CN113313578A - Credit card consumption staging decision method and device - Google Patents
Credit card consumption staging decision method and device Download PDFInfo
- Publication number
- CN113313578A CN113313578A CN202110662992.4A CN202110662992A CN113313578A CN 113313578 A CN113313578 A CN 113313578A CN 202110662992 A CN202110662992 A CN 202110662992A CN 113313578 A CN113313578 A CN 113313578A
- Authority
- CN
- China
- Prior art keywords
- training data
- consumption
- decision tree
- decision
- model
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000003066 decision tree Methods 0.000 claims abstract description 95
- 238000012545 processing Methods 0.000 claims abstract description 13
- 238000012549 training Methods 0.000 claims description 114
- 238000003062 neural network model Methods 0.000 claims description 31
- 238000013528 artificial neural network Methods 0.000 claims description 24
- 238000005457 optimization Methods 0.000 claims description 19
- 238000013138 pruning Methods 0.000 claims description 16
- 238000010276 construction Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 description 26
- 238000010586 diagram Methods 0.000 description 4
- 238000013507 mapping Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000004308 accommodation Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000002945 steepest descent method Methods 0.000 description 1
Images
Classifications
-
- 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
- G06Q40/03—Credit; Loans; Processing thereof
-
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
Abstract
The invention discloses a credit card consumption staging decision method and a device, which are used for acquiring the current consumption transaction related information of a credit card, extracting transaction characteristic data from the current consumption transaction related information, carrying out discretization and classification processing on the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes, and inputting the characteristic category attribute set into a pre-constructed decision tree model to obtain the staging decision result of the current consumption transaction. According to the invention, intelligent decision can be made according to the relevant information of each consumption transaction to determine whether the consumption transaction is staged, and the whole decision process does not need manual participation, so that the labor cost is reduced, and the decision efficiency of consumption staging is improved.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a credit card consumption staging decision method and a credit card consumption staging decision device.
Background
Credit cards, also known as debit cards, are a form of payment for cashless transactions and are simple credit services. The credit card consumption stage means that after the card holder is charged by swiping the card, if the single consumption amount reaches the minimum consumption amount specified by the bank, the card holder can apply to the bank to split the consumption into several repayment, and the bank divides the consumption into several times to be charged into the credit card account of the card holder.
In the existing credit card consumption stage, when a customer service determines that a customer meets a stage condition according to the consumption amount of the customer, the customer carries out telemarketing on the customer and recommends the customer to handle the credit card consumption stage, and if the customer agrees the consumption stage, a bank charges a certain amount of commission charge through the stage number.
However, the existing credit card consumption staging decision process requires a lot of manual involvement, and thus, a lot of labor cost is required.
Disclosure of Invention
In view of the above, the invention discloses a credit card consumption staging decision method and device, so as to realize intelligent decision according to the relevant information of each consumption transaction, determine whether the consumption transaction is staged, and avoid manual participation in the whole decision process, thereby reducing the labor cost and improving the decision efficiency of consumption staging.
A credit card consumption staging decision method, comprising:
acquiring the relevant information of the consumption transaction of the credit card;
extracting transaction characteristic data from the consumption transaction related information;
discretizing and classifying the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes;
inputting the feature category attribute set into a pre-constructed decision tree model to obtain a staging decision result of the current consumption transaction, wherein the staging decision result comprises: agreeing to the credit card consumption installment or disagreeing with the credit card consumption installment.
Optionally, the information related to the current consumption transaction includes: current consumption transaction basic information and consumption customer information.
Optionally, the process of constructing the decision tree model includes:
acquiring historical consumption transaction related information in a preset time period;
extracting the characteristic category attribute corresponding to each consumption transaction from the historical consumption transaction related information to obtain a training data set containing the characteristic category attributes corresponding to all consumption transactions;
placing the training data set into a root node;
and based on a loss function, dividing the training data set into a plurality of training data subsets, distributing each training data subset to a corresponding leaf node until all the training data subsets are distributed, and constructing to obtain the decision tree model.
Optionally, the segmenting the training data set into a plurality of training data subsets based on a loss function, and allocating each training data subset to a corresponding leaf node until all the training data subsets are allocated, so as to construct the decision tree model, which specifically includes:
selecting an optimal feature attribute from the training data set put into the root node based on the loss function;
segmenting the training data set based on the optimal characteristic attributes to obtain a plurality of training data subsets;
constructing leaf nodes, and dividing each training data subset into corresponding leaf nodes;
selecting a new optimal feature attribute from the subset of training data placed into each of the leaf nodes based on the loss function;
segmenting the training data subset based on the new optimal feature attributes;
and repeating the process until all the training data subsets are distributed to the corresponding leaf nodes to obtain the decision tree model.
Optionally, after the decision tree model is constructed, the method further includes:
and carrying out pruning optimization on the decision tree model by using the BP neural network to obtain an optimized decision tree model.
Optionally, the pruning optimization is performed on the decision tree model by using the BP neural network to obtain an optimized decision tree model, which specifically includes:
taking the non-leaf nodes of the decision tree model as an input layer of the BP neural network, and taking whether credit card consumption stages are agreed or not as an output layer of the BP neural network to construct a neural network model;
randomly initializing weights in the neural network model, wherein the weights represent the importance degree of corresponding leaf nodes in the decision tree model;
training the neural network model based on the weight to obtain an error parameter of the neural network model;
adjusting the weight value based on the error parameter to obtain the latest weight value;
and training the neural network model again based on the latest weight, repeating the training until the weight is converged or the iteration times reach the preset times to obtain the final weight, completing pruning optimization of the decision tree model, and obtaining the optimized decision tree model.
A credit card consumption staging decision making apparatus comprising:
the acquisition unit is used for acquiring the relevant information of the current consumption transaction of the credit card;
the extraction unit is used for extracting transaction characteristic data from the consumption transaction related information;
the processing unit is used for carrying out discretization and classification processing on the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes;
a decision result determining unit, configured to input the feature category attribute set to a pre-constructed decision tree model to obtain a staging decision result of the current consumption transaction, where the staging decision result includes: agreeing to the credit card consumption installment or disagreeing with the credit card consumption installment.
Optionally, the method further includes: a model construction unit;
the model construction unit is used for constructing the decision tree model;
the model building unit specifically comprises:
the acquisition subunit is used for acquiring historical consumption transaction related information within a preset time period;
the extraction subunit is used for extracting the characteristic category attribute corresponding to each consumption transaction from the historical consumption transaction related information to obtain a training data set containing the characteristic category attributes corresponding to all consumption transactions;
a root node determining subunit, configured to place the training data set into a root node;
and the model determining subunit is used for dividing the training data set into a plurality of training data subsets based on a loss function, distributing each training data subset to a corresponding leaf node until all the training data subsets are distributed, and constructing to obtain the decision tree model.
Optionally, the model determination subunit is specifically configured to:
selecting an optimal feature attribute from the training data set put into the root node based on the loss function;
segmenting the training data set based on the optimal characteristic attributes to obtain a plurality of training data subsets;
constructing leaf nodes, and dividing each training data subset into corresponding leaf nodes;
selecting a new optimal feature attribute from the subset of training data placed into each of the leaf nodes based on the loss function;
segmenting the training data subset based on the new optimal feature attributes;
and repeating the process until all the training data subsets are distributed to the corresponding leaf nodes to obtain the decision tree model.
Optionally, the model building unit further includes:
and the model optimization subunit is used for pruning and optimizing the decision tree model by using a BP (back propagation) neural network after the decision tree model is constructed and obtained, so as to obtain the optimized decision tree model.
Optionally, the model optimization subunit is specifically configured to:
taking the non-leaf nodes of the decision tree model as an input layer of the BP neural network, and taking whether credit card consumption stages are agreed or not as an output layer of the BP neural network to construct a neural network model;
randomly initializing weights in the neural network model, wherein the weights represent the importance degree of corresponding leaf nodes in the decision tree model;
training the neural network model based on the weight to obtain an error parameter of the neural network model;
adjusting the weight value based on the error parameter to obtain the latest weight value;
and training the neural network model again based on the latest weight, repeating the training until the weight is converged or the iteration times reach the preset times to obtain the final weight, completing pruning optimization of the decision tree model, and obtaining the optimized decision tree model.
According to the technical scheme, the invention discloses a credit card consumption stage decision method and a credit card consumption stage decision device, which are used for acquiring the relevant information of the current consumption transaction of a credit card, extracting transaction characteristic data from the relevant information of the current consumption transaction, carrying out discretization and classification processing on the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes, and inputting the characteristic category attribute set into a pre-constructed decision tree model to obtain a stage decision result of the current consumption transaction. According to the invention, intelligent decision can be made according to the relevant information of each consumption transaction to determine whether the consumption transaction is staged, and the whole decision process does not need manual participation, so that the labor cost is reduced, and the decision efficiency of consumption staging is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the disclosed drawings without creative efforts.
FIG. 1 is a flowchart of a credit card consumption staging decision method according to an embodiment of the present invention;
FIG. 2 is a flowchart of a method for constructing a decision tree model according to an embodiment of the present invention;
FIG. 3 is a flowchart of a method for pruning and optimizing a decision tree model using a BP neural network to obtain an optimized decision tree model according to an embodiment of the present invention;
FIG. 4 is a schematic structural diagram of a credit card consumption staging decision device according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a model building unit according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the invention discloses a credit card consumption staging decision method and a device, which are used for acquiring the relevant information of the current consumption transaction of a credit card, extracting transaction characteristic data from the relevant information of the current consumption transaction, carrying out discretization and classification processing on the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes, and inputting the characteristic category attribute set into a pre-constructed decision tree model to obtain the staging decision result of the current consumption transaction. According to the invention, intelligent decision can be made according to the relevant information of each consumption transaction to determine whether the consumption transaction is staged, and the whole decision process does not need manual participation, so that the labor cost is reduced, and the decision efficiency of consumption staging is improved.
Referring to fig. 1, a flowchart of a credit card consumption staging decision method disclosed in an embodiment of the present invention includes:
step S101, acquiring the relevant information of the current consumption transaction of the credit card;
the consumption transaction in this embodiment refers to consumption of shopping, traveling, dining, transportation, education, and the like, which occurs on a POS machine or an online payment platform of a merchant by a user using a credit card as a payment settlement tool.
The information related to the consumption transaction may specifically include: current consumption transaction basic information and consumption customer information.
The current consumption transaction basic information includes consumption amount, consumption time, consumption place, consumption mode, consumption channel, consumption category, consumption attribute and the like.
Wherein, the consumption category is actually a transaction category, including: the invention only selects the related data of credit card transaction category.
The consumption attributes include: catering consumption, local traffic consumption, shopping consumption, educational study consumption, healthcare consumption, lodging consumption and other consumption.
The consuming client information includes: age of the customer, gender of the customer, marital status of the customer, credit rating of the customer, monthly transaction amount, staging frequency, etc.
Step S102, transaction characteristic data is extracted from the consumption transaction related information;
the transaction characteristic data includes, for example, the gender of the customer, and the consumption amount of the large amount or the small amount.
Step S103, discretizing and classifying the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes;
the value of each feature attribute of the feature category attributes is in a fixed range, and data can be classified according to the feature category attributes, for example, the feature category attributes are gender and are classified into male and female.
It should be noted that the decision tree model requires that the input data is a feature category attribute, and therefore discretization and classification processing needs to be performed on the transaction feature data to obtain a feature category attribute set including a plurality of feature category attributes. For example, the consumption amount is segmented into: the consumption time is divided into four sections, namely, 300, 500, 2000: five intervals of [23:00-09:00, 09:00-12:00, 12:00-14:00, 14:00-17:00, 17:00-23:00 ].
And step S104, inputting the characteristic category attribute set into a pre-constructed decision tree model to obtain a stage decision result of the current consumption transaction.
Wherein the staging decision result comprises: agreeing to the credit card consumption installment or disagreeing with the credit card consumption installment.
In summary, the invention discloses a credit card consumption staging decision method, which comprises the steps of obtaining the relevant information of the current consumption transaction of a credit card, extracting transaction characteristic data from the relevant information of the current consumption transaction, carrying out discretization and classification processing on the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes, and inputting the characteristic category attribute set into a pre-constructed decision tree model to obtain the staging decision result of the current consumption transaction. According to the invention, intelligent decision can be made according to the relevant information of each consumption transaction to determine whether the consumption transaction is staged, and the whole decision process does not need manual participation, so that the labor cost is reduced, and the decision efficiency of consumption staging is improved.
In order to further optimize the above embodiment, referring to fig. 2, a flowchart of a method for constructing a decision tree model disclosed in the embodiment of the present invention includes:
step S201, obtaining historical consumption transaction related information in a preset time period;
the value of the preset time period is determined according to actual needs, and the invention is not limited herein.
The historical consumption transaction related information comprises consumption transaction related information corresponding to each consumption transaction in a preset time period.
Step S202, extracting the characteristic category attribute corresponding to each consumption transaction from the historical consumption transaction related information to obtain a training data set containing the characteristic category attributes corresponding to all consumption transactions;
the process of extracting the feature category attribute from the historical consumption transaction related information may refer to the corresponding content in the embodiment shown in fig. 1, and is not described herein again.
Step S203, putting the training data set into a root node;
this embodiment starts with a root node and builds a decision tree model based on a training data set.
And S204, based on a loss function, dividing the training data set into a plurality of training data subsets, distributing each training data subset to a corresponding leaf node until all the training data subsets are distributed, and constructing to obtain the decision tree model.
The loss function (loss function) is a function that maps a random event or a value of a random variable related to the random event to a non-negative real number to represent a "risk" or a "loss" of the random event. In application, the loss function is usually associated with the optimization problem as a learning criterion, i.e. the model is solved and evaluated by minimizing the loss function.
The invention segments a training data set into a plurality of training data subsets to construct a decision tree model based on a loss function.
Step S204 may specifically include:
selecting an optimal feature attribute from the training data set put into the root node based on a loss function;
segmenting the training data set based on the optimal characteristic attributes to obtain a plurality of training data subsets;
constructing leaf nodes, and dividing each training data subset into corresponding leaf nodes;
selecting a new optimal feature attribute from the subset of training data placed into each of the leaf nodes based on the loss function;
segmenting the training data subset based on the new optimal feature attributes;
and repeating the process until all the training data subsets are distributed to the corresponding leaf nodes to obtain the decision tree model.
The method comprises the steps of starting from a root node, gradually determining each leaf node by segmenting a training data set based on a loss function until all training data subsets are correctly classified or have no proper characteristics, and distributing all training data subsets to one leaf node to form a decision tree model.
Generally, an overfitting phenomenon may occur in a generated decision tree model, and the overfitting phenomenon refers to a situation that the decision tree model has a high decision accuracy in training data, but has a low decision accuracy in a test set.
In order to avoid the over-fitting phenomenon of the decision tree model, after step S204, the method may further include:
and (4) pruning and optimizing the decision tree model by using the BP neural network to obtain an optimized decision tree model.
Bp (back propagation) neural network, a multi-layer feedforward network trained according to an error inverse propagation algorithm, is one of the most widely applied neural network models. The BP neural network is able to learn and store a large number of input-output pattern mappings without prior disclosure of mathematical equations describing such mappings. The learning rule is that the steepest descent method is used, and the weight and the threshold value of the network are continuously adjusted through back propagation, so that the error square sum of the network is minimum.
Preferably, the decision tree model used in step S104 is preferably an optimized decision tree model.
In order to further optimize the above embodiment, referring to fig. 3, a flowchart of a method for performing pruning optimization on a decision tree model by using a BP neural network to obtain an optimized decision tree model is disclosed in the embodiment of the present invention, and the method includes:
s301, taking a non-leaf node of the decision tree model as an input layer of the BP neural network, taking whether to agree with credit card consumption staging as an output layer of the BP neural network, and constructing a neural network model;
non-leaf nodes of the decision tree model, such as customer category "female", consumption attribute "accommodation consumption", and the like.
Step S302, randomly initializing weights in the neural network model;
wherein the weight value represents the importance degree of the corresponding leaf node in the decision tree model.
Step S303, training the neural network model based on the weight to obtain an error parameter of the neural network model;
step S304, adjusting the weight value based on the error parameter to obtain the latest weight value;
step S305, training the neural network model again based on the latest weight value, repeating the training until the weight value is converged or the iteration times reach the preset times to obtain the final weight value, completing pruning optimization of the decision tree model, and obtaining the optimized decision tree model.
Corresponding to the embodiment of the method, the invention also discloses a credit card consumption staging decision device.
Referring to fig. 4, a schematic structural diagram of a credit card consumption staging decision device disclosed in an embodiment of the present invention includes:
an obtaining unit 401, configured to obtain relevant information of the current consumption transaction of the credit card;
the consumption transaction in this embodiment refers to consumption of shopping, traveling, dining, transportation, education, and the like, which occurs on a POS machine or an online payment platform of a merchant by a user using a credit card as a payment settlement tool.
The information related to the consumption transaction may specifically include: current consumption transaction basic information and consumption customer information.
The current consumption transaction basic information includes consumption amount, consumption time, consumption place, consumption mode, consumption channel, consumption category, consumption attribute and the like.
Wherein, the consumption category is actually a transaction category, including: the invention only selects the related data of credit card transaction category.
The consumption attributes include: catering consumption, local traffic consumption, shopping consumption, educational study consumption, healthcare consumption, lodging consumption and other consumption.
The consuming client information includes: age of the customer, gender of the customer, marital status of the customer, credit rating of the customer, monthly transaction amount, staging frequency, etc.
An extracting unit 402, configured to extract transaction characteristic data from the relevant information of the current consumption transaction;
the transaction characteristic data includes, for example, the gender of the customer, and the consumption amount of the large amount or the small amount.
The processing unit 403 is configured to perform discretization and classification processing on the transaction feature data to obtain a feature category attribute set including a plurality of feature category attributes;
the value of each feature attribute of the feature category attributes is in a fixed range, and data can be classified according to the feature category attributes, for example, the feature category attributes are gender and are classified into male and female.
A decision result determining unit 404, configured to input the feature category attribute set to a pre-constructed decision tree model to obtain a staging decision result of the current consumption transaction, where the staging decision result includes: agreeing to the credit card consumption installment or disagreeing with the credit card consumption installment.
In summary, the invention discloses a credit card consumption staging decision device, which is used for acquiring the relevant information of the current consumption transaction of a credit card, extracting transaction characteristic data from the relevant information of the current consumption transaction, carrying out discretization and classification processing on the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes, and inputting the characteristic category attribute set into a pre-constructed decision tree model to obtain the staging decision result of the current consumption transaction. According to the invention, intelligent decision can be made according to the relevant information of each consumption transaction to determine whether the consumption transaction is staged, and the whole decision process does not need manual participation, so that the labor cost is reduced, and the decision efficiency of consumption staging is improved.
In the above embodiment, the decision device may further include: a model construction unit;
the model construction unit is used for constructing the decision tree model.
For further optimizing the above embodiment, referring to fig. 5, a schematic structural diagram of a model building unit disclosed in the embodiment of the present invention is provided, where the model building unit specifically includes:
an obtaining subunit 501, configured to obtain historical consumption transaction related information within a preset time period;
the value of the preset time period is determined according to actual needs, and the invention is not limited herein.
The historical consumption transaction related information comprises consumption transaction related information corresponding to each consumption transaction in a preset time period.
An extracting subunit 502, configured to extract a feature category attribute corresponding to each consumption transaction from the historical consumption transaction related information, to obtain a training data set including feature category attributes corresponding to all consumption transactions;
a root node determining subunit 503, configured to place the training data set into a root node;
a model determining subunit 504, configured to divide the training data set into a plurality of training data subsets based on a loss function, allocate each training data subset to a corresponding leaf node, and construct the decision tree model until all the training data subsets are allocated.
The loss function (loss function) is a function that maps a random event or a value of a random variable related to the random event to a non-negative real number to represent a "risk" or a "loss" of the random event. In application, the loss function is usually associated with the optimization problem as a learning criterion, i.e. the model is solved and evaluated by minimizing the loss function.
The invention segments a training data set into a plurality of training data subsets to construct a decision tree model based on a loss function.
The model determining subunit 504 is specifically configured to:
selecting an optimal feature attribute from the training data set put into the root node based on the loss function;
segmenting the training data set based on the optimal characteristic attributes to obtain a plurality of training data subsets;
constructing leaf nodes, and dividing each training data subset into corresponding leaf nodes;
selecting a new optimal feature attribute from the subset of training data placed into each of the leaf nodes based on the loss function;
segmenting the training data subset based on the new optimal feature attributes;
and repeating the process until all the training data subsets are distributed to the corresponding leaf nodes to obtain the decision tree model.
Generally, an overfitting phenomenon may occur in a generated decision tree model, and the overfitting phenomenon refers to a situation that the decision tree model has a high decision accuracy in training data, but has a low decision accuracy in a test set.
In order to avoid the over-fitting phenomenon of the decision tree model, the model construction unit may further include:
and the model optimization subunit is used for pruning and optimizing the decision tree model by using a BP (back propagation) neural network after the decision tree model is constructed and obtained, so as to obtain the optimized decision tree model.
Wherein the model optimization subunit may be specifically configured to:
taking the non-leaf nodes of the decision tree model as an input layer of the BP neural network, and taking whether credit card consumption stages are agreed or not as an output layer of the BP neural network to construct a neural network model;
randomly initializing weights in the neural network model, wherein the weights represent the importance degree of corresponding leaf nodes in the decision tree model;
training the neural network model based on the weight to obtain an error parameter of the neural network model;
adjusting the weight value based on the error parameter to obtain the latest weight value;
and training the neural network model again based on the latest weight, repeating the training until the weight is converged or the iteration times reach the preset times to obtain the final weight, completing pruning optimization of the decision tree model, and obtaining the optimized decision tree model.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, 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 identical elements in a process, method, article, or apparatus that comprises the element.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (11)
1. A credit card consumption staging decision method, comprising:
acquiring the relevant information of the consumption transaction of the credit card;
extracting transaction characteristic data from the consumption transaction related information;
discretizing and classifying the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes;
inputting the feature category attribute set into a pre-constructed decision tree model to obtain a staging decision result of the current consumption transaction, wherein the staging decision result comprises: agreeing to the credit card consumption installment or disagreeing with the credit card consumption installment.
2. The decision method of claim 1, wherein the current consumption transaction-related information comprises: current consumption transaction basic information and consumption customer information.
3. The decision method according to claim 1, wherein the construction process of the decision tree model comprises:
acquiring historical consumption transaction related information in a preset time period;
extracting the characteristic category attribute corresponding to each consumption transaction from the historical consumption transaction related information to obtain a training data set containing the characteristic category attributes corresponding to all consumption transactions;
placing the training data set into a root node;
and based on a loss function, dividing the training data set into a plurality of training data subsets, distributing each training data subset to a corresponding leaf node until all the training data subsets are distributed, and constructing to obtain the decision tree model.
4. The decision method according to claim 3, wherein the constructing the decision tree model by dividing the training data set into a plurality of training data subsets based on a loss function and allocating each of the training data subsets to a corresponding leaf node until all of the training data subsets are allocated comprises:
selecting an optimal feature attribute from the training data set put into the root node based on the loss function;
segmenting the training data set based on the optimal characteristic attributes to obtain a plurality of training data subsets;
constructing leaf nodes, and dividing each training data subset into corresponding leaf nodes;
selecting a new optimal feature attribute from the subset of training data placed into each of the leaf nodes based on the loss function;
segmenting the training data subset based on the new optimal feature attributes;
and repeating the process until all the training data subsets are distributed to the corresponding leaf nodes to obtain the decision tree model.
5. The decision method of claim 3, further comprising, after constructing the decision tree model,:
and carrying out pruning optimization on the decision tree model by using the BP neural network to obtain an optimized decision tree model.
6. The decision method according to claim 5, wherein the pruning optimization of the decision tree model by using the BP neural network to obtain the optimized decision tree model specifically comprises:
taking the non-leaf nodes of the decision tree model as an input layer of the BP neural network, and taking whether credit card consumption stages are agreed or not as an output layer of the BP neural network to construct a neural network model;
randomly initializing weights in the neural network model, wherein the weights represent the importance degree of corresponding leaf nodes in the decision tree model;
training the neural network model based on the weight to obtain an error parameter of the neural network model;
adjusting the weight value based on the error parameter to obtain the latest weight value;
and training the neural network model again based on the latest weight, repeating the training until the weight is converged or the iteration times reach the preset times to obtain the final weight, completing pruning optimization of the decision tree model, and obtaining the optimized decision tree model.
7. A credit card consumption staging decision-making device, comprising:
the acquisition unit is used for acquiring the relevant information of the current consumption transaction of the credit card;
the extraction unit is used for extracting transaction characteristic data from the consumption transaction related information;
the processing unit is used for carrying out discretization and classification processing on the transaction characteristic data to obtain a characteristic category attribute set containing a plurality of characteristic category attributes;
a decision result determining unit, configured to input the feature category attribute set to a pre-constructed decision tree model to obtain a staging decision result of the current consumption transaction, where the staging decision result includes: agreeing to the credit card consumption installment or disagreeing with the credit card consumption installment.
8. The decision-making device according to claim 7, further comprising: a model construction unit;
the model construction unit is used for constructing the decision tree model;
the model building unit specifically comprises:
the acquisition subunit is used for acquiring historical consumption transaction related information within a preset time period;
the extraction subunit is used for extracting the characteristic category attribute corresponding to each consumption transaction from the historical consumption transaction related information to obtain a training data set containing the characteristic category attributes corresponding to all consumption transactions;
a root node determining subunit, configured to place the training data set into a root node;
and the model determining subunit is used for dividing the training data set into a plurality of training data subsets based on a loss function, distributing each training data subset to a corresponding leaf node until all the training data subsets are distributed, and constructing to obtain the decision tree model.
9. The decision apparatus according to claim 8, wherein the model determination subunit is specifically configured to:
selecting an optimal feature attribute from the training data set put into the root node based on the loss function;
segmenting the training data set based on the optimal characteristic attributes to obtain a plurality of training data subsets;
constructing leaf nodes, and dividing each training data subset into corresponding leaf nodes;
selecting a new optimal feature attribute from the subset of training data placed into each of the leaf nodes based on the loss function;
segmenting the training data subset based on the new optimal feature attributes;
and repeating the process until all the training data subsets are distributed to the corresponding leaf nodes to obtain the decision tree model.
10. The decision-making device according to claim 8, wherein the model construction unit further comprises:
and the model optimization subunit is used for pruning and optimizing the decision tree model by using a BP (back propagation) neural network after the decision tree model is constructed and obtained, so as to obtain the optimized decision tree model.
11. The decision-making device according to claim 10, wherein the model optimization subunit is specifically configured to:
taking the non-leaf nodes of the decision tree model as an input layer of the BP neural network, and taking whether credit card consumption stages are agreed or not as an output layer of the BP neural network to construct a neural network model;
randomly initializing weights in the neural network model, wherein the weights represent the importance degree of corresponding leaf nodes in the decision tree model;
training the neural network model based on the weight to obtain an error parameter of the neural network model;
adjusting the weight value based on the error parameter to obtain the latest weight value;
and training the neural network model again based on the latest weight, repeating the training until the weight is converged or the iteration times reach the preset times to obtain the final weight, completing pruning optimization of the decision tree model, and obtaining the optimized decision tree model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110662992.4A CN113313578B (en) | 2021-06-15 | 2021-06-15 | Decision method and device for credit card consumption stage |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110662992.4A CN113313578B (en) | 2021-06-15 | 2021-06-15 | Decision method and device for credit card consumption stage |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113313578A true CN113313578A (en) | 2021-08-27 |
| CN113313578B CN113313578B (en) | 2024-03-08 |
Family
ID=77379001
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110662992.4A Active CN113313578B (en) | 2021-06-15 | 2021-06-15 | Decision method and device for credit card consumption stage |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113313578B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105930934A (en) * | 2016-04-27 | 2016-09-07 | 北京物思创想科技有限公司 | Prediction model demonstration method and device and prediction model adjustment method and device |
| CN107346515A (en) * | 2017-06-01 | 2017-11-14 | 银联智策顾问（上海）有限公司 | A kind of credit card Forecasting Methodology and device by stages |
| CN108416395A (en) * | 2018-03-27 | 2018-08-17 | 东北大学 | A kind of Interactive Decision-Making tree constructing method based on attribute loop |
| CN109685635A (en) * | 2018-09-11 | 2019-04-26 | 深圳平安财富宝投资咨询有限公司 | Methods of risk assessment, air control server-side and the storage medium of financial business |
| CN109767312A (en) * | 2018-12-10 | 2019-05-17 | 江西师范大学 | A kind of training of credit evaluation model, appraisal procedure and device |
| CN110659922A (en) * | 2018-06-28 | 2020-01-07 | 马上消费金融股份有限公司 | Client screening method, device, server and computer readable storage medium |
| CN110751515A (en) * | 2019-10-16 | 2020-02-04 | 北京百分点信息科技有限公司 | Decision-making method and device based on user consumption behaviors, electronic equipment and storage medium |
| CN112329874A (en) * | 2020-11-12 | 2021-02-05 | 京东数字科技控股股份有限公司 | Data service decision method and device, electronic equipment and storage medium |
| CN113609193A (en) * | 2021-08-11 | 2021-11-05 | 中国工商银行股份有限公司 | Method and device for training prediction model for predicting customer transaction behavior |
-
2021
- 2021-06-15 CN CN202110662992.4A patent/CN113313578B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105930934A (en) * | 2016-04-27 | 2016-09-07 | 北京物思创想科技有限公司 | Prediction model demonstration method and device and prediction model adjustment method and device |
| CN108960514A (en) * | 2016-04-27 | 2018-12-07 | 第四范式（北京）技术有限公司 | It shows the method, apparatus of prediction model and adjusts the method, apparatus of prediction model |
| CN107346515A (en) * | 2017-06-01 | 2017-11-14 | 银联智策顾问（上海）有限公司 | A kind of credit card Forecasting Methodology and device by stages |
| CN108416395A (en) * | 2018-03-27 | 2018-08-17 | 东北大学 | A kind of Interactive Decision-Making tree constructing method based on attribute loop |
| CN110659922A (en) * | 2018-06-28 | 2020-01-07 | 马上消费金融股份有限公司 | Client screening method, device, server and computer readable storage medium |
| CN109685635A (en) * | 2018-09-11 | 2019-04-26 | 深圳平安财富宝投资咨询有限公司 | Methods of risk assessment, air control server-side and the storage medium of financial business |
| CN109767312A (en) * | 2018-12-10 | 2019-05-17 | 江西师范大学 | A kind of training of credit evaluation model, appraisal procedure and device |
| CN110751515A (en) * | 2019-10-16 | 2020-02-04 | 北京百分点信息科技有限公司 | Decision-making method and device based on user consumption behaviors, electronic equipment and storage medium |
| CN112329874A (en) * | 2020-11-12 | 2021-02-05 | 京东数字科技控股股份有限公司 | Data service decision method and device, electronic equipment and storage medium |
| CN113609193A (en) * | 2021-08-11 | 2021-11-05 | 中国工商银行股份有限公司 | Method and device for training prediction model for predicting customer transaction behavior |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113313578B (en) | 2024-03-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102567391B (en) | Method and device for building classification forecasting mixed model | |
| CN107944874A (en) | Air control method, apparatus and system based on transfer learning | |
| CN109300039A (en) | The method and system of intellectual product recommendation are carried out based on artificial intelligence and big data | |
| CN111967971B (en) | Bank customer data processing method and device | |
| CN111566684B (en) | Method and system for optimizing user grouping of advertisements | |
| CN105225135B (en) | Potential customer identification method and device | |
| Ustinovichius | Determination of efficiency of investments in construction | |
| CN111210072B (en) | Prediction model training and user resource limit determining method and device | |
| CN109214446A (en) | Potentiality good performance personnel kind identification method, system, terminal and computer readable storage medium | |
| CN111581516A (en) | Investment product recommendation method and related device | |
| CN110147389A (en) | Account number treating method and apparatus, storage medium and electronic device | |
| WO2017079824A1 (en) | Markov decision process-based decision support tool for financial planning, budgeting, and forecasting | |
| CN110276677A (en) | Refund prediction technique, device, equipment and storage medium based on big data platform | |
| CN109598404A (en) | Automatically to the method and apparatus for issuing the progress data processing of sales task list | |
| CN115409577A (en) | Intelligent container repurchase prediction method and system based on user behavior and environmental information | |
| CN109919675A (en) | Communication user upshift prediction probability recognition methods neural network based and system | |
| Spray | Reorganise, replace or expand? The role of the supply-chain in first-time exporting | |
| CN112836750A (en) | System resource allocation method, device and equipment | |
| CN113313578B (en) | Decision method and device for credit card consumption stage | |
| CN112508608B (en) | Popularization activity configuration method, system, computer equipment and storage medium | |
| CN115600818A (en) | Multi-dimensional scoring method and device, electronic equipment and storage medium | |
| CN108268898A (en) | A kind of electronic invoice user clustering method based on K-Means | |
| CN113191570A (en) | Fund planning recommendation method, device and equipment based on deep learning | |
| Kohansal et al. | An integrated MILP-MCDM decision framework for uncertain multi-criteria facilities location problem of glass industries | |
| CN109308565B (en) | Crowd performance grade identification method and device, storage medium and computer equipment |
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 |