CN111768247A - Order-placing rate prediction method, device and readable storage medium - Google Patents
Order-placing rate prediction method, device and readable storage medium Download PDFInfo
- Publication number
- CN111768247A CN111768247A CN202010625584.7A CN202010625584A CN111768247A CN 111768247 A CN111768247 A CN 111768247A CN 202010625584 A CN202010625584 A CN 202010625584A CN 111768247 A CN111768247 A CN 111768247A
- Authority
- CN
- China
- Prior art keywords
- user behavior
- behavior data
- target
- rate
- data
- 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.)
- Pending
Links
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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- 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
-
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a method, equipment and a readable storage medium for predicting a drop rate, wherein the method for predicting the drop rate comprises the following steps: the method comprises the steps of obtaining user behavior data, determining a user behavior feature combination corresponding to the user behavior data, dynamically matching a first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set, and predicting the order placing rate of a target user corresponding to the user behavior data based on the user behavior data and the first target prediction model to obtain a first predicted order placing rate. The method and the device solve the technical problem of low accuracy of the order placing rate prediction.
Description
Technical Field
The present application relates to the field of artificial intelligence, and in particular, to a method and an apparatus for predicting a placing rate, and a readable storage medium.
Background
With the continuous development of financial technologies, especially internet technology and finance, more and more technologies (such as distributed, Blockchain, artificial intelligence and the like) are applied to the financial field, but the financial industry also puts higher requirements on the technologies, such as higher requirements on the distribution of backlog of the financial industry.
With the continuous development of computer software and artificial intelligence, the application of the neural network model is more and more extensive, at present, the neural network model is commonly used for predicting the ordering rate corresponding to the user behavior, wherein, the order placing rate includes the click rate of the user to a certain article, the purchase probability and the like, however, in some specific user behavior scenes, each local user generates a series of continuous user behaviors, and the series of user behaviors generated by each local user are often different and have great difference, and the feature dimensions of the collected user behavior data are different, and currently, the order taking rate of the user is usually predicted based on a fixed neural network model, and then, the order taking rate prediction is carried out on the target users with the user behavior data with different characteristic dimensions based on the fixed neural network model, so that the accuracy of the order taking rate prediction is low.
Disclosure of Invention
The present application mainly aims to provide a method, a device and a readable storage medium for predicting a drop rate, and aims to solve the technical problem of low accuracy of the drop rate prediction in the prior art.
In order to achieve the above object, the present application provides a method for predicting a lower order rate, where the method for predicting a lower order rate is applied to a device for predicting a lower order rate, and the method for predicting a lower order rate includes:
acquiring user behavior data and determining a user behavior characteristic combination corresponding to the user behavior data;
dynamically matching a first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set;
and predicting the order placing rate of the target user corresponding to the user behavior data based on the user behavior data and the first target prediction model to obtain a first predicted order placing rate.
Optionally, the step of determining a user behavior feature combination corresponding to the user behavior data includes:
determining a user behavior path diagram corresponding to the user behavior data, and acquiring each user behavior node corresponding to the user behavior path diagram;
and determining the user behavior characteristic combination based on the user behavior characteristics corresponding to the user behavior nodes.
Optionally, after the step of performing order rate prediction on the target user corresponding to the user behavior data based on the user behavior data and the first target prediction model to obtain a first predicted order rate, the order rate prediction method further includes:
acquiring newly-added user behavior data, and determining second user behavior data based on the newly-added user behavior data and the user behavior data;
and dynamically matching a second target prediction model corresponding to the second user behavior data in the preset model set, and predicting the order placing rate of the target user based on the second user behavior data and the second target prediction model to obtain a second predicted order placing rate.
Optionally, the preset model set comprises at least one target model,
before the step of dynamically matching the first target prediction model corresponding to the user behavior data based on the user behavior feature combination and the preset model set, the order prediction method includes:
acquiring each preset feature combination and a model to be trained, and determining a training sample set corresponding to each preset feature combination;
and respectively carrying out iterative training on the model to be trained on the basis of each training sample set to obtain the target model corresponding to each preset feature combination so as to obtain the preset model set.
Optionally, the step of determining a training sample set corresponding to each preset feature combination includes:
acquiring local user behavior data, and extracting feature combination representation data corresponding to each preset feature combination from the local user behavior data on the basis of each target feature corresponding to each preset feature combination;
and acquiring a sample label set, and determining each training sample set based on the sample label set and each feature combination representation data.
Optionally, the user behavior data includes at least one stage of user behavior data, and one stage of user behavior data corresponds to one user behavior stage,
after the step of obtaining the user behavior data, the order prediction method further includes:
determining a stage user behavior characteristic combination corresponding to each stage user behavior data;
dynamically matching the stage prediction models corresponding to the stage user behavior data respectively based on the stage user behavior feature combination and a preset stage prediction model set;
and respectively carrying out order rate prediction on the target user corresponding to the user behavior data in each user behavior stage based on each stage prediction model and each stage user behavior data to obtain the order rate of the target user in each stage of the user behavior stage.
The present application further provides a unit rate prediction apparatus, the unit rate prediction apparatus is a virtual apparatus, and the unit rate prediction apparatus is applied to a unit rate prediction device, and the unit rate prediction apparatus includes:
the first determining module is used for acquiring user behavior data and determining a user behavior characteristic combination corresponding to the user behavior data;
the second determining module is used for dynamically matching a first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set;
and the first prediction module is used for predicting the order placing rate of the target user corresponding to the user behavior data based on the user behavior data and the first target prediction model to obtain a first predicted order placing rate.
Optionally, the first prediction module comprises:
a first extraction unit, configured to extract user behavior feature data from the user behavior data based on the user behavior feature combination;
and the prediction unit is used for inputting the user behavior characteristic data into the first target prediction model, classifying the user behavior characteristic data, and predicting the order placing rate of the target user to obtain the first predicted order placing rate.
Optionally, the first determining module includes:
the acquisition unit is used for determining a user behavior path diagram corresponding to the user behavior data and acquiring each user behavior node corresponding to the user behavior path diagram;
and the first determining unit is used for determining the user behavior characteristic combination based on the user behavior characteristics corresponding to the user behavior nodes.
Optionally, the second determining module includes:
the second determining unit is used for determining each preset feature combination corresponding to the preset model set and determining a first target feature combination corresponding to the user behavior feature combination in each preset feature combination;
a third determining unit, configured to determine the first target prediction model in the preset model set based on the first target feature combination.
Optionally, the unit rate prediction apparatus further includes:
the third determining module is used for acquiring newly-added user behavior data and determining second user behavior data based on the newly-added user behavior data and the user behavior data;
and the second prediction module is used for dynamically matching a second target prediction model corresponding to the second user behavior data in the preset model set, and predicting the order placing rate of the target user based on the second user behavior data and the second target prediction model to obtain a second predicted order placing rate.
Optionally, the unit rate prediction apparatus further includes:
the fourth determining module is used for acquiring each preset feature combination and the model to be trained and determining a training sample set corresponding to each preset feature combination;
and the iterative training module is used for respectively carrying out iterative training on the model to be trained on the basis of each training sample set to obtain the target model corresponding to each preset feature combination so as to obtain the preset model set.
Optionally, the fourth determining module includes:
the second extraction unit is used for acquiring local user behavior data and extracting feature combination representation data corresponding to each preset feature combination from each local user behavior data on the basis of each target feature corresponding to each preset feature combination;
and the fourth determining unit is used for acquiring a sample label set and determining each training sample set based on the sample label set and each feature combination representation data.
Optionally, the unit rate prediction apparatus further includes:
a fifth determining module, configured to determine a combination of stage user behavior characteristics corresponding to each stage user behavior data;
the sixth determining module is used for respectively and dynamically matching the stage prediction models corresponding to the stage user behavior data based on the stage user behavior feature combination and the preset stage prediction model set;
and the third prediction module is used for performing order rate prediction on the target user corresponding to the user behavior data in each user behavior phase respectively based on each phase prediction model and each phase user behavior data to obtain the order rate of the target user in each user behavior phase.
The present application further provides a device for predicting a drop rate, where the device for predicting a drop rate is an entity device, and the device for predicting a drop rate includes: a memory, a processor, and a program of the drop rate prediction method stored on the memory and executable on the processor, the program of the drop rate prediction method when executed by the processor implementing the steps of the drop rate prediction method as described above.
The present application also provides a readable storage medium having stored thereon a program for implementing a method of order placement prediction, which when executed by a processor implements the steps of the method as described above.
Compared with the technical means of predicting the order rate of target users with user behavior data of different characteristic dimensions by adopting a fixed neural network model based on the prior art, the method and the device have the advantages that the user behavior feature combination corresponding to the user behavior data is analyzed in advance, the first target prediction model matched with the characteristic dimension of the user behavior data is dynamically matched based on the user behavior feature combination, the user behavior data and the prediction model are consistent in the data characteristic dimension, the purpose of predicting the order rate based on the prediction model and the user behavior data matched with the characteristic dimension is realized, and the problems that when the characteristic dimension of the user behavior data of each local user is different in the prior art, namely, the characteristic dimension of each user behavior data is in dynamic change are overcome, the order placing rate prediction is carried out on a series of different user behavior data based on a fixed neural network model, so that the technical defect that the order placing rate prediction accuracy is low is caused, and the order placing rate prediction accuracy is improved, so that the technical problem that the order placing rate prediction accuracy is low is solved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and together with the description, serve to explain the principles of the application.
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a schematic flow chart of a first embodiment of a single rate prediction method according to the present application;
FIG. 2 is a schematic flow chart of the user behavior path diagram in the single rate prediction method of the present application;
FIG. 3 is a flowchart illustrating a second embodiment of a single rate prediction method according to the present application;
fig. 4 is a schematic diagram of a matrix representation form corresponding to the local user behavior data in the single rate prediction method of the present application;
FIG. 5 is a flowchart illustrating a third embodiment of a single rate prediction method according to the present application
Fig. 6 is a schematic device structure diagram of a hardware operating environment according to an embodiment of the present application.
The objectives, features, and advantages of the present application will be further described with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In a first embodiment of the present invention, with reference to fig. 1, a method for predicting a lower order rate includes:
step S10, acquiring user behavior data, and determining a user behavior feature combination corresponding to the user behavior data;
in this embodiment, it should be noted that, on the preset time line, each local user in the lower single-rate prediction device generates a series of user behaviors, and the series of user behaviors corresponding to each local user are different, so that the data feature dimension of each local user is different, and further, for all local users, the data feature dimension corresponding to the user behavior data is in dynamic change, for example, assuming that the user a generates the user behavior a and the user behavior B, the user B generates the user behavior a, the user behavior B, and the user behavior c, and the user behavior feature corresponding to the user behavior a is x, the user behavior feature corresponding to the user behavior B is y, and the user behavior feature corresponding to the user behavior c is z, the data feature dimension corresponding to the user behavior data at the user a is 2, and the data feature dimension corresponding to the user behavior data at the user B is 3, and then the data characteristic dimensions corresponding to the user A and the user B are different, that is, for all the acquired user behavior data, the data characteristic dimensions are dynamically changed between 2 and 3.
It should be noted that the user behavior feature combination is a feature combination corresponding to each user behavior feature corresponding to the user behavior data, and the user behavior data is associated data of the current user behavior, where the current user behavior is a user behavior generated at the current time point, for example, if the user a consults the article a with a WeChat before the current time point and the current time point, and then purchases the associated article B of the article a, and the like, the current user behavior is respectively WeChat with the associated article B, and the user behavior data is content of WeChat communication and association degree of the associated article B with the article a, and the like.
The method comprises the steps of obtaining user behavior data, determining a user behavior feature combination corresponding to the user behavior data, specifically obtaining the user behavior data, generating a user behavior path diagram corresponding to the user behavior data according to the time sequence of occurrence of each user behavior corresponding to the user behavior data, and determining the user behavior feature combination corresponding to the user behavior data based on the user behavior path diagram.
Wherein the step of determining the user behavior feature combination corresponding to the user behavior data comprises:
step S11, determining a user behavior path graph corresponding to the user behavior data, and acquiring each user behavior node corresponding to the user behavior path graph;
in this embodiment, it should be noted that the user behavior path graph is a path graph that records each user behavior and a time sequence of occurrence of each user behavior, and the user behavior path graph includes a user behavior node corresponding to each user behavior.
Determining a user behavior path diagram corresponding to the user behavior data, and acquiring each user behavior node corresponding to the user behavior path diagram, specifically, generating the user behavior path diagram corresponding to the user behavior data based on the time sequence of occurrence of each user behavior in the user behavior data, so as to cluster associated data of each user behavior in the user behavior data, wherein each data associated with the user behavior is gathered on the corresponding user behavior node, and further acquire each user behavior node corresponding to the user behavior path diagram, as shown in fig. 2, the user behavior path diagram is a schematic diagram of the user behavior path diagram, wherein an enterprise, a service, an entry, a page, an article, a question, a dialogue, a WeChat, a telephone call, a score, and a list are names of the user behavior nodes.
Step S12, determining the user behavior feature combination based on the user behavior features corresponding to the user behavior nodes.
In this embodiment, the user behavior feature combination is determined based on the user behavior features corresponding to each of the user behavior nodes, specifically, the user behavior feature combination is generated based on the feature codes of the user behavior features corresponding to each of the user behavior nodes and the time sequence of occurrence of the user behavior corresponding to each of the user behavior nodes, for example, if the sequence of occurrence of the user behavior of the user H is user behavior a, user behavior b, and user behavior C, and the feature code of the user behavior feature corresponding to the user behavior a is x, the feature code of the user behavior feature corresponding to the user behavior b is y, and the feature code of the user behavior feature corresponding to the user behavior C is z, the user behavior feature combination is a vector (x, y, z).
Step S20, dynamically matching a first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set;
in this embodiment, it should be noted that the user behavior feature combination is a combination of the user behavior features, and may be represented by a feature combination vector formed by feature codes corresponding to the user behavior features, where the preset model set at least includes one prediction model, and the preset model set is a set of prediction models trained based on different preset user behavior feature combinations, and one preset user behavior feature combination corresponds to one prediction model, for example, assuming that the preset model set includes a prediction model M and a prediction model N, the prediction model M is trained based on a user behavior feature combination (x1, x2), and the prediction model N is trained based on a user behavior feature combination (x1, x2, x3), where x1, x2, and x3 are all feature codes of user behavior features.
And dynamically matching a first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set, specifically, dynamically matching a corresponding first target prediction model in the preset model set based on a feature combination vector corresponding to the user behavior feature combination.
The step of dynamically matching a first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set comprises:
step S21, determining each preset feature combination corresponding to the preset model set, and determining a first target feature combination corresponding to the user behavior feature combination in each preset feature combination;
in this embodiment, it should be noted that the preset feature combination is a preset combination composed of different user behavior features.
Determining each preset feature combination corresponding to the preset model set, and determining a first target feature combination corresponding to the user behavior feature combination in each preset feature combination, specifically, obtaining each preset feature combination corresponding to the preset model set, and comparing each preset feature combination with the user behavior feature combination respectively, to obtain a bit coincidence rate corresponding to each preset feature combination and the user behavior feature combination, and further determining a maximum target bit coincidence rate in each bit coincidence rate, and taking the preset feature combination corresponding to the target bit coincidence rate as the first target feature combination corresponding to the user behavior feature combination, for example, assuming that the user behavior feature combination is a vector (x1, x2, x3, x4), each preset feature combination is a vector a (x1, x2, x3), vector B (x1, x2) and vector C (x1, x3), the bit matching rate corresponding to vector a is 75%, the bit matching rate corresponding to vector B is 50%, the bit matching rate corresponding to vector C is 50%, and the first target feature combination is vector a (x1, x2, x 3).
Step S22, determining the first target prediction model in the preset model set based on the first target feature combination.
In this embodiment, based on the first target feature combination, the first target prediction model is determined in the preset model set, and specifically, a prediction model corresponding to the first target feature combination in the preset model set is used as the first target prediction model.
Step S30, based on the user behavior data and the first target prediction model, performing order placement rate prediction on a target user corresponding to the user behavior data to obtain a first predicted order placement rate.
In this embodiment, based on the user behavior data and the first target prediction model, order placement rate prediction is performed on a target user corresponding to the user behavior data to obtain a first predicted order placement rate, specifically, a first target feature combination corresponding to the user behavior data is determined, target feature data corresponding to the first target feature combination is extracted, and then the target feature data is input into the first target prediction model to perform order placement rate prediction on the target user based on the target feature data to obtain a first predicted order placement rate, where the first target prediction model is a model trained and optimized based on the first target feature combination.
The step of predicting the order rate of the target user corresponding to the user behavior data based on the user behavior data and the first target prediction model to obtain a first predicted order rate includes:
step S31, extracting user behavior feature data from the user behavior data based on the user behavior feature combination;
in this embodiment, based on the user behavior feature combination, user behavior feature data is extracted from the user behavior data, specifically, the user behavior feature combination is compared with each preset feature combination corresponding to the preset model set, so as to determine a first target feature combination corresponding to the user behavior feature combination in each preset feature combination, and further, based on each target user behavior feature in the first target feature combination, user behavior feature data is extracted from the user behavior data, where the user behavior feature data is a feature value corresponding to each target user behavior feature in the first target feature combination.
Step S32, inputting the user behavior feature data into the first target prediction model, and classifying the user behavior feature data to predict the order placing rate of the target user, so as to obtain the first predicted order placing rate.
In this embodiment, the user behavior feature data may be represented by a user behavior feature representation matrix, where each column of the user behavior feature representation matrix corresponds to one target user behavior feature, for example, it is assumed that a certain column of the user behavior feature representation matrix is a vector D (a, b, c), where the vectors (a, b, c) are feature values, and if the target behavior feature corresponding to the vector D is telephone communication, the vector (a, b, c) represents voice data of telephone communication.
Inputting the user behavior feature data into the first target prediction model, classifying the user behavior feature data to perform order rate prediction on the target user to obtain the first predicted order rate, specifically, inputting the user behavior feature representation matrix into the first target prediction model, and performing data processing on the user behavior feature representation matrix, where the data processing includes convolution, pooling, full connection, and the like to classify the user behavior feature data to obtain a classification representation vector, and determining the first predicted order rate based on the classification representation vector, for example, assuming that the classification representation vector is (1, 0.9), where 1 is the identification of the target user for identifying the type of the target user, and 0.9 is the first predicted order rate.
After the step of predicting the order rate of the target user corresponding to the user behavior data based on the user behavior data and the first target prediction model to obtain a first predicted order rate, the order rate prediction method further includes:
step S40, acquiring newly added user behavior data, and determining second user behavior data based on the newly added user behavior data and the user behavior data;
in this embodiment, it should be noted that the newly added user behavior data is data corresponding to a user behavior generated by a user at a later time point of the current time point.
Acquiring newly added user behavior data, determining second user behavior data based on the newly added user behavior data and the user behavior data, specifically acquiring newly added user behavior data, and solving a union set of the newly added user behavior data and the user behavior data to acquire the second user behavior data.
Step S50, dynamically matching a second target prediction model corresponding to the second user behavior data in the preset model set, and performing order placement rate prediction on the target user based on the second user behavior data and the second target prediction model to obtain a second predicted order placement rate.
In this embodiment, a second target prediction model corresponding to the second user behavior data is dynamically matched in the preset model set, and based on the second user behavior data and the second target prediction model, an order rate prediction is performed on the target user to obtain a second predicted order rate, specifically, a second user behavior feature combination corresponding to the second user behavior data is obtained, a second target feature combination corresponding to the second user behavior feature combination is determined in each preset feature combination corresponding to the preset model set, and then the prediction model corresponding to the second target feature combination in the preset model set is used as the second target prediction model, further, based on the second target feature combination, second user behavior feature data is determined in the second user behavior data, and further based on the second user behavior feature data and the second target prediction model, and predicting the order placing rate of the target user to obtain a second predicted order placing rate, and determining a prediction model trained on the target feature combination with the same feature dimension based on the feature combination of the user behavior data even if the feature dimension of the user behavior data is in dynamic change, so that the user behavior data is always matched with the prediction model, and the accurate prediction of the order placing rate is realized.
Compared with the technical means of performing order rate prediction on target users with user behavior data of different feature dimensions by adopting a fixed neural network model in the prior art, the method provided by the embodiment dynamically matches a first target prediction model matched with the feature dimensions of the user behavior data by analyzing the user behavior feature combination corresponding to the user behavior data in advance and further based on the user behavior feature combination, so that the user behavior data and the prediction model are consistent in the data feature dimensions, the aim of predicting the order rate based on the prediction model and the user behavior data matched with the feature dimensions is fulfilled, and the problems that in the prior art, when the feature dimensions of the user behavior data of each local user are different, namely, the feature dimensions of the user behavior data are in dynamic change are overcome, the order placing rate prediction is carried out on a series of different user behavior data based on a fixed neural network model, so that the technical defect that the order placing rate prediction accuracy is low is caused, and the order placing rate prediction accuracy is improved, so that the technical problem that the order placing rate prediction accuracy is low is solved.
Further, referring to fig. 3, in another embodiment of the present application, based on the first embodiment of the present application, the preset model set at least includes a target model,
before the step of dynamically matching the first target prediction model corresponding to the user behavior data based on the user behavior feature combination and the preset model set, the order prediction method includes:
step A10, acquiring each preset feature combination and a model to be trained, and determining a training sample set corresponding to each preset feature combination;
in this embodiment, it should be noted that the model to be trained is an untrained neural network model, and one of the first target prediction models corresponds to one of the preset feature combinations.
The method comprises the steps of obtaining each preset feature combination and a model to be trained, determining a training sample set corresponding to each preset feature combination, specifically obtaining each preset feature combination and a model to be trained, and extracting the training sample set corresponding to each preset feature combination from local user data based on each user behavior feature corresponding to each preset feature combination.
Wherein, the step of determining the training sample set corresponding to each preset feature combination comprises:
step A11, acquiring local user behavior data, and extracting feature combination representation data corresponding to each preset feature combination from each local user behavior data based on each target feature corresponding to each preset feature combination;
in this embodiment, it should be noted that the preset feature combination at least includes one target feature, where the local user behavior data is user behavior data corresponding to each local user, where the local user behavior data may be represented by a matrix, as shown in fig. 4, a schematic diagram of a matrix representation form corresponding to the local user behavior data is shown, where columns x11 and x12 of the matrix and the like all correspond to user behavior features, rows user1 to user of the matrix all correspond to training samples, where one training sample corresponds to one local user, 0 and 1 are feature variable codes corresponding to the user behavior features, and are used to identify feature values of the user behavior features, for example, 1 represents that an enterprise operation range is large, 0 represents that an enterprise operation range is small, and null represents that the user does not generate the user behavior corresponding to the user behavior features, and y is a sample label corresponding to the local user, y being equal to 1 indicates that the user placed an order, and y being equal to 0 indicates that the user did not place an order.
Acquiring local user behavior data, extracting feature combination representation data corresponding to each preset feature combination from the local user behavior data based on each target feature corresponding to each preset feature combination, specifically acquiring the local user behavior data, and further executing the following steps for each preset feature combination:
determining target features corresponding to the preset feature combination, further extracting, based on the target features, target user behavior feature data corresponding to the target features from each local user behavior data, obtaining training sample data corresponding to each target feature from each local user behavior data, and further using each training sample data as the feature combination representation data, where it is to be noted that, if all features corresponding to the local user behavior data do not include all target features, the training sample corresponding to each target feature cannot be extracted, for example, it is assumed that the preset feature combination includes a feature x1 and a feature x2, a feature x1, a feature x2, and a feature x3 exist in the local user a, a feature x1 exists in the local user B, and a feature x1, a feature x2, a feature x2, and a feature x2 exist in the local user C, Feature x3 and feature x4 are extracted from user behavior data corresponding to the local user a and the local user C, respectively, feature data about the feature x1 and the feature x2 are extracted, and then training sample data (x1a, x2a) corresponding to the local user a and training sample data (x1B, x2B) corresponding to the local user B are obtained, where x1a is feature data of the feature x1 in the user behavior data corresponding to the local user a, x2a is feature data of the feature x2 in the user behavior data corresponding to the local user a, x1B is feature data of the feature x1 in the user behavior data corresponding to the local user B, and x2B is feature data of the feature x2 in the user behavior data corresponding to the local user B.
Step A12, obtaining a sample label set, and determining each training sample set based on the sample label set and each feature combination representation data.
In this embodiment, it should be noted that the sample label set is a set of sample labels corresponding to the local users.
Obtaining a sample label set, and determining each training sample set based on the sample label set and each feature combination representing data, specifically, executing the following steps corresponding to each feature combination representing data:
obtaining a sample label corresponding to each local user and training sample data corresponding to each local user in the feature combination representing data, and further generating a training sample corresponding to each local user from the training sample data corresponding to each local user and the corresponding sample label, where the training sample includes a training sample number set and a sample label, and further obtaining a training sample set corresponding to a preset feature combination corresponding to the feature combination representing data by each local user, and further obtaining each training sample set, for example, if the training sample data is a vector (a, b), and the sample label is 1, then the corresponding training sample is (a, b, 1).
Step A20, based on each training sample set, performing iterative training on the model to be trained respectively to obtain the target model corresponding to each preset feature combination, so as to obtain the preset model set.
In this embodiment, based on each training sample set, iterative training is performed on the model to be trained, so as to obtain the target model corresponding to each preset feature combination, so as to obtain the preset model set, specifically, the following steps are performed for each training sample set:
extracting a sample representation matrix of the iteration from the training sample set, wherein each row of the sample representation matrix corresponds to a target local user, each column of the sample representation matrix corresponds to a user behavior feature, the sample representation matrix is a matrix for storing the training sample data corresponding to each target local user, the sample representation matrix is further input into the model to be trained, and data processing is performed on the sample representation matrix, wherein the data processing includes convolution, pooling, full connection and the like, so as to obtain an output sample label vector, the output sample label vector includes a model classification label for classifying each training sample data by the model to be trained, and then a model training loss is calculated based on the sample label vector and a real sample label vector corresponding to each target local user, further judging whether the model training loss is converged, if the model training loss is converged, then the model to be trained satisfies a preset iteration end condition, then using the model to be trained as a target model corresponding to the training sample set, further obtaining the preset model set, if the model training loss is not converged, then the model to be trained does not satisfy the preset iteration end condition, then optimizing the model to be trained based on the model training loss, for example, optimizing the model to be trained and the like by calculating a gradient, and re-obtaining a sample representation matrix, so as to perform iterative training on the optimized model to be trained again, until the model training loss obtained by calculation is converged, obtaining the target model, further obtaining the preset model set, wherein the preset model set comprises target models corresponding to each training sample set, the preset iteration ending conditions comprise model loss convergence, maximum iteration times and the like.
In this embodiment, each preset feature combination and a model to be trained are obtained, a training sample set corresponding to each preset feature combination is determined, and then iterative training is performed on the model to be trained respectively based on each training sample set, so as to obtain the target model corresponding to each preset feature combination, so as to obtain the preset model set, that is, this embodiment provides a method for training the preset model set, that is, before training the model, each preset feature combination is obtained first, and then each preset feature combination is obtained, a training sample set corresponding to each preset feature combination is obtained in each local user behavior data, and then each corresponding target model is trained and optimized respectively based on each training sample set, so as to obtain the preset model set, and even if the feature dimension difference of each user is very large, by determining the user behavior feature combination corresponding to each user behavior data in a targeted manner, the corresponding target prediction models can be accurately matched for the user behavior feature combinations in the preset model set so as to accurately predict the user behavior results corresponding to the user behavior data, further, the condition that the order placing rate prediction is inaccurate due to the fact that the feature dimensions of the user behavior data are in dynamic changes is avoided, and the accuracy of the order placing rate prediction is improved, so that a foundation is laid for solving the technical problem that the accuracy of the order placing rate prediction is low.
Further, referring to fig. 5, based on the first and second embodiments of the present application, in another embodiment of the present application, the user behavior data at least includes a stage user behavior data, and one stage user behavior data corresponds to a user behavior stage,
after the step of obtaining the user behavior data, the order prediction method further includes:
step B10, determining the stage user behavior feature combination corresponding to each stage user behavior data;
in this embodiment, it should be noted that, on the preset time line, a series of user behavior data generated by the target user on the time line will form a user behavior sequence, where the user behavior sequence includes at least one user behavior phase, and the phase user behavior feature data is user behavior data collected in corresponding user behavior phases, for example, assuming that, on the preset time line, a user behavior a occurs to a user, a user behavior feature X is generated to generate user behavior data X0, and then a user behavior B occurs to generate user behavior feature Y, and a user behavior data Y0 is generated, then the phase user behavior feature group corresponding to the first user behavior phase is X, the user behavior data is X0, the phase user behavior feature group corresponding to the second user behavior phase is (X, Y), the generated user behavior data is (X0, y 0).
Furthermore, because the user behavior characteristics corresponding to the user behavior data in each stage are usually different, and further the data characteristic dimensions of the user behavior data in each stage are usually different, on a preset time line, the characteristic dimensions of the user behavior data corresponding to the target user are in dynamic change, and further on the basis of a fixed neural network model, the order rate of the target user in each user behavior stage is predicted on the preset time line, and since the characteristic dimensions corresponding to the stage user behavior data are not matched with the characteristic dimensions corresponding to the neural network model, the error of the order rate prediction is easily caused, and further the accuracy of the order rate prediction is low.
Determining a stage user behavior feature combination corresponding to each stage user behavior data, specifically, generating a user behavior path diagram corresponding to each stage user behavior data based on the occurrence time sequence of the stage user behavior corresponding to each stage user behavior data, wherein one stage user behavior data corresponds to a user behavior node in the user behavior path diagram, the occurrence time sequence of writing of each stage user is the connection sequence of each user behavior node, and further determining the stage user behavior feature combination corresponding to each stage user behavior data based on the user behavior path diagram.
Step B20, dynamically matching the stage prediction models corresponding to the stage user behavior data respectively based on the stage user behavior feature combinations and a preset stage prediction model set;
in this embodiment, it should be noted that the preset stage prediction model set is a set of stage prediction models trained based on stage user behavior data corresponding to different user behavior stages, where the stage prediction models are used to predict ordering rates corresponding to stage user behaviors, and user behavior stages corresponding to the stage prediction models are the same as user behavior stages corresponding to training samples used for training and optimizing the stage prediction models.
Based on each stage user behavior feature combination and a preset stage prediction model set, dynamically matching a stage prediction model corresponding to each stage user behavior data, specifically, based on a stage user behavior feature combination vector corresponding to each stage user behavior feature combination, dynamically matching a corresponding stage prediction model for each stage user behavior feature combination in the preset stage prediction model set, where the stage user behavior feature combination vector is a combination of feature codes corresponding to each stage user behavior feature in the stage user behavior feature combination, for example, assuming that the stage user behavior feature combination includes 2 stage user behavior features and the feature codes are X and Y, respectively, the stage user behavior feature combination vector is (X, Y).
Step B30, based on each stage prediction model and each stage user behavior data, performing order rate prediction on the target user corresponding to the user behavior data in each user behavior stage, respectively, to obtain the stage order rate of the target user in each user behavior stage.
In this embodiment, based on each of the phase prediction models and each of the phase user behavior data, performing order rate prediction on a target user corresponding to the user behavior data in each of the user behavior phases respectively, to obtain an order rate of the target user in each of the user behavior phases, specifically, for each of the phase user behavior data and its corresponding phase prediction model, the following steps are performed:
inputting the stage user behavior data into the stage prediction model, and performing data processing on the stage user behavior data, wherein the data processing includes convolution, pooling, full connection and the like, so as to classify the stage user behavior data, obtain a classification label corresponding to the stage user behavior data, further query a corresponding order rate based on the classification label, further obtain the order rate of the target user in the user behavior stage corresponding to the stage user behavior data, and further obtain the order rate of the target user in the stage corresponding to each user behavior stage. For example, suppose that a target user first generates a user behavior a corresponding to user behavior data a in a first user behavior stage, then generates a user behavior B corresponding to user behavior data B in a second user behavior stage, and finally generates a user behavior C corresponding to user behavior data C in a third user behavior stage, then the stage user behavior data corresponding to the first user behavior stage is a, the stage user behavior data corresponding to the second user behavior stage is (a, B), the stage user behavior data corresponding to the third user behavior stage is (a, B, C), and then after the ordering rate prediction is performed, a classification label corresponding to a is m, (a, B) a classification label corresponding to n, (a, B, C) a classification label corresponding to k, and the ordering rate corresponding to m is 50%, the ordering rate corresponding to n is 70%, the ordering rate corresponding to k is 80%, and then the order rate of the target user is 50% in the stage of the first user behavior stage, 70% in the stage of the second user behavior stage and 80% in the stage of the third user behavior stage.
Compared with the technical means of predicting the order rate of the target user in each user behavior stage by adopting a fixed neural network model based on the prior art, the method for predicting the order rate of the target user in each user behavior stage provided by the embodiment analyzes the stage user behavior feature combination corresponding to the user behavior data of each stage in advance after acquiring the user behavior data of the target user in each stage, and dynamically matches the stage prediction model corresponding to the user behavior data of each stage based on the user behavior feature combination of each stage and a preset stage prediction model set, so that the user behavior data of each stage is consistent with the stage prediction model in the data feature dimension, the stage user behavior data and the stage prediction model matched based on the feature dimension are realized, the order rate of the target user in each user behavior stage is accurately predicted, the method overcomes the technical defect that the accuracy of the order placing rate prediction is low because the characteristic dimensionality of the stage user behavior data corresponding to each user behavior stage is in dynamic change, a neural network model is not matched with the input stage user behavior data, and the accuracy of the order placing rate prediction is low.
Referring to fig. 6, fig. 6 is a schematic device structure diagram of a hardware operating environment according to an embodiment of the present application.
As shown in fig. 6, the unit rate prediction apparatus may include: a processor 1001, such as a CPU, a memory 1005, and a communication bus 1002. The communication bus 1002 is used for realizing connection communication between the processor 1001 and the memory 1005. The memory 1005 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 1005 may alternatively be a memory device separate from the processor 1001 described above.
Optionally, the order prediction device may further include a rectangular user interface, a network interface, a camera, an RF (Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, and the like. The rectangular user interface may comprise a Display screen (Display), an input sub-module such as a Keyboard (Keyboard), and the optional rectangular user interface may also comprise a standard wired interface, a wireless interface. The network interface may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface).
Those skilled in the art will appreciate that the drop rate prediction device configuration shown in fig. 6 does not constitute a limitation of the drop rate prediction device and may include more or fewer components than shown, or some components may be combined, or a different arrangement of components.
As shown in fig. 6, a memory 1005, which is a kind of computer storage medium, may include therein an operating system, a network communication module, and a program of the order prediction method. The operating system is a program that manages and controls the hardware and software resources of the order prediction device, supports the execution of the order prediction method program, and other software and/or programs. The network communication module is used for realizing communication among the components in the memory 1005 and communication with other hardware and software in the unit rate prediction method system.
In the order-dropping-rate prediction apparatus shown in fig. 6, the processor 1001 is configured to execute a program of the order-dropping-rate prediction method stored in the memory 1005, and implement the steps of the order-dropping-rate prediction method described in any one of the above.
The specific implementation of the order rate prediction device in the present application is substantially the same as the embodiments of the order rate prediction method, and is not described herein again.
The embodiment of the present application further provides a lower single rate prediction apparatus, where the lower single rate prediction apparatus is applied to a lower single rate prediction device, and the lower single rate prediction apparatus includes:
the first determining module is used for acquiring user behavior data and determining a user behavior characteristic combination corresponding to the user behavior data;
the second determining module is used for dynamically matching a first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set;
and the first prediction module is used for predicting the order placing rate of the target user corresponding to the user behavior data based on the user behavior data and the first target prediction model to obtain a first predicted order placing rate.
Optionally, the first prediction module comprises:
a first extraction unit, configured to extract user behavior feature data from the user behavior data based on the user behavior feature combination;
and the prediction unit is used for inputting the user behavior characteristic data into the first target prediction model, classifying the user behavior characteristic data, and predicting the order placing rate of the target user to obtain the first predicted order placing rate.
Optionally, the first determining module includes:
the acquisition unit is used for determining a user behavior path diagram corresponding to the user behavior data and acquiring each user behavior node corresponding to the user behavior path diagram;
and the first determining unit is used for determining the user behavior characteristic combination based on the user behavior characteristics corresponding to the user behavior nodes.
Optionally, the second determining module includes:
the second determining unit is used for determining each preset feature combination corresponding to the preset model set and determining a first target feature combination corresponding to the user behavior feature combination in each preset feature combination;
a third determining unit, configured to determine the first target prediction model in the preset model set based on the first target feature combination.
Optionally, the unit rate prediction apparatus further includes:
the third determining module is used for acquiring newly-added user behavior data and determining second user behavior data based on the newly-added user behavior data and the user behavior data;
and the second prediction module is used for dynamically matching a second target prediction model corresponding to the second user behavior data in the preset model set, and predicting the order placing rate of the target user based on the second user behavior data and the second target prediction model to obtain a second predicted order placing rate.
Optionally, the unit rate prediction apparatus further includes:
the fourth determining module is used for acquiring each preset feature combination and the model to be trained and determining a training sample set corresponding to each preset feature combination;
and the iterative training module is used for respectively carrying out iterative training on the model to be trained on the basis of each training sample set to obtain the target model corresponding to each preset feature combination so as to obtain the preset model set.
Optionally, the fourth determining module includes:
the second extraction unit is used for acquiring local user behavior data and extracting feature combination representation data corresponding to each preset feature combination from each local user behavior data on the basis of each target feature corresponding to each preset feature combination;
and the fourth determining unit is used for acquiring a sample label set and determining each training sample set based on the sample label set and each feature combination representation data.
Optionally, the unit rate prediction apparatus further includes:
a fifth determining module, configured to determine a combination of stage user behavior characteristics corresponding to each stage user behavior data;
the sixth determining module is used for respectively and dynamically matching the stage prediction models corresponding to the stage user behavior data based on the stage user behavior feature combination and the preset stage prediction model set;
and the third prediction module is used for performing order rate prediction on the target user corresponding to the user behavior data in each user behavior phase respectively based on each phase prediction model and each phase user behavior data to obtain the order rate of the target user in each user behavior phase.
The specific implementation of the order rate prediction apparatus of the present application is substantially the same as the embodiments of the order rate prediction method, and is not described herein again.
The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings, or which are directly or indirectly applied to other related technical fields, are included in the scope of the present application.
Claims (10)
1. A method for predicting a bill drop rate is characterized in that the method for predicting the bill drop rate comprises the following steps:
acquiring user behavior data and determining a user behavior characteristic combination corresponding to the user behavior data;
dynamically matching a first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set;
and predicting the order placing rate of the target user corresponding to the user behavior data based on the user behavior data and the first target prediction model to obtain a first predicted order placing rate.
2. The method for predicting the single rate as claimed in claim 1, wherein the step of dynamically matching the first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set comprises:
extracting user behavior feature data from the user behavior data based on the user behavior feature combination;
inputting the user behavior characteristic data into the first target prediction model, classifying the user behavior characteristic data to predict the order placing rate of the target user, and obtaining the first predicted order placing rate.
3. The order prediction method as claimed in claim 1, wherein the step of determining the combination of the user behavior characteristics corresponding to the user behavior data comprises:
determining a user behavior path diagram corresponding to the user behavior data, and acquiring each user behavior node corresponding to the user behavior path diagram;
and determining the user behavior characteristic combination based on the user behavior characteristics corresponding to the user behavior nodes.
4. The method for predicting the single rate as claimed in claim 1, wherein the step of dynamically matching the first target prediction model corresponding to the user behavior data based on the user behavior feature combination and a preset model set comprises:
determining each preset feature combination corresponding to the preset model set, and determining a first target feature combination corresponding to the user behavior feature combination in each preset feature combination;
determining the first target prediction model in the preset model set based on the first target feature combination.
5. The order rate prediction method of claim 1, wherein after the step of predicting the order rate of the target user corresponding to the user behavior data based on the user behavior data and the first target prediction model to obtain a first predicted order rate, the order rate prediction method further comprises:
acquiring newly-added user behavior data, and determining second user behavior data based on the newly-added user behavior data and the user behavior data;
and dynamically matching a second target prediction model corresponding to the second user behavior data in the preset model set, and predicting the order placing rate of the target user based on the second user behavior data and the second target prediction model to obtain a second predicted order placing rate.
6. The method of claim 1, wherein the predetermined set of models comprises at least one target model,
before the step of dynamically matching the first target prediction model corresponding to the user behavior data based on the user behavior feature combination and the preset model set, the order prediction method includes:
acquiring each preset feature combination and a model to be trained, and determining a training sample set corresponding to each preset feature combination;
and respectively carrying out iterative training on the model to be trained on the basis of each training sample set to obtain the target model corresponding to each preset feature combination so as to obtain the preset model set.
7. The method of claim 6, wherein the step of determining the training sample set corresponding to each of the predetermined feature combinations comprises:
acquiring local user behavior data, and extracting feature combination representation data corresponding to each preset feature combination from the local user behavior data on the basis of each target feature corresponding to each preset feature combination;
and acquiring a sample label set, and determining each training sample set based on the sample label set and each feature combination representation data.
8. The order rate prediction method as claimed in claim 1, wherein the user behavior data comprises at least one stage user behavior data, and one stage user behavior data corresponds to one user behavior stage,
after the step of obtaining the user behavior data, the order prediction method further includes:
determining a stage user behavior characteristic combination corresponding to each stage user behavior data;
dynamically matching the stage prediction models corresponding to the stage user behavior data respectively based on the stage user behavior feature combination and a preset stage prediction model set;
and respectively carrying out order rate prediction on the target user corresponding to the user behavior data in each user behavior stage based on each stage prediction model and each stage user behavior data to obtain the order rate of the target user in each stage of the user behavior stage.
9. An order-dropping-rate prediction apparatus, characterized in that the order-dropping-rate prediction apparatus comprises: a memory, a processor, and a program stored on the memory for implementing the downward single rate prediction method,
the memory is used for storing a program for realizing the single rate prediction method;
the processor is configured to execute a program implementing the order taking prediction method to implement the steps of the order taking prediction method according to any one of claims 1 to 8.
10. A readable storage medium having stored thereon a program for implementing a method of order prediction, the program being executable by a processor to implement the steps of the method of order prediction as claimed in any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010625584.7A CN111768247A (en) | 2020-06-30 | 2020-06-30 | Order-placing rate prediction method, device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010625584.7A CN111768247A (en) | 2020-06-30 | 2020-06-30 | Order-placing rate prediction method, device and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111768247A true CN111768247A (en) | 2020-10-13 |
Family
ID=72724265
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010625584.7A Pending CN111768247A (en) | 2020-06-30 | 2020-06-30 | Order-placing rate prediction method, device and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111768247A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112527852A (en) * | 2021-02-07 | 2021-03-19 | 北京淇瑀信息科技有限公司 | User dynamic support strategy allocation method and device and electronic equipment |
| CN112926690A (en) * | 2021-03-31 | 2021-06-08 | 北京奇艺世纪科技有限公司 | Data processing method, device, equipment and storage medium |
| CN113283822A (en) * | 2021-07-23 | 2021-08-20 | 支付宝(杭州)信息技术有限公司 | Feature processing method and device |
-
2020
- 2020-06-30 CN CN202010625584.7A patent/CN111768247A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112527852A (en) * | 2021-02-07 | 2021-03-19 | 北京淇瑀信息科技有限公司 | User dynamic support strategy allocation method and device and electronic equipment |
| CN112926690A (en) * | 2021-03-31 | 2021-06-08 | 北京奇艺世纪科技有限公司 | Data processing method, device, equipment and storage medium |
| CN112926690B (en) * | 2021-03-31 | 2023-09-01 | 北京奇艺世纪科技有限公司 | Data processing method, device, equipment and storage medium |
| CN113283822A (en) * | 2021-07-23 | 2021-08-20 | 支付宝(杭州)信息技术有限公司 | Feature processing method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111768247A (en) | Order-placing rate prediction method, device and readable storage medium | |
| CN111160469B (en) | Active learning method of target detection system | |
| CN109992601B (en) | To-do information pushing method and device and computer equipment | |
| CN107220217A (en) | Characteristic coefficient training method and device that logic-based is returned | |
| CN110264274B (en) | Guest group dividing method, model generating method, device, equipment and storage medium | |
| CN112801773A (en) | Enterprise risk early warning method, device, equipment and storage medium | |
| CN111783867A (en) | Machine learning algorithm selection method and device | |
| CN109872230B (en) | Test method and device of financial data analysis system, medium and electronic equipment | |
| CN113805871A (en) | Front-end code generation method and device and computer equipment | |
| CN111784401A (en) | Order taking rate prediction method, device, equipment and readable storage medium | |
| CN111768242A (en) | Order-placing rate prediction method, device and readable storage medium | |
| CN114692889A (en) | Meta-feature training model for machine learning algorithm | |
| CN113887655A (en) | Model chain regression prediction method, device, equipment and computer storage medium | |
| CN111784246B (en) | Logistics path estimation method | |
| CN112712181A (en) | Model construction optimization method, device, equipment and readable storage medium | |
| CN112633414A (en) | Feature selection optimization method, device and readable storage medium | |
| CN117370162A (en) | Test tool management method, device, equipment and storage medium | |
| CN113010687B (en) | Exercise label prediction method and device, storage medium and computer equipment | |
| CN112860736A (en) | Big data query optimization method and device and readable storage medium | |
| CN113468076A (en) | Application program exception testing method, device, equipment and storage medium | |
| US11366833B2 (en) | Augmenting project data with searchable metadata for facilitating project queries | |
| CN113656292A (en) | Multi-dimensional cross-space-time basic software performance bottleneck detection method | |
| JP6611268B2 (en) | Identification device, analysis system, identification method, and identification program | |
| CN113706040B (en) | Risk identification method, apparatus, device and storage medium | |
| CN109614328B (en) | Method and apparatus for processing test data |
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 |