount of events may be used as the basic index through statistical analysis, and the basic indexes such as the amount of money, the amount of cases, the amount of audit events, and the amount of events may be accumulated.
For example, based on the example of step S104, the single-dimensional information includes BIN codes, categories, and time periods, the two-dimensional information includes BIN code-time periods, category-time periods, and BIN code-categories, the single-dimensional information may be the BIN codes, the time windows corresponding to the grids including the BIN codes may be combined, and the event types included in the grids including the BIN codes may be comprehensively analyzed, so that basic indexes such as money amount, case rate, event amount, and the like may be determined, and a basic index pool may be formed.
In step S108, the basic index is subjected to derivation processing by an index derivation algorithm, so as to obtain a derived index corresponding to the target data.
The index derivation algorithm may be a rule or algorithm that is based on a basic index and develops a new index different from the basic index by processing the basic index in a certain specified manner, and the index derivation algorithm may include multiple types, and a rule or algorithm that can develop the basic index to obtain the new index in actual application may be used as the index derivation algorithm, specifically, a genetic algorithm, and the like. The derived indicator may be a new indicator based on the evolution of the base indicator.
In the implementation, based on the foregoing related content, the related variables (i.e., grid variables) of the current grid mainly take the form of proportional variables, such as the case rate in the predetermined time period before the current time mentioned above, or the usage rate of the new user, and the like, and the above proportional variables only depend on the experience of the technician, so that the limitation is large, in order to expand the grid variables and enable more choices in the subsequent processing, a new index can be derived based on the basic index, specifically, an index derivation algorithm to be used can be set in the server in advance, in the practical application, the index derivation algorithm can provide multiple choices, different index derivation algorithms can be selected according to the practical requirements, for example, in order to accelerate the derivation processing efficiency, an index derivation algorithm with relatively simplified processing logic can be selected, and if necessary, a better derivation effect can be obtained, and a fine index derivation algorithm related to the processing process can be selected.
When the basic index corresponding to the target data is obtained through the processing in the step S106, the server may invoke an originally set or selected index derivation algorithm, perform derivation processing on the basic index through a processing logic process of the index derivation algorithm to obtain a corresponding new index, may evaluate the derived new index, determine whether the new index meets the requirement, whether to continue to derive the new index, and the like, may perform elimination processing on the new index that does not meet the requirement, and if it is necessary to continue to derive the new index, may continue to derive the new index through the processing logic until the derived new index passes the above evaluation, and finally, may use the new index that passes the evaluation as the derived index corresponding to the target data.
After the derivative index corresponding to the target data is obtained, the derivative index and the basic index can be combined, so that the index corresponding to the target data can be obtained, and the index corresponding to the target data is expanded to a certain extent. After the obtained index corresponding to the target data is obtained, further subsequent related processing can be performed based on the index, for example, corresponding features can be determined based on the index, risk prevention and control rules can be created or updated based on the determined features, and the like.
The embodiment of the specification provides a data derivation method, which includes generating one or more grids covering target data by performing grid processing on the target data, then generating single-dimensional information and multi-dimensional information according to the information contained in each grid, determining a basic index corresponding to the target data according to the single-dimensional information and the multi-dimensional information, and time periods and event types contained in each grid, and finally performing derivation processing on the basic index through an index derivation algorithm to obtain a derived index corresponding to the target data, so that the grids are automatically divided through discrete target data, the obtained grids are listed, meanwhile, the single-dimensional information and the multi-dimensional information are generated according to the grids to further construct the basic index, and the basic index is determined not only through the single-dimensional information but also through the multi-dimensional information, therefore, the content contained in the basic index is wider, a data base is provided for subsequent processing, in addition, the aggregation of elements in the basic index is realized by means of an index derivative algorithm, and then the derivative index corresponding to the target data is generated, so that the index corresponding to the target data can be expanded to a certain extent, and further the characteristic corresponding to the index can be more precise and comprehensive, so that the accuracy of the subsequent processing is improved, for example, the obtained risk prevention and control rule can be more accurate, and the like.
Example two
As shown in fig. 2, an execution subject of the method may be a terminal device or a server, where the terminal device may be a device such as a personal computer, or may also be a mobile terminal device such as a mobile phone and a tablet computer, and the terminal device may be a terminal device used by a user. The server may be an independent server, or a server cluster composed of a plurality of servers, and the server may be a background server of a certain service, or a background server of a certain website (such as an online shopping website or a payment application), and the like. The method can be used for deriving more related data based on basic data and the like, and in practical application, the method can be applied to various scenes, for example, in a wind control system, in the process of creating fraud prevention and control rules, more training data or characteristics can be acquired through the embodiment of the specification, or the method can also be applied to risk detection and the like.
In order to improve the data processing efficiency, in this embodiment, the execution main body is taken as an example for description, and for the case of the terminal device, processing may be performed according to the following related contents, which is not described herein again. The method may specifically comprise the steps of:
in step S202, the target data is subjected to gridding processing in a randomly divided manner, and one or more grids covering the target data are generated.
In implementation, the target data may be obtained through multiple ways, where the target data may be historical data of a certain service, and for details, reference may be made to relevant content of step S102 in the foregoing embodiment, and details are not described here again. After the target data is acquired, a code script preset in the server can be executed, so that the server automatically performs gridding processing on the target data in a random division mode to divide the target data, each divided part of data can form a grid, one or more grids covering the target data can be obtained in the mode, and each grid can include part of data in the target data.
In step S204, dimension information is generated from information contained in each mesh, and the dimension information includes single-dimension information and multi-dimension information.
In implementation, the specific processing procedure of the step S204 may refer to the related content of the step S104 in the first embodiment, and in practical application, for some dimension information, the related information including the predetermined field in the grid may be subjected to predetermined processing, so that the general content or the accurate content and the like corresponding to the dimension information may be determined, for example, for a time period dimension included in the dimension information, because for a transaction event, the information of the transaction event often includes transaction time, the time field commonly used in the event such as the transaction event may be subjected to predetermined processing, so as to reasonably divide the time period of the transaction in the grid and determine the time period meeting the requirement or the general purpose. For example, the time of 0 o 'clock to 6 o' clock in the united states is the low peak period of the transaction, but the time period is a high-risk period when the account number is stolen, and at this time, the time period corresponding to the country is 12 o 'clock to 18 o' clock. In this case, the time field may be determined by being associated with the corresponding data set, and 12 o 'clock to 18 o' clock may be divided into the same interval. Meanwhile, in the process of determining the grids, the dimension information can be chosen or rejected according to the event magnitude covered by the grids, so that the dimension information which is not suitable or not in accordance with the requirements can be filtered, and only the dimension information which is suitable or in accordance with the requirements is reserved.
In step S206, a basic index corresponding to the target data is determined according to the dimension information, and the time period and the event type included in each grid.
The event type may include a payment class and an operation class, wherein the event of the payment class may include a transaction event, a transfer event, a code scanning event, and the like, and the event of the operation class may include a login event, a password modification event, a registration event, a browsing event, and the like.
The content of the step S206 is the same as the content of the step S106 in the first embodiment, and the specific processing procedure of the step S206 may refer to the related content of the step S106 in the first embodiment, which is not described herein again.
In step S208, it is determined whether the basic indicator satisfies a first selection condition corresponding to the indicator derivation algorithm.
The index derivation algorithm may include a genetic algorithm, a random walk algorithm, and a violence derivation algorithm, among others. The first selection condition may include multiple conditions, for example, the iteration number condition and/or the data amount satisfy a predetermined data amount threshold, and may be specifically determined according to a selected index derivation algorithm, which is not limited in the embodiment of the present specification.
In implementation, in order to expand grid variables and enable more choices in subsequent processing, a new index may be derived based on a basic index, and specifically, an index derivation algorithm that needs to be used may be set in the server in advance. Since a myriad of new indicators can be derived by the indicator derivation algorithm, in practical applications, only a certain number of new indicators may be derived, and therefore, the indicator derivation algorithm may include a selection condition (i.e., a first selection condition) for selecting a certain number of new indicators. The basic index may be first matched with the first selection condition, and if the basic index satisfies the first selection condition, it may be indicated that the basic index does not need to be subjected to derivation processing, and subsequent processing may be performed directly through the basic index. If the basic index does not satisfy the first selection condition, it may indicate that the current basic index cannot satisfy the subsequent processing requirement and needs to be subjected to the derivation process, and at this time, the following processing of step S210 may be performed.
In step S210, if not, performing at least one derivation process on the basic index, and determining whether the index obtained by the derivation process satisfies the first selection condition after performing the derivation process each time until it is determined that the index obtained by the derivation process satisfies the first selection condition or the number of times of performing the derivation process reaches the predetermined number threshold.
The predetermined number threshold may be set according to an actual situation, specifically, 100 times or 200 times, and the like, which is not limited in this specification.
In the implementation, for example, taking the index derivation algorithm as the random walk algorithm as an example, the basic index may be used as an initial iteration point, the step length of each iteration and the numerical value of the control precision (which may be a very small positive number for controlling the end algorithm) may be set, and the iteration number (which may be a first selection condition) is set.
The embodiment of the present specification provides only one optional implementation manner, and specifically, the number of derivative processes and the like may be preset as a first selection condition, and then, any two indexes included in the basic index may be superimposed or correspondingly calculated to obtain a new index, or any three or more indexes in the basic index may be superimposed or correspondingly calculated to obtain a new index until the set number of derivative processes is reached.
In step S212, the index obtained by the last derivation process is determined as the derivation index corresponding to the target data.
It should be noted that, if the basic index meets the first selection condition, it can be shown that the basic index does not need to be derived, and the subsequent processing can be performed directly through the basic index.
The embodiment of the specification provides a data derivation method, which includes generating one or more grids covering target data by performing grid processing on the target data, then generating single-dimensional information and multi-dimensional information according to the information contained in each grid, determining a basic index corresponding to the target data according to the single-dimensional information and the multi-dimensional information, and time periods and event types contained in each grid, and finally performing derivation processing on the basic index through an index derivation algorithm to obtain a derived index corresponding to the target data, so that the grids are automatically divided through discrete target data, the obtained grids are listed, meanwhile, the single-dimensional information and the multi-dimensional information are generated according to the grids to further construct the basic index, and the basic index is determined not only through the single-dimensional information but also through the multi-dimensional information, therefore, the content contained in the basic index is wider, a data base is provided for subsequent processing, in addition, the aggregation of elements in the basic index is realized by means of an index derivative algorithm, and then the derivative index corresponding to the target data is generated, so that the index corresponding to the target data can be expanded to a certain extent, and further the characteristic corresponding to the index can be more precise and comprehensive, so that the accuracy of the subsequent processing is improved, for example, the obtained risk prevention and control rule can be more accurate, and the like.
EXAMPLE III
As shown in fig. 3, an execution subject of the method may be a terminal device or a server, where the terminal device may be a device such as a personal computer, or may also be a mobile terminal device such as a mobile phone and a tablet computer, and the terminal device may be a terminal device used by a user. The server may be an independent server, or a server cluster composed of a plurality of servers, and the server may be a background server of a certain service, or a background server of a certain website (such as an online shopping website or a payment application), and the like. The method can be used for deriving more related data based on basic data and the like, and in practical application, the method can be applied to various scenes, for example, in a wind control system, in the process of creating fraud prevention and control rules, more training data can be acquired through the embodiment of the specification, or the method can also be applied to risk detection and the like.
In order to improve the data processing efficiency, in this embodiment, the execution main body is taken as an example for description, and for the case of the terminal device, processing may be performed according to the following related contents, which is not described herein again. In the embodiment of the present specification, in order to select the optimized index (including the derived index) based on the basic index, a genetic algorithm may be used. The genetic algorithm is a calculation method for simulating the natural selection of a biological evolution theory and the biological evolution process of a genetic mechanism, and is a method for expanding a population and searching an optimal solution by simulating the natural evolution process. Genetic algorithms start with a population representing a possible potential solution set to the problem, and a population is composed of a certain number of individuals (i.e., base indices) that are encoded (or genetically encoded). Each individual is actually a chromosome-bearing entity. Chromosomes, which are the main carriers of genetic material, are a collection of multiple genes that determine the external appearance of an individual's shape. After the initial population (i.e. basic index) is generated, evolution can be carried out generation by generation according to the principle of survival and superiority and inferiority of fittest, in each generation, individuals are selected according to the fitness of individuals in a problem domain, combined intersection and/or variation are carried out by means of genetic operators of natural genetics, a population (i.e. derivative index) representing a new solution set is generated, the process can lead the subsequent generation population (i.e. derivative index) of the population as naturally evolved to be more adaptive to the environment than the previous generation population, and the final generation population (i.e. derivative index) can be used as the final derivative index after decoding.
The method may specifically comprise the steps of:
in step S302, the target data is subjected to gridding processing in a randomly divided manner, and one or more grids covering the target data are generated.
In step S304, dimension information is generated from information contained in each mesh, the dimension information including single-dimension information and multi-dimension information.
In step S306, a basic index corresponding to the target data is determined according to the dimension information, and the time period and the event type included in each grid.
The event type may include a payment class and an operation class, the event of the payment class may include a transaction event, a transfer event, a code scanning event, and the like, and the event of the operation class may include a login event, a password modification event, a registration event, a browsing event, and the like.
In step S308, the basic index is encoded to obtain basic data.
In practice, since genetic algorithms cannot directly deal with parameters of problem space, it is necessary to convert basic indexes into chromosomes or individuals of genetic space, which are composed of genes in a certain structure, wherein the conversion operation is coding and can also be called (problem) Representation (replication). Therefore, the basic index can be encoded to convert the basic index into basic data of a genetic space.
In step S310, fitness calculation is performed on the basic data to obtain first data.
In the implementation, the fitness in the genetic algorithm is considered to represent the adaptive capacity of a certain individual to the environment, and the fitness can be represented by a fitness function, which can also be referred to as an evaluation function, and the fitness function in the genetic algorithm can be used as an index for judging the degree of goodness or badness of the individual in the population, and is evaluated according to an objective function of the problem to be solved. The genetic algorithm does not need other external information in the search evolution process, and only uses the fitness to evaluate the quality of the individual and is used as the basis of subsequent operation. In the genetic algorithm, the fitness function needs to be compared and ranked, and the selection probability is calculated on the basis, so that the value of the fitness function can take a positive value.
Through the above related content, the fitness function may be selected in advance, and the selection of the fitness function may be implemented in various ways, for example, the fitness function is determined through target data, or the fitness function is set through experience, or a style of the fitness function may be specified through a service requirement, and the like, which is not limited in the embodiment of the present specification. After the fitness function is selected in the above manner, each of the basic indexes can be input into the set fitness function for calculation, and the obtained calculation result is the first data. Through the method, the first data of each index in the basic indexes can be obtained.
It should be noted that, in order to make the fitness (or fitness function) in the embodiment of the present disclosure have a more comprehensive Importance Indicator capability, the fitness in the embodiment of the present disclosure may be obtained by aggregating a plurality of predetermined indicators, where the plurality of predetermined indicators may be set according to actual conditions, and the embodiment of the present disclosure provides an optional combination of indicators, that is, the plurality of predetermined indicators may include at least two of an area auc (area Under Curve) Under a receptive ROC (Receiver Operating Characteristic Curve), an information quantity iv (information value), and an Embedding layer Embedding.
In step S312, it is determined whether the first data satisfies a first selection condition corresponding to a genetic algorithm.
In implementation, the server obtains first data corresponding to the basic data (i.e., the initial population) through the processing in step S308, and then compares the first data of each index with a preset evaluation threshold, and if the first data of a certain index is greater than or equal to the evaluation threshold, the index may be set as a passing index (or a passing label may be added to the index in the basic index, etc.), and if the first data of a certain index is less than the evaluation threshold, the index may be set as a discarding index (or a discarding label may be added to the index in the basic index). After the processing is completed, the server may count the number of indexes subjected to fitness calculation (or evaluation), obtain first data of each index, calculate an average value of the first data of the indexes subjected to the fitness calculation, compare the number of the indexes subjected to the fitness calculation with related information in a first selection condition, compare the obtained average value with related information in the first selection condition, and if the results of the two comparisons are both passed, determine that the first data satisfies the first selection condition, and if at least one of the results of the two comparisons is failed (that is, the number of the indexes subjected to the fitness calculation does not reach the first selection condition, and/or the obtained average value does not reach the first selection condition), determine that the first data does not satisfy the first selection condition.
If it is determined that the first data does not satisfy the first selection condition, it may indicate that the base index does not satisfy the first selection condition, and at this time, the following processes of step S314 to step S316 may be performed. If the first data is judged to meet the first selection condition, the basic index can be shown to meet the first selection condition, and at the moment, subsequent processing can be executed directly through the basic index without derivative processing.
In step S314, if the first data does not satisfy the first selection condition, the basic data is derived at least once, and after each derivation process, whether the data obtained by the derivation process satisfies the first selection condition is determined until it is determined that the data obtained by the derivation process satisfies the first selection condition or the number of derivation processes performed reaches the predetermined number threshold.
Wherein, the derivation process may include a crossover process and/or a mutation process. The cross processing may be a process of randomly exchanging partial contents between any two pieces of existing basic data to form new data (and then generating a new indicator), and the mutation processing may be a process of randomly transforming partial contents of any one piece of existing basic data to form new data (and then generating a new indicator).
In practice, if it is determined through the above processing procedure that the first data does not satisfy the first selection condition, the first data may be discarded, and meanwhile, according to the genetic algorithm, the information included in the data in the basic data corresponding to the first data may not be single information, and may include multiple information (corresponding to a combination of multiple genes), so that the data may be processed in a form of gene recombination (including gene crossing and gene mutation) with other data in the basic data to form derived data, and therefore, a cross function and a mutation function may be set, the generated basic data may be subjected to cross processing and/or mutation processing through the cross function and/or the mutation function, and finally the data obtained through the cross processing and/or the mutation processing may be obtained, and then the fitness calculation may be performed on the data obtained through the cross processing and/or the mutation processing, obtaining second data, matching the obtained second data with the first selection condition to determine whether the second data satisfies the first selection condition, if the second data satisfies the first selection condition, the process of the following step S316 may be performed, if the second data does not satisfy the first selection condition, the data obtained by the cross processing and/or mutation processing can be discarded, thereby finishing the derivation operation of the basic data, or, the data obtained by the cross processing and/or mutation processing which does not meet the first selection condition can be processed by the cross processing and/or mutation processing again to generate new data, then, whether the new data meets the first selection condition is judged again, and the above processing procedure can be repeated until the new data meets the first selection condition or the number of times of the executed crossover processing and/or mutation processing reaches the preset number threshold.
In step S316, the index corresponding to the data obtained by the last crossover processing and/or mutation processing is determined as the derived index corresponding to the target data.
The embodiment of the specification provides a data derivation method, which includes generating one or more grids covering target data by performing grid processing on the target data, then generating single-dimensional information and multi-dimensional information according to the information contained in each grid, determining a basic index corresponding to the target data according to the single-dimensional information and the multi-dimensional information, and time periods and event types contained in each grid, and finally performing derivation processing on the basic index through an index derivation algorithm to obtain a derived index corresponding to the target data, so that the grids are automatically divided through discrete target data, the obtained grids are listed, meanwhile, the single-dimensional information and the multi-dimensional information are generated according to the grids to further construct the basic index, and the basic index is determined not only through the single-dimensional information but also through the multi-dimensional information, therefore, the content contained in the basic index is wider, a data base is provided for subsequent processing, in addition, the aggregation of elements in the basic index is realized by means of an index derivative algorithm, and then the derivative index corresponding to the target data is generated, so that the index corresponding to the target data can be expanded to a certain extent, and further the characteristic corresponding to the index can be more precise and comprehensive, so that the accuracy of the subsequent processing is improved, for example, the obtained risk prevention and control rule can be more accurate, and the like.
Example four
Based on the same idea, the data derivation method provided in the embodiments of the present specification further provides a data derivation apparatus, as shown in fig. 4.
The data derivation means comprises: a gridding module 401, a dimension generation module 402, a base index determination module 403, and an index derivation module 404, wherein:
a gridding module 401, configured to perform gridding processing on target data to generate one or more grids covering the target data;
a dimension generating module 402, configured to generate dimension information according to information included in each grid, where the dimension information includes single-dimension information and multi-dimension information;
a basic index determining module 403, configured to determine a basic index corresponding to the target data according to the dimension information, and a time period and an event type included in each grid;
and an index derivation module 404, configured to perform derivation processing on the basic index through an index derivation algorithm to obtain a derived index corresponding to the target data.
In the embodiments of the present specification, the index derivation algorithm includes a genetic algorithm, a random walk algorithm, and a violence derivation algorithm.
In this embodiment of the present specification, the index derivation module 404 includes:
the judgment unit is used for judging whether the basic indexes meet first selection conditions corresponding to the index derivation algorithm;
the index derivation unit is used for performing at least one derivation process on the basic index if the basic index is not satisfied, and judging whether the index obtained by the derivation process satisfies the first selection condition after the derivation process is performed each time until the index obtained by the derivation process satisfies the first selection condition or the number of times of the derivation process performed reaches a preset number threshold; and determining the index obtained by the last derivation treatment as the derivation index corresponding to the target data.
In an embodiment of this specification, the determining unit is configured to:
coding the basic indexes to obtain basic data;
carrying out fitness calculation on the basic data to obtain first data;
judging whether the first data meet the first selection condition or not;
if the first data meets the first selection condition, the basic index meets the first selection condition;
if the first data does not meet the first selection condition, the basic index does not meet the first selection condition;
the index derivation unit is used for performing cross processing and/or mutation processing on the basic data corresponding to the basic indexes.
In the embodiments of the present specification, the fitness is obtained by aggregating a plurality of predetermined indexes.
In the embodiment of the present specification, the plurality of predetermined indexes include at least two of an area AUC under a receptivity ROC curve, an information amount IV, and an Embedding layer Embedding.
In an embodiment of the present specification, the event types include a payment class and an operation class, the events of the payment class include a transaction event, a transfer event, and a code scanning event, and the events of the operation class include a login event, a password modification event, a registration event, and a browsing event.
In this embodiment of the present specification, the gridding module 401 is configured to perform gridding processing on target data in a random division manner to generate one or more grids covering the target data.
The embodiment of the present specification provides a data derivation apparatus, which generates one or more grids covering target data by performing grid processing on target data, then generates single-dimensional information and multi-dimensional information according to information included in each grid, determines a basic index corresponding to the target data according to the single-dimensional information and the multi-dimensional information, and a time period and an event type included in each grid, and finally performs derivation processing on the basic index through an index derivation algorithm to obtain a derived index corresponding to the target data, so that the grids are automatically divided by discrete target data, the obtained grids are listed, and meanwhile, the single-dimensional information and the multi-dimensional information are generated according to the grids to construct the basic index, which is determined not only by the single-dimensional information but also by the multi-dimensional information, therefore, the content contained in the basic index is wider, a data base is provided for subsequent processing, in addition, the aggregation of elements in the basic index is realized by means of an index derivative algorithm, and then the derivative index corresponding to the target data is generated, so that the index corresponding to the target data can be expanded to a certain extent, and further the characteristic corresponding to the index can be more precise and comprehensive, so that the accuracy of the subsequent processing is improved, for example, the obtained risk prevention and control rule can be more accurate, and the like.
EXAMPLE five
Based on the same idea, the data derivation apparatus provided in the embodiments of the present specification further provides a data derivation device, as shown in fig. 5.
The derivative device of the data may be the server or the terminal device provided in the above embodiments.
The derivative devices of the data may have large differences due to different configurations or performances, and may include one or more processors 501 and memories 502, and the memories 502 may store one or more stored applications or data. Memory 502 may be, among other things, transient or persistent storage. The application program stored in memory 502 may include one or more modules (not shown), each of which may include a series of computer-executable instructions in a derivative device for data. Still further, the processor 501 may be arranged in communication with the memory 502 to execute a series of computer-executable instructions in the memory 502 on a device that is derivative of the data. The derivation of data may also include one or more power supplies 503, one or more wired or wireless network interfaces 504, one or more input-output interfaces 505, and one or more keyboards 506.
In particular, in this embodiment, the apparatus for deriving data includes a memory, and one or more programs, where the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for the apparatus for deriving data, and the one or more programs configured to be executed by the one or more processors include computer-executable instructions for:
performing gridding processing on target data to generate one or more grids covering the target data;
generating dimension information according to information contained in each grid, wherein the dimension information comprises single-dimension information and multi-dimension information;
determining a basic index corresponding to the target data according to the dimension information, and the time period and the event type contained in each grid;
and deriving the basic indexes through an index derivation algorithm to obtain derived indexes corresponding to the target data.
In the embodiments of the present specification, the index derivation algorithm includes a genetic algorithm, a random walk algorithm, and a violence derivation algorithm.
In an embodiment of this specification, the deriving the basic index through an index derivation algorithm to obtain a derived index corresponding to the target data includes:
judging whether the basic indexes meet first selection conditions corresponding to the index derivation algorithm or not;
if not, performing at least one derivation process on the basic index, and judging whether the index obtained by the derivation process meets the first selection condition after the derivation process is performed each time until the index obtained by the derivation process meets the first selection condition or the number of the performed derivation processes reaches a preset number threshold; and determining the index obtained by the last derivation treatment as the derivation index corresponding to the target data.
In an embodiment of this specification, the determining whether the basic indicator meets a first selection condition corresponding to the indicator derivation algorithm includes:
coding the basic indexes to obtain basic data;
carrying out fitness calculation on the basic data to obtain first data;
judging whether the first data meet the first selection condition or not;
if the first data meets the first selection condition, the basic index meets the first selection condition;
if the first data does not meet the first selection condition, the basic index does not meet the first selection condition;
the deriving the basic index comprises the following steps:
and performing cross processing and/or mutation processing on the basic data corresponding to the basic indexes.
In the embodiments of the present specification, the fitness is obtained by aggregating a plurality of predetermined indexes.
In the embodiment of the present specification, the plurality of predetermined indexes include at least two of an area AUC under a receptivity ROC curve, an information amount IV, and an Embedding layer Embedding.
In an embodiment of the present specification, the event types include a payment class and an operation class, the events of the payment class include a transaction event, a transfer event, and a code scanning event, and the events of the operation class include a login event, a password modification event, a registration event, and a browsing event.
In this embodiment of this specification, the performing gridding processing on target data to generate one or more grids covering the target data includes:
and carrying out gridding processing on the target data in a random division mode to generate one or more grids covering the target data.
The embodiment of the present specification provides a data derivation apparatus, which generates one or more grids covering target data by performing grid processing on target data, then generates single-dimensional information and multi-dimensional information according to information included in each grid, determines a basic index corresponding to the target data according to the single-dimensional information and the multi-dimensional information, and a time period and an event type included in each grid, and finally performs derivation processing on the basic index through an index derivation algorithm to obtain a derived index corresponding to the target data, so that the grids are automatically divided by discrete target data, the obtained grids are listed, and meanwhile, the single-dimensional information and the multi-dimensional information are generated according to the grids to construct the basic index, which is determined not only by the single-dimensional information but also by the multi-dimensional information, therefore, the content contained in the basic index is wider, a data base is provided for subsequent processing, in addition, the aggregation of elements in the basic index is realized by means of an index derivative algorithm, and then the derivative index corresponding to the target data is generated, so that the index corresponding to the target data can be expanded to a certain extent, and further the characteristic corresponding to the index can be more precise and comprehensive, so that the accuracy of the subsequent processing is improved, for example, the obtained risk prevention and control rule can be more accurate, and the like.
The foregoing description has been directed to specific embodiments of this disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
In the 90 s of the 20 th century, improvements in a technology could clearly distinguish between improvements in hardware (e.g., improvements in circuit structures such as diodes, transistors, switches, etc.) and improvements in software (improvements in process flow). However, as technology advances, many of today's process flow improvements have been seen as direct improvements in hardware circuit architecture. Designers almost always obtain the corresponding hardware circuit structure by programming an improved method flow into the hardware circuit. Thus, it cannot be said that an improvement in the process flow cannot be realized by hardware physical modules. For example, a Programmable Logic Device (PLD), such as a Field Programmable Gate Array (FPGA), is an integrated circuit whose Logic functions are determined by programming the Device by a user. A digital system is "integrated" on a PLD by the designer's own programming without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Furthermore, nowadays, instead of manually making an Integrated Circuit chip, such Programming is often implemented by "logic compiler" software, which is similar to a software compiler used in program development and writing, but the original code before compiling is also written by a specific Programming Language, which is called Hardware Description Language (HDL), and HDL is not only one but many, such as abel (advanced Boolean Expression Language), ahdl (alternate Hardware Description Language), traffic, pl (core universal Programming Language), HDCal (jhdware Description Language), lang, Lola, HDL, laspam, hardward Description Language (vhr Description Language), vhal (Hardware Description Language), and vhigh-Language, which are currently used in most common. It will also be apparent to those skilled in the art that hardware circuitry that implements the logical method flows can be readily obtained by merely slightly programming the method flows into an integrated circuit using the hardware description languages described above.
The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, an Application Specific Integrated Circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic for the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may thus be considered a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be regarded as being both a software module for performing the method and a structure within a hardware component.
The systems, devices, modules or units illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. One typical implementation device is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.
For convenience of description, the above devices are described as being divided into various units by function, and are described separately. Of course, the functionality of the various elements may be implemented in the same one or more software and/or hardware implementations in implementing one or more embodiments of the present description.
As will be appreciated by one skilled in the art, embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, one or more embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of the present description may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
Embodiments of the present description are described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the description. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
As will be appreciated by one skilled in the art, embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, one or more embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of the present description may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
One or more embodiments of the present description may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. One or more embodiments of the specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only an example of the present specification, and is not intended to limit the present specification. Various modifications and alterations to this description will become apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification should be included in the scope of the claims of the present specification.