CN114022196A - Advertisement putting method, device, electronic device and storage medium - Google Patents
Advertisement putting method, device, electronic device and storage medium Download PDFInfo
- Publication number
- CN114022196A CN114022196A CN202111227516.6A CN202111227516A CN114022196A CN 114022196 A CN114022196 A CN 114022196A CN 202111227516 A CN202111227516 A CN 202111227516A CN 114022196 A CN114022196 A CN 114022196A
- Authority
- CN
- China
- Prior art keywords
- advertisement
- prediction
- data
- information
- prediction 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.)
- 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
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- 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
-
- 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
Abstract
The application relates to an advertisement putting method, an advertisement putting device, an electronic device and a storage medium, wherein the advertisement putting method comprises the following steps: receiving advertisement request information, and performing data preprocessing on the advertisement request information to obtain first characteristic data; processing the first characteristic data by using an advertisement prediction model to obtain a first prediction classification label corresponding to the advertisement request information, wherein the advertisement prediction model is obtained by training according to the acquired advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the characteristic data according to the characteristic data, and the prediction classification label comprises the probability that the characteristic data corresponds to various preset advertisement categories; and determining a prediction result of the advertisement request information according to the first prediction classification label, and returning the corresponding advertisement for display by combining a preset sorting rule. Through the method and the device, the problems of low advertisement putting precision and low putting efficiency in the related technology are solved, and efficient and accurate matching of the advertisements is realized.
Description
Technical Field
The present application relates to the field of advertisement delivery technologies, and in particular, to an advertisement delivery method, an advertisement delivery apparatus, an electronic apparatus, and a storage medium.
Background
With the development of network technology and digital economy, internet advertisement placement develops rapidly. Compare traditional advertisement mode, internet advertisement puts in high click rate, high conversion rate to and carry out advertisement putting selectively to customer's demand based on big data, the audience is more accurate, and advertisement putting effect is better. When the service is normalized gradually, the service needs to be further improved, and at this time, users of different categories and advertisements need to be matched more accurately to meet better user experience, and meanwhile, the advertisement putting effect is improved.
In the conventional advertisement delivery strategy, although accurate delivery can be achieved by analyzing the user preference, advertisement request and other relevant information, the user access relevant information needs to be analyzed regularly to evaluate the advertisement delivery effect, and the advertisement delivery strategy is adjusted according to the delivery effect, so that time and labor are wasted; importantly, the advertisement needs to be operated firstly and then adjusted according to the effect, so that accurate matching can be achieved only by operating for a period of time, the user experience and the advertisement timeliness are influenced, and even the gold time for advertisement putting can be missed; and if more accurate effect evaluation is required to be obtained, and the adjusted advertisement delivery strategy can be accurately matched, the analysis fineness also needs to be further improved, and the operation is more complex and time-consuming.
At present, no effective solution is provided aiming at the problems of low advertisement putting precision and low putting efficiency in the related technology.
Disclosure of Invention
The embodiment of the application provides an advertisement putting method, an advertisement putting device, an electronic device and a storage medium, and aims to at least solve the problems that in the related technology, the advertisement putting needs low precision and the putting efficiency is low.
In a first aspect, an embodiment of the present application provides an advertisement delivery method, including: receiving advertisement request information, and performing data preprocessing on the advertisement request information to obtain first characteristic data, wherein the data preprocessing comprises: carrying out independent thermalization treatment; processing the first feature data by using an advertisement prediction model to obtain a first prediction classification label corresponding to the advertisement request information, wherein the advertisement prediction model is obtained by training according to an acquired advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the feature data according to the feature data, and the prediction classification label comprises the probability that the feature data corresponds to various preset advertisement categories; and determining a prediction result of the advertisement request information according to the first prediction classification label, and returning a corresponding advertisement for display by combining a preset sorting rule.
In some embodiments, determining a predicted outcome for the advertisement request information based on the first predicted category label includes: detecting probability values of prediction probabilities of the first feature data corresponding to various preset advertisement categories in the first prediction classification label; and selecting the preset advertisement category corresponding to the prediction probability with the maximum probability value as a prediction result.
In some of these embodiments, the method further comprises: collecting the advertisement access log; wherein the advertisement access log comprises: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information; performing data preprocessing on the advertisement access log to obtain second characteristic data; and training by using the second characteristic data to obtain the advertisement prediction model.
In some embodiments, the data preprocessing the advertisement access log to obtain second feature data includes: performing data cleaning on the advertisement access log; and carrying out independent heating processing on the data after data cleaning, and combining each preset characteristic dimension information to obtain the second characteristic data.
In some of these embodiments, the advertisement prediction model comprises one of the following prediction models: a logistic regression prediction model, a gradient lifting decision tree prediction model and a deep neural network prediction model.
In some embodiments, the advertisement prediction model includes a logistic regression prediction model, and processing the first feature data using the advertisement prediction model to obtain a first prediction classification tag corresponding to the advertisement request information includes: detecting data information corresponding to each preset feature dimension in the first feature data by using the advertisement prediction model, and calculating a weight value of the influence of the data information corresponding to each preset feature dimension on a prediction result; determining probability values of prediction probabilities of the first feature data corresponding to various preset advertisement categories based on step functions; and determining a first prediction classification label corresponding to the advertisement request information according to the probability value.
In some embodiments, after determining a prediction result of the advertisement request information according to the first prediction classification tag and returning a corresponding advertisement for display by combining a preset sorting rule, the method further includes: recording the advertisement access log; wherein the advertisement access log comprises: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information.
In a second aspect, an embodiment of the present application provides an advertisement delivery device, including:
the data processing module is used for receiving advertisement request information and carrying out data preprocessing on the advertisement request information to obtain first characteristic data, wherein the data preprocessing comprises the following steps: and (4) performing independent thermalization treatment.
And the prediction module is used for processing the first characteristic data by utilizing an advertisement prediction model to obtain a first prediction classification label corresponding to the advertisement request information, wherein the advertisement prediction model is obtained by training according to the acquired advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the characteristic data according to the characteristic data, and the prediction classification label comprises the probability that the characteristic data corresponds to various preset advertisement categories.
And the pushing module is used for determining the prediction result of the advertisement request information according to the first prediction classification label and returning the corresponding advertisement for displaying by combining a preset sorting rule.
In a third aspect, an embodiment of the present application provides an electronic apparatus, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the advertisement delivery method according to the first aspect is implemented.
In a fourth aspect, the present application provides a storage medium, on which a computer program is stored, where the program is executed by a processor to implement the advertisement delivery method according to the first aspect.
Compared with the related art, the advertisement delivery method, the advertisement delivery device, the electronic device and the storage medium provided by the embodiment of the application obtain the first characteristic data by receiving the advertisement request information and performing data preprocessing on the advertisement request information, wherein the data preprocessing comprises: carrying out independent thermalization treatment; processing the first feature data by using an advertisement prediction model to obtain a first prediction classification label corresponding to the advertisement request information, wherein the advertisement prediction model is obtained by training according to an acquired advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the feature data according to the feature data, and the prediction classification label comprises the probability that the feature data corresponds to various preset advertisement categories; and determining a prediction result of the advertisement request information according to the first prediction classification label, and returning a corresponding advertisement for display by combining a preset sorting rule. The problems of low advertisement putting precision and low putting efficiency in the related technology are solved, efficient and accurate matching of advertisements is achieved, and meanwhile user experience and advertisement putting effect are improved.
The details of one or more embodiments of the application are set forth in the accompanying drawings and the description below to provide a more thorough understanding of the application.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
fig. 1 is a block diagram of a hardware configuration of a terminal of an advertisement delivery method according to an embodiment of the present application;
FIG. 2 is a flow chart of a method of advertisement delivery according to an embodiment of the present application;
FIG. 3 is a flow diagram of the construction of an advertisement prediction model according to an embodiment of the present application;
FIG. 4 is a flow chart of a method of advertisement delivery in accordance with a preferred embodiment of the present application;
FIG. 5 is a diagram illustrating an actual scenario application of the advertisement delivery method according to the preferred embodiment of the present application;
fig. 6 is a block diagram of an advertisement delivery apparatus according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described and illustrated below with reference to the accompanying drawings and embodiments. 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. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments provided in the present application without any inventive step are within the scope of protection of the present application. Moreover, it should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints, which may vary from one implementation to another.
Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of ordinary skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments without conflict.
Unless defined otherwise, technical or scientific terms referred to herein shall have the ordinary meaning as understood by those of ordinary skill in the art to which this application belongs. Reference to "a," "an," "the," and similar words throughout this application are not to be construed as limiting in number, and may refer to the singular or the plural. The present application is directed to the use of the terms "including," "comprising," "having," and any variations thereof, which are intended to cover non-exclusive inclusions; for example, a process, method, system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or elements, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. Reference to "connected," "coupled," and the like in this application is not intended to be limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. Reference herein to "a plurality" means greater than or equal to two. "and/or" describes an association relationship of associated objects, meaning that three relationships may exist, for example, "A and/or B" may mean: a exists alone, A and B exist simultaneously, and B exists alone. Reference herein to the terms "first," "second," "third," and the like, are merely to distinguish similar objects and do not denote a particular ordering for the objects.
Before describing and explaining embodiments of the present application, a description will be given of the related art used in the present application as follows:
and (3) logistic regression: the logistic regression is an algorithm model, which is an algorithm model that drills the weight of each feature dimension on the result according to data to predict the result. Logistic regression is a set of linear regression plus a step function.
Independent heating treatment: the one-hot process is a process of converting the feature information into one-hot code representation. The one-hot code is used for representing the existence of the characteristics by 0 or 1, and the one-hot processing is a process for converting N pieces of characteristic information into the N-dimensional one-hot code.
The method provided by the embodiment can be executed in a terminal, a computer or a similar operation device. Taking an example of the operation on a terminal, fig. 1 is a hardware structure block diagram of the terminal of the advertisement delivery method according to the embodiment of the present application. As shown in fig. 1, the terminal 10 may include one or more (only one shown in fig. 1) processors 102 (the processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, and optionally may also include a transmission device 106 for communication functions and an input-output device 108. It will be understood by those skilled in the art that the structure shown in fig. 1 is only an illustration and is not intended to limit the structure of the terminal. For example, the terminal 10 may also include more or fewer components than shown in FIG. 1, or have a different configuration than shown in FIG. 1.
The memory 104 may be used to store computer programs, for example, software programs and modules of application software, such as computer programs corresponding to the advertisement delivery method in the embodiment of the present invention, and the processor 102 executes various functional applications and data processing by running the computer programs stored in the memory 104, so as to implement the method described above. The memory 104 may include high speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include memory located remotely from the processor 102, which may be connected to the terminal 10 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The transmission device 106 is used to receive or transmit data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider of the terminal 10. In one example, the transmission device 106 includes a Network adapter (NIC) that can be connected to other Network devices through a base station to communicate with the internet. In one example, the transmission device 106 may be a Radio Frequency (RF) module, which is used to communicate with the internet in a wireless manner.
The present embodiment provides an advertisement delivery method operating in the terminal, and fig. 2 is a flowchart of an advertisement delivery method according to an embodiment of the present application, and as shown in fig. 2, the flowchart includes the following steps:
step S201, receiving advertisement request information, and performing data preprocessing on the advertisement request information to obtain first characteristic data, wherein the data preprocessing comprises: and (4) performing independent thermalization treatment.
In this embodiment, each time an advertisement is displayed on a page, an application requests an advertisement from a system, and the request information includes an advertisement request, returned advertisement information, callback information of display and click, user basic information, user terminal information, user GIS information, characteristic information of the advertisement, interaction information between the user and the advertisement, and the like; therefore, the advertisement request information is divided into key characteristic dimensions, so that the advertisement prediction model is convenient to process, and the efficiency and the accuracy of prediction are improved.
Step S202, processing the first characteristic data by using an advertisement prediction model to obtain a first prediction classification label corresponding to the advertisement request information, wherein the advertisement prediction model is obtained by training according to the collected advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the characteristic data according to the characteristic data, and the prediction classification label comprises the probability that the characteristic data corresponds to various preset advertisement categories.
In the embodiment, a first prediction classification label corresponding to advertisement request information is obtained by processing first feature data by using an advertisement prediction model, wherein the advertisement prediction model is obtained by training according to an acquired advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the feature data according to the feature data, and the prediction classification label comprises probabilities of the feature data corresponding to various preset advertisement categories; the method and the device realize efficient and accurate matching of advertisement putting, and can accurately match the advertisement once advertisement request information is received because the advertisement prediction model is trained before the advertisement putting.
And step S203, determining a prediction result of the advertisement request information according to the first prediction classification label, and returning a corresponding advertisement to display by combining a preset sorting rule.
In the foregoing steps S201 to S203, the first feature data is obtained by receiving the advertisement request information and performing data preprocessing on the advertisement request information, where the data preprocessing includes: carrying out independent thermalization treatment; processing the first characteristic data by using an advertisement prediction model to obtain a first prediction classification label corresponding to the advertisement request information, wherein the advertisement prediction model is obtained by training according to the acquired advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the characteristic data according to the characteristic data, and the prediction classification label comprises the probability that the characteristic data corresponds to various preset advertisement categories; and determining a prediction result of the advertisement request information according to the first prediction classification label, and returning the corresponding advertisement for display by combining a preset sorting rule. The problems of low advertisement putting precision and low putting efficiency in the related technology are solved, efficient and accurate matching of advertisements is achieved, and meanwhile user experience and advertisement putting effect are improved.
It should be noted that, in this embodiment, the advertisement prediction model is constructed by using a logistic regression algorithm-based model, and logistic regression itself is supervised learning, so that the prediction model can solve the user of what type, in what space (geographic coordinates), in what time period, and for what kind of advertisements, so as to improve the performance of the model and improve the matching accuracy of advertisement delivery.
In some embodiments, determining a predicted outcome for the advertisement request information based on the first predicted category label includes:
step 1, detecting probability values of prediction probabilities of the first feature data corresponding to various preset advertisement categories in the first prediction classification labels.
And 2, selecting a preset advertisement category corresponding to the prediction probability with the maximum probability value as a prediction result.
Detecting probability values of the prediction probabilities of the first feature data corresponding to various preset advertisement categories in the first prediction classification label in the steps; the preset advertisement category corresponding to the prediction probability with the maximum probability value is selected as the prediction result, matching of the preset feature dimension and the advertisement category feature with the highest adaptability is achieved, the prediction accuracy is improved, and then the advertisement putting accuracy can be improved.
In some of these embodiments, the method further comprises the steps of:
step 1, collecting an advertisement access log; wherein the advertisement access log comprises: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information.
And 2, carrying out data preprocessing on the advertisement access log to obtain second characteristic data.
And 3, training by utilizing the second characteristic data to obtain an advertisement prediction model.
Acquiring an advertisement access log in the steps; wherein the advertisement access log comprises: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information; carrying out data preprocessing on the advertisement access log to obtain second characteristic data; the advertisement prediction model is obtained by training the second characteristic data, the training of the advertisement prediction model is realized, the logistic regression is supervised learning, and the training data is the advertisement access log, so that the advertisement prediction model can solve the user of which type, under which space (geographic coordinates) and in which time period, the good sensitivity of the advertisement of which type is achieved, the matching accuracy of advertisement putting is improved, and meanwhile, the model is trained in advance, so that the embodiment can obtain better user experience at the beginning of use.
In some embodiments, the data preprocessing the advertisement access log to obtain the second feature data includes:
and step 1, carrying out data cleaning on the advertisement access log.
And 2, performing independent heating processing on the data after data cleaning, and combining all preset feature dimension information to obtain second feature data.
Performing data cleaning on the advertisement access log in the steps; carrying out independent thermalization processing on the data after data cleaning, and combining each preset characteristic dimension information to obtain second characteristic data; the method and the device realize the splitting of the advertisement access log into key characteristic dimensions and are also beneficial to improving the performance of the advertisement prediction model.
In some of these embodiments, the advertisement prediction model comprises one of the following prediction models: a logistic regression prediction model, a gradient lifting decision tree prediction model and a deep neural network prediction model.
In the advertisement prediction model in this embodiment, not only a logistic regression algorithm but also a Gradient Boosting Decision Tree (GBDT), a deep neural network, and the like may be used for the selection of the algorithm, so that a better effect can be achieved.
In some embodiments, the advertisement prediction model includes a logistic regression prediction model, and the advertisement prediction model is used to process the first feature data to obtain a first prediction classification tag corresponding to the advertisement request information, including the following steps:
step 1, detecting data information corresponding to each preset feature dimension in the first feature data by using an advertisement prediction model, and calculating a weight value of the influence of the data information corresponding to each preset feature dimension on a prediction result.
And 2, determining the probability values of the prediction probabilities of the first characteristic data corresponding to various preset advertisement categories based on the step function.
And step 3, determining a first prediction classification label corresponding to the advertisement request information according to the probability value.
Detecting data information corresponding to each preset characteristic dimension in the first characteristic data by using an advertisement prediction model in the steps, and calculating a weight value of the influence of the data information corresponding to each preset characteristic dimension on a prediction result; obtaining the probability values of the prediction probabilities of the first characteristic data corresponding to various preset advertisement categories by combining the step functions; determining a first prediction classification label corresponding to the advertisement request information according to the probability value; and the accurate matching of the advertisements is realized.
In some embodiments, after determining a prediction result of the advertisement request information according to the first prediction classification tag and returning a corresponding advertisement for display by combining with a preset sorting rule, the method further includes:
recording an advertisement access log; wherein the advertisement access log comprises: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information.
By recording the advertisement access log after each advertisement display, the adjustment and optimization of the advertisement prediction model are facilitated.
A method for constructing and training a neural network model (advertisement prediction model) will be described below, and fig. 3 is a flow chart of constructing an advertisement prediction model according to an embodiment of the present application, and as shown in fig. 3, the method includes the following steps:
step S301, collecting advertisement access logs. Usually, in a proper position in the application, a pit and a buried point are provided as points, and when a user uses the application, each time an advertisement on a page is displayed, the application can initiate an advertisement request message to the system. The ad access log may be synchronized to a computing platform (Hadoop or spark) through a log synchronization tool (flash or Logitail).
Step S302, an advertisement prediction model is constructed.
Step S303, carrying out data preprocessing on the advertisement access log to obtain second characteristic data serving as a data set, and randomly dividing the data set into a training set and a verification set according to a preset proportion. And firstly, cleaning the data of the advertisement access log, carrying out independent heating processing on the data after the data cleaning, and combining each preset characteristic dimension information to obtain second characteristic data.
And step S304, inputting the data sets into the constructed advertisement prediction model in batches, and training to obtain the trained advertisement prediction model. The logistic regression is a set of linear regression and a step function, so that the weight values for simulating the influence of each feature dimension on the result are mainly calculated through training data through a formula, and then the probability value of the occurrence of the result is calculated through the step function:
z＝ω0+ω1x1+ω2x2+…+ωnxn
where e represents the base of the natural logarithm, ω0、ω1、ω2……ωnRepresenting individual estimated values, x1、x2……xnRepresenting the respective data, and z represents the weight value corresponding to a set of estimated values.
And obtaining a trained advertisement prediction model after training.
The embodiments of the present application are described and illustrated below by means of preferred embodiments.
Fig. 4 is a flowchart of an advertisement delivery method according to a preferred embodiment of the present application. As shown in fig. 4, the advertisement delivery method includes the following steps:
step S401, collecting advertisement access logs.
In this embodiment, the advertisement access log includes: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information. Usually, in a proper position in the application, a pit and a buried point are provided as points, and when a user uses the application, each time an advertisement on a page is displayed, the application can initiate an advertisement request message to the system. The ad access log may be synchronized to a computing platform (Hadoop or spark) through a log synchronization tool (flash or Logitail).
Step S402, constructing an advertisement prediction model based on a logistic regression algorithm.
Step S403, carrying out data preprocessing on the advertisement access log to obtain second characteristic data serving as a data set, and randomly dividing the data set into a training set and a verification set according to a ratio of 7: 3. And firstly, cleaning the data of the advertisement access log, carrying out independent heating processing on the data after the data cleaning, and combining each preset characteristic dimension information to obtain second characteristic data.
And S404, inputting the data sets into the constructed advertisement prediction model in batches, and training to obtain the trained advertisement prediction model. The logistic regression is a set of linear regression and a step function, so that the weight values for simulating the influence of each feature dimension on the result are mainly calculated through training data through a formula, and then the probability value of the occurrence of the result is calculated through the step function:
z＝ω0+ω1x1+ω2x2+…+ωnxn
and obtaining a trained advertisement prediction model after training.
Step S405, receiving advertisement request information, and performing data preprocessing on the advertisement request information to obtain first characteristic data.
Step S406, the first characteristic data is processed by using the advertisement prediction model, and a first prediction classification label corresponding to the advertisement request information is obtained.
In step S404, a trained advertisement prediction model is obtained, and therefore, when new advertisement request information comes in, the information attached to the request needs to be split into preset feature dimensions, then the advertisement classification features with the highest preset goodness are found through the feature dimensions, then the advertisement classification features are assembled into entries, then the prediction is performed by the advertisement prediction model, the most fit classification is found according to the predicted value in a descending order, and then the most fit advertisement is taken according to the sorting rule and returned to the front end for display.
Step S407, determining a prediction result of the advertisement request information according to the first prediction classification label, and returning a corresponding advertisement to display by combining a preset sorting rule. The sequencing rule can be configured through a system, and operators can adjust the final display content at any time according to requirements.
Step S408, recording an advertisement access log. The recorded advertisement access logs can be synchronized to the advertisement prediction model, and the advertisement prediction model is further optimized.
In this embodiment, the description of the actual scene application is shown in fig. 5, and mainly includes: log data, feature processing, model training, advertisement listing and advertisement recalling.
Log data: the advertisement data is divided into advertisement basic data and is written into a log when the advertisement returns; the advertisement callback data is behavior data, and the interaction actions performed by the user when seeing the advertisement are divided into a display behavior and a click behavior. And the front end can call back the service when the action occurs, and the back end writes the log for keeping the file after finishing the service.
Characteristic processing: logistic regression itself is supervised learning, enabling the established advertisement prediction model to solve what types of users, under what space (geographical coordinates), what time periods, how sensitive to what categories of advertisements are. Therefore, behaviors and basic information need to be decomposed and refined from the advertisement access log according to preset feature dimensions, and all advertisement access log information needs to be converted into feature files and stored in Hadoop for subsequent training.
Model training: in the actual process, the obtained feature file is required to be used for training, the existing model is loaded for continuous training, then the training effect of the advertisement prediction model is verified by using the verification set, and when the predicted error rate is within the preset value, the advertisement prediction model is cached. And subsequently, storing the adaptive relation of the user, the space, the time and the advertisement into a cache through a task model.
Putting advertisements on shelves: when an advertiser creates an advertisement on a DSP (demand side platform), after the advertisement is written into a mysql library, the advertisement is monitored by canal (canal translates to a water channel/pipeline/ditch, and the main purpose is based on mysql database increment log analysis), the canal sends a message to inform a task of putting the advertisement on shelf, and when the task obtains the demand of putting the advertisement on shelf, the advertisement is put on shelf in an ES advertisement library. In fig. 5, HDFS denotes a Hadoop distributed file system, API denotes an application program interface, and CRM denotes customer relationship management.
And (3) advertising recall: when the device of the user is on a page with advertisements, the application can initiate an advertisement filling requirement, the user can be decomposed into tags, the advertisement candidate set is extracted from the advertisement library according to the space geography, time and the like of the user at that time, and then appropriate advertisements are selected and displayed to the user.
It should be noted that the steps illustrated in the above-described flow diagrams or in the flow diagrams of the figures may be performed in a computer system, such as a set of computer-executable instructions, and that, although a logical order is illustrated in the flow diagrams, in some cases, the steps illustrated or described may be performed in an order different than here. For example, steps S401 and S402, and steps S401 and S405.
The embodiment further provides an advertisement delivery device, which is used for implementing the above embodiments and preferred embodiments, and the description of the device is omitted. As used hereinafter, the terms "module," "unit," "subunit," and the like may implement a combination of software and/or hardware for a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
Fig. 6 is a block diagram of an advertisement delivery apparatus according to an embodiment of the present application, and as shown in fig. 6, the apparatus includes: a data processing module 61, a prediction module 62 and a push module 63.
The data processing module 61 is configured to receive the advertisement request information, and perform data preprocessing on the advertisement request information to obtain first feature data, where the data preprocessing includes: and (4) performing independent thermalization treatment.
And the prediction module 62 and the data processing module 61 are coupled and configured to process the first feature data by using an advertisement prediction model, and obtain a first prediction classification label corresponding to the advertisement request information, where the advertisement prediction model is obtained by training according to the collected advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the feature data according to the feature data, and the prediction classification label includes probabilities that the feature data corresponds to various preset advertisement categories.
And the pushing module 63 is coupled to the predicting module 62, and is configured to determine a prediction result of the advertisement request information according to the first prediction classification tag, and return a corresponding advertisement to display by combining with a preset ordering rule.
In some embodiments, the pushing module 63 is configured to detect, in the first prediction classification tag, probability values of prediction probabilities that the first feature data correspond to various preset advertisement categories; and selecting a preset advertisement category corresponding to the prediction probability with the maximum probability value as a prediction result.
In some of these embodiments, the data processing module 61 is configured to collect advertisement access logs; wherein the advertisement access log comprises: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information; carrying out data preprocessing on the advertisement access log to obtain second characteristic data; and training by using the second characteristic data to obtain an advertisement prediction model.
In some embodiments, the data processing module 61 is configured to perform data cleansing on the advertisement access log; and carrying out independent heating processing on the data after data cleaning, and combining each preset characteristic dimension information to obtain second characteristic data.
In some embodiments, the prediction module 62 is configured to detect data information corresponding to each preset feature dimension in the first feature data by using an advertisement prediction model, and calculate a weight value of an influence of the data information corresponding to each preset feature dimension on the prediction result; determining probability values of the prediction probabilities of the first feature data corresponding to various preset advertisement categories based on a step function; and determining a first prediction classification label corresponding to the advertisement request information according to the probability value.
In some of these embodiments, the data processing module 61 is configured to record an advertisement access log; wherein the advertisement access log comprises: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information.
The above modules may be functional modules or program modules, and may be implemented by software or hardware. For a module implemented by hardware, the modules may be located in the same processor; or the modules can be respectively positioned in different processors in any combination.
The present embodiment also provides an electronic device comprising a memory having a computer program stored therein and a processor configured to execute the computer program to perform the steps of any of the above method embodiments.
Optionally, the electronic apparatus may further include a transmission device and an input/output device, wherein the transmission device is connected to the processor, and the input/output device is connected to the processor.
Optionally, in this embodiment, the processor may be configured to execute the following steps by a computer program:
s1, receiving the advertisement request information, and performing data preprocessing on the advertisement request information to obtain first characteristic data, wherein the data preprocessing comprises: and (4) performing independent thermalization treatment.
And S2, processing the first characteristic data by using an advertisement prediction model to obtain a first prediction classification label corresponding to the advertisement request information, wherein the advertisement prediction model is obtained by training according to the collected advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the characteristic data according to the characteristic data, and the prediction classification label comprises the probability that the characteristic data corresponds to various preset advertisement categories.
And S3, determining the prediction result of the advertisement request information according to the first prediction classification label, and returning the corresponding advertisement for display by combining with a preset sorting rule.
It should be noted that, for specific examples in this embodiment, reference may be made to examples described in the foregoing embodiments and optional implementations, and details of this embodiment are not described herein again.
In addition, in combination with the advertisement delivery method in the foregoing embodiments, the embodiments of the present application may provide a storage medium to implement. The storage medium having stored thereon a computer program; the computer program, when executed by a processor, implements any of the advertisement delivery methods in the above embodiments.
It should be understood by those skilled in the art that various features of the above-described embodiments can be combined in any combination, and for the sake of brevity, all possible combinations of features in the above-described embodiments are not described in detail, but rather, all combinations of features which are not inconsistent with each other should be construed as being within the scope of the present disclosure.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. An advertisement delivery method, the method comprising:
receiving advertisement request information, and performing data preprocessing on the advertisement request information to obtain first characteristic data, wherein the data preprocessing comprises: carrying out independent thermalization treatment;
processing the first feature data by using an advertisement prediction model to obtain a first prediction classification label corresponding to the advertisement request information, wherein the advertisement prediction model is obtained by training according to an acquired advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the feature data according to the feature data, and the prediction classification label comprises the probability that the feature data corresponds to various preset advertisement categories;
and determining a prediction result of the advertisement request information according to the first prediction classification label, and returning a corresponding advertisement for display by combining a preset sorting rule.
2. The advertisement delivery method according to claim 1, wherein determining the prediction result of the advertisement request information according to the first prediction classification tag comprises:
detecting probability values of prediction probabilities of the first feature data corresponding to various preset advertisement categories in the first prediction classification label;
and selecting the preset advertisement category corresponding to the prediction probability with the maximum probability value as a prediction result.
3. The method of advertisement delivery according to claim 1, further comprising:
collecting the advertisement access log; wherein the advertisement access log comprises: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information;
performing data preprocessing on the advertisement access log to obtain second characteristic data;
and training by using the second characteristic data to obtain the advertisement prediction model.
4. The advertisement delivery method according to claim 3, wherein the data preprocessing is performed on the advertisement access log to obtain second feature data, and the method comprises:
performing data cleaning on the advertisement access log;
and carrying out independent heating processing on the data after data cleaning, and combining each preset characteristic dimension information to obtain the second characteristic data.
5. The method of claim 1, wherein the advertisement prediction model comprises one of the following prediction models: a logistic regression prediction model, a gradient lifting decision tree prediction model and a deep neural network prediction model.
6. The advertisement delivery method of claim 5, wherein the advertisement prediction model comprises a logistic regression prediction model, and the obtaining the first prediction classification label corresponding to the advertisement request information by processing the first feature data with the advertisement prediction model comprises:
detecting data information corresponding to each preset feature dimension in the first feature data by using the advertisement prediction model, and calculating a weight value of the influence of the data information corresponding to each preset feature dimension on a prediction result;
determining probability values of prediction probabilities of the first feature data corresponding to various preset advertisement categories based on step functions;
and determining a first prediction classification label corresponding to the advertisement request information according to the probability value.
7. The advertisement delivery method according to claim 1, wherein after determining the prediction result of the advertisement request information according to the first prediction classification tag and returning the corresponding advertisement to be displayed in combination with a preset sorting rule, the method further comprises:
recording the advertisement access log; wherein the advertisement access log comprises: advertisement request and response information, display and callback information, user related information, terminal equipment related information, geographic position GIS information and user and advertisement interaction information.
8. An advertisement delivery device, comprising:
the data processing module is used for receiving advertisement request information and carrying out data preprocessing on the advertisement request information to obtain first characteristic data, wherein the data preprocessing comprises the following steps: carrying out independent thermalization treatment;
the prediction module is used for processing the first characteristic data by utilizing an advertisement prediction model to obtain a first prediction classification label corresponding to the advertisement request information, wherein the advertisement prediction model is obtained by training according to an acquired advertisement access log, the advertisement prediction model is trained to obtain a prediction classification label corresponding to the characteristic data according to the characteristic data, and the prediction classification label comprises the probability that the characteristic data corresponds to various preset advertisement categories;
and the pushing module is used for determining the prediction result of the advertisement request information according to the first prediction classification label and returning the corresponding advertisement for displaying by combining a preset sorting rule.
9. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, and wherein the processor is configured to execute the computer program to perform the advertisement delivery method of any one of claims 1 to 7.
10. A storage medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the advertisement delivery method of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111227516.6A CN114022196A (en) | 2021-10-21 | 2021-10-21 | Advertisement putting method, device, electronic device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111227516.6A CN114022196A (en) | 2021-10-21 | 2021-10-21 | Advertisement putting method, device, electronic device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114022196A true CN114022196A (en) | 2022-02-08 |
Family
ID=80057049
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111227516.6A Pending CN114022196A (en) | 2021-10-21 | 2021-10-21 | Advertisement putting method, device, electronic device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114022196A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114782106A (en) * | 2022-05-07 | 2022-07-22 | 南京欣威视通信息科技股份有限公司 | Elevator advertisement system for realizing accurate delivery of advertisement based on scene analysis |
| CN114826715A (en) * | 2022-04-15 | 2022-07-29 | 咪咕文化科技有限公司 | Network protection method, device, equipment and storage medium |
-
2021
- 2021-10-21 CN CN202111227516.6A patent/CN114022196A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114826715A (en) * | 2022-04-15 | 2022-07-29 | 咪咕文化科技有限公司 | Network protection method, device, equipment and storage medium |
| CN114826715B (en) * | 2022-04-15 | 2024-03-22 | 咪咕文化科技有限公司 | Network protection method, device, equipment and storage medium |
| CN114782106A (en) * | 2022-05-07 | 2022-07-22 | 南京欣威视通信息科技股份有限公司 | Elevator advertisement system for realizing accurate delivery of advertisement based on scene analysis |
| CN114782106B (en) * | 2022-05-07 | 2023-09-26 | 南京欣威视通信息科技股份有限公司 | Elevator advertisement system for realizing accurate advertisement delivery based on scene analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110198310B (en) | Network behavior anti-cheating method and device and storage medium | |
| CN107102941B (en) | Test case generation method and device | |
| CN108280115B (en) | Method and device for identifying user relationship | |
| CN104808892A (en) | Application icon ordering method, device, system and associated equipment | |
| CN107426328B (en) | Information pushing method and device | |
| CN105160545B (en) | Method and device for determining release information style | |
| CN111538907B (en) | Object recommendation method, system and device | |
| CN114022196A (en) | Advertisement putting method, device, electronic device and storage medium | |
| CN105187237A (en) | Method and device for searching associated user identifications | |
| US11257019B2 (en) | Method and system for search provider selection based on performance scores with respect to each search query | |
| CN106021455A (en) | Image characteristic relationship matching method, apparatus and system | |
| CN111148018B (en) | Method and device for identifying and positioning regional value based on communication data | |
| CN110363427A (en) | Model quality evaluation method and apparatus | |
| CN109446431A (en) | For the method, apparatus of information recommendation, medium and calculate equipment | |
| CN105550248A (en) | User information mining method and apparatus | |
| CN110209921B (en) | Method and device for pushing media resource, storage medium and electronic device | |
| CN108737486A (en) | Information-pushing method and device | |
| CN111191133B (en) | Service search processing method, device and equipment | |
| CN111026969B (en) | Content recommendation method and device, storage medium and server | |
| CN110675179A (en) | Marketing information processing method and device, electronic equipment and readable storage medium | |
| CN111552835A (en) | File recommendation method and device and server | |
| KR101028810B1 (en) | Apparatus and method for analyzing advertisement target | |
| CN110704754A (en) | Push model optimization method and device executed by user terminal | |
| CN114020960A (en) | Music recommendation method, device, server and storage medium | |
| CN109087124A (en) | A kind of application program Value Prediction Methods and device |
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 |