CN113570422A

CN113570422A - Creative guide information generation method and device, computer equipment and storage medium

Info

Publication number: CN113570422A
Application number: CN202111125599.8A
Authority: CN
Inventors: 刘凯; 杨秀金
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2021-09-26
Filing date: 2021-09-26
Publication date: 2021-10-29
Anticipated expiration: 2041-09-26
Also published as: CN113570422B

Abstract

The application relates to a method and a device for generating creative guidance information, computer equipment and a storage medium. The method comprises the following steps: acquiring a pre-delivered creative material and a delivery dimension of the creative material; the creative material is multimedia information displayed to a user; based on the delivery dimension, inquiring similar creative materials matched with the creative materials; acquiring a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result; and generating creative guidance information which contains the similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generation mode. By adopting the method, the generation efficiency of the creative guidance information can be effectively improved.

Description

Creative guide information generation method and device, computer equipment and storage medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for generating creative guidance information, a computer device, and a storage medium.

Background

With the development of computer technology and internet technology, more and more creative promotion tasks need to be put in order to help users to quickly screen out product information meeting the needs of the users from various creative promotion tasks. For example, the creative promotion task may be an advertisement, and when a merchant puts the advertisement, it is usually necessary to perform a putting experiment in different user accounts in advance to determine a putting creative and a putting plan.

However, in the current generation method of creative guidance information, it is usually necessary to know each dimension data related to the creative promotion task to be delivered in advance or to perform a delivery experiment to determine how to deliver the creative promotion task.

Disclosure of Invention

In view of the above, it is necessary to provide a creative guidance information generation method, a creative guidance information generation device, a computer device, and a storage medium, which can effectively improve the generation efficiency of creative guidance information.

A method of generating creative guidance information, the method comprising: acquiring a pre-delivered creative material and a delivery dimension of the creative material; the creative material is multimedia information displayed to a user; based on the delivery dimension, inquiring similar creative materials matched with the creative materials; acquiring a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result; and generating creative guidance information which contains the similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generation mode.

An apparatus for generating creative guide information, the apparatus comprising: the creative material acquisition module is used for acquiring pre-delivered creative materials and delivery dimensions of the creative materials; the creative material is multimedia information displayed to a user; the similar material query module is used for querying similar creative materials matched with the creative materials based on the release dimension; the generation mode acquisition module is used for acquiring a creative guidance information generation mode according to the delivery form and audience group type of the creative material; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result; and the generating module is used for generating creative guidance information which contains the similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generating mode.

In one embodiment, the creative material acquisition module includes: the creative material receiving unit is used for receiving the uploaded pre-posted creative material; the delivery dimension display unit is used for displaying different candidate delivery dimensions in the creative guidance information generation interface; the delivery dimension screening unit is used for responding to dimension screening operation and screening out delivery dimensions of the creative materials from candidate delivery dimensions displayed on the creative guidance information generation interface; the delivery dimension comprises at least one of delivery flow position information, delivery conversion types and target coverage crowd.

In one embodiment, the similar material query module is further configured to query a first creative material corresponding to the delivery traffic placement information, a second creative material corresponding to the delivery conversion type, and a third creative material corresponding to the target coverage group when the delivery dimensions are delivery traffic placement information, a delivery conversion type, and a target coverage group; the similar material inquiry module comprises: and the creative material merging unit is used for merging the first creative material, the second creative material and the third creative material to obtain similar creative materials matched with the creative materials.

In one embodiment, the apparatus further comprises: the release dimension acquisition module is used for acquiring release dimensions corresponding to released creative promotion tasks within a preset release time period; the creative data acquisition module is used for acquiring creative data which has an association relation with the released creative promotion task; the storage module is used for forming creative delivery data according to the delivered creative promotion tasks and the creative data; storing the creative delivery data in a creative library; the similar material query module is further used for querying similar creative materials matched with the creative materials in creative delivery data of the creative library based on the delivery dimensions.

In one embodiment, the apparatus further comprises: the delivery data query module is used for querying delivery effect data corresponding to the first creative material, the second creative material and the third creative material; the generating module is also used for generating initial creative guidance information matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generating mode; and the adding module is used for adding the similar creative materials serving as creative samples and the putting effect data into the initial creative guidance information to obtain the creative guidance information.

In one embodiment, the generating module comprises: the guidance information data generation unit is used for processing the similar creative materials by utilizing the creative guidance information generation mode to generate creative guidance information data; the creative guide information template inquiry unit is used for inquiring a corresponding creative guide information template based on the delivery form and audience group type of the creative material; and the adding unit is used for adding the creative guidance information data to the creative guidance information template to obtain initial creative guidance information.

In one embodiment, the apparatus further comprises: the historical guidance information acquisition module is used for acquiring at least one piece of relevant historical creative guidance information in a creative guidance information base based on the delivery form and audience group type of the creative material; the analysis module is used for analyzing the creative dimension, the creative index and the delivery result of at least one piece of historical creative guidance information to obtain a calculation mode of the creative dimension, the creative index and the delivery result of the historical creative guidance information; the coding module is used for coding the obtained calculation mode to obtain a data processing mode; the storage module is also used for storing the obtained data processing mode according to the configuration type and the function type; the generation module is also used for forming the creative generation mode according to the delivery form and audience group type of the creative materials in the historical creative guidance information.

In one embodiment, the apparatus further comprises: the characteristic labeling module is used for respectively performing characteristic labeling on the first creative material, the second creative material and the third creative material; the determining module is used for determining creative material intersection in the marked first creative material, the marked second creative material and the marked third creative material; and taking the intersection of the creative materials as similar creative materials matched with the creative materials.

In one embodiment, the similar material query module comprises: the characteristic vector acquisition unit is used for extracting characteristic vectors corresponding to all the frames of images in the video material; performing dimension reduction processing on the feature vector to obtain a dimension reduction feature vector of each frame of image; performing feature conversion on the dimensionality reduction feature vector to obtain a feature vector of the video material; and inquiring similar creative materials matched with the feature vectors of the video materials based on the delivery dimensions.

In an embodiment, the feature vector obtaining unit is further configured to perform dimension reduction processing on the frame rate feature vector and the definition feature vector, respectively, to obtain a dimension reduction frame rate feature vector and a dimension reduction definition feature vector; performing feature conversion processing on the dimensionality reduction frame rate feature vector and the dimensionality reduction definition feature vector to obtain a frame rate feature vector and a definition feature vector of the video material; and merging the frame rate characteristic vector and the definition characteristic vector of the video material to obtain the characteristic vector of the video material.

In one embodiment, the encoding module is further configured to encode the feature vector of the video material to obtain a video feature code corresponding to the video material; the similar material inquiry module is also used for inquiring the creative material according to the video characteristic coding to obtain the similar creative material matched with the video material.

In one embodiment, the apparatus further comprises: the characteristic extraction module is used for extracting the characteristics of stock creative materials in the creative library to obtain the characteristic vector of each stock creative material; the coding module is also used for coding the feature vectors of the stock creative materials to obtain material feature codes corresponding to the stock creative materials; the storage module is also used for storing each material characteristic code in the creative library; the characteristic code searching module is used for searching a target material characteristic code matched with the video characteristic code in the creative library; and the similar material acquisition module is used for acquiring similar creative materials corresponding to the target material feature codes in the creative library.

In one embodiment, the apparatus further comprises: the check code acquisition module is used for acquiring a check code corresponding to the video material; the similar material query module is also used for querying creative materials based on the check codes and the feature codes to obtain similar creative materials matched with the video materials.

A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program: acquiring a pre-delivered creative material and a delivery dimension of the creative material; the creative material is multimedia information displayed to a user; based on the delivery dimension, inquiring similar creative materials matched with the creative materials; acquiring a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result; and generating creative guidance information which contains the similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generation mode.

A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of: acquiring a pre-delivered creative material and a delivery dimension of the creative material; the creative material is multimedia information displayed to a user; based on the delivery dimension, inquiring similar creative materials matched with the creative materials; acquiring a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result; and generating creative guidance information which contains the similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generation mode.

The creative guidance information generation method, the creative guidance information generation device, the computer equipment and the storage medium acquire pre-delivered creative materials and delivery dimensions of the creative materials; the creative material is multimedia information displayed to the user; inquiring similar creative materials matched with the creative materials based on the delivery dimension; acquiring a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result; and generating creative guide information which contains similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guide information generation mode. Because similar creative materials matched with the creative materials can be inquired through similar identification, new edition creative guide information which is not delivered before can be automatically generated according to related delivery data of similar creatives and creative guide information generation modes corresponding to delivery forms of pre-delivered creative materials and audience groups, and delivery suggestions which are delivered before can be carried out on users. From this, the human effect of operation output intention guide information has been improved, a large amount of work that originally manual handling changed into systematic automated processing to the creation efficiency of intention guide information has effectively been improved.

Drawings

FIG. 1 is a diagram of an application environment of a method for generating creative guidance information in one embodiment;

FIG. 2 is a schematic flow diagram illustrating a method for generating creative guidance information in one embodiment;

FIG. 3 is a schematic flow diagram illustrating the steps of obtaining pre-posted creative materials and posting dimensions for creative materials in one embodiment;

FIG. 4 is a schematic diagram of an interface for a user to upload creative material and to filter impression dimensions in one embodiment;

FIG. 5 is a flowchart illustrating the step of merging the first creative material, the second creative material, and the third creative material to obtain similar creative materials matching the creative materials in one embodiment;

FIG. 6 is a diagram illustrating an embodiment of the creative promotion task delivery in an image-text format;

FIG. 7 is a diagram illustrating an embodiment of the creative promotion task delivery in video form;

FIG. 8 is a diagram illustrating an embodiment of an creative promotion task placement in the form of a voting card;

FIG. 9 is a flow diagram that illustrates the processing of an advertiser after logging onto an advertising platform, according to one embodiment;

FIG. 10 is a block diagram showing the configuration of a creative guide information generation apparatus in one embodiment;

FIG. 11 is a diagram illustrating an internal structure of a computer device in one embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail 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.

The method for generating creative guidance information provided by the application can be applied to the application environment shown in fig. 1. As shown in fig. 1, the application environment includes a terminal 102 and a server 104, and the application environment may be an environment in which the terminal 102 interacts with the server 104. Wherein the terminal 102 communicates with the server 104 via a network. The server 104 may obtain the pre-posted creative material uploaded by the terminal 102 and the posting dimensions corresponding to the creative material, where the creative material is multimedia information to be presented to the user. Further, the server 104 queries similar creative materials matched with the creative materials based on the delivery dimensions, and obtains creative guide information generation modes according to delivery forms and audience group types of the creative materials, the creative guide information generation modes include creative generation modes and data processing modes obtained by coding calculation modes based on historical creative guide information, wherein the creative guide information generation modes include creative generation modes and calculation modes based on creative dimensions, creative indexes and delivery results, and the server 104 generates creative guide information which contains the similar creative materials and is matched with the creative dimensions, the creative indexes and the delivery results based on the creative guide information generation modes. After the server 104 generates creative guide information including similar creative materials and matching with the creative dimension, the creative index and the delivery result based on the creative guide information generation manner, the server 104 may store the creative guide information in association with the pre-delivered creative material, for example, the creative guide information may be used as a creative tag of the pre-delivered creative material. The server 104 can also transmit the generated creative guide information to the corresponding terminal 102 so that the user can retrieve the creative guide information associated with the pre-delivered creative material. The pre-posted creative material and the posting dimension corresponding to the creative material may be directly obtained from the database by the server 104, or may be sent to the server 104 by other devices, such as a user terminal. The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, and the server 104 may be implemented by an independent server or a server cluster formed by a plurality of servers. It is understood that the method for generating creative guidance information provided by the embodiment of the present application may also be executed by a terminal. The server 104 may also be a service node in a blockchain system, a Peer-To-Peer (P2P) network is formed among the service nodes in the blockchain system, and the P2P Protocol is an application layer Protocol operating on a Transmission Control Protocol (TCP).

Machine Learning (ML) is a multi-domain cross discipline, and relates to a plurality of disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like. The special research on how a computer simulates or realizes the learning behavior of human beings so as to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve the performance of the computer. Machine learning is the core of artificial intelligence, is the fundamental approach for computers to have intelligence, and is applied to all fields of artificial intelligence. Machine learning and deep learning generally include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and formal education learning. The scheme provided by the embodiment of the present application relates to various types in the machine learning technology, and is not specifically limited herein.

In one embodiment, as shown in fig. 2, a method for generating creative guidance information is provided, which is described by taking the method as an example applied to the server in fig. 1, and includes the following steps:

step 202, acquiring pre-launched creative materials and launching dimensions of the creative materials; creative material is multimedia information that is presented to a user.

The pre-delivered creative material refers to a material that is not delivered yet, for example, a related material corresponding to a software product that is not disclosed yet. The creative materials refer to creative materials, which may be multimedia information presented to a user, that is, the creative materials may include multiple types of creative materials, which refer to different types of material forms, such as one or more of a video form, a text form, a picture form, a voice form, and the like. For example, the creative material may include at least one of video material, picture material, text material, and voice material, among others.

The delivery dimensions refer to different dimensions set when the creative material is delivered, for example, the delivery dimensions may include at least one of delivery traffic slot information, delivery conversion types, target coverage crowd, and the like. It is understood that the delivery dimension may be other dimensions set by the user, and the type included in the delivery dimension is not specifically limited.

Specifically, the server can obtain the pre-posted creative material uploaded by the user and the posting dimension selected by the user and corresponding to the creative material. The server can also receive a creative guide information generation instruction and process the creative materials stored in the server as pre-delivered creative materials. For example, the user may generate a corresponding creative guidance information generation instruction through a trigger operation, and when the server receives the creative guidance information generation instruction, the server may use a creative material stored in the server and corresponding to the user identifier as a pre-delivered creative material.

For example, a user a logs in a cloud platform by inputting an account password in a webpage, and the cloud platform may acquire all pre-delivered creative materials uploaded by the user a according to a user identifier, that is, the user a. Meanwhile, the user A can select the releasing dimensionality meeting the requirement of the user A from a plurality of candidate releasing dimensionalities displayed on the page, and the server can respond to the dimensionality selection operation of the user A and screen out the releasing dimensionality corresponding to the creative material to be released.

And step 204, inquiring similar creative materials matched with the creative materials based on the delivery dimensions.

The similar creative material refers to a material having similarity with the pre-delivered creative material, that is, a material having similarity with the pre-delivered creative material meeting a similarity threshold condition. For example, if the material a is an image with a face, and the material B is also an image with a face, after the server performs similarity matching between the material a and the material B, it is determined that the similarity between the material a and the material B satisfies a preset similarity threshold condition, and then the server may use the material a and the material B as similar materials.

Specifically, after the server obtains the pre-delivered creative materials and the delivery dimensions of the creative materials, the server may query similar creative materials matched with the pre-delivered creative materials in the database based on the delivery dimensions. When the server queries the similar creative materials matched with the pre-cast creative materials in the database, the server may perform query matching based on different identifiers or character strings, for example, the server may perform query based on an identification code obtained by encoding each material in advance, or may perform query based on a feature sequence obtained by performing feature tagging on each material in advance, where the identifier of the query is not specifically limited. In addition, the similar creative materials can be materials of different types, and the server can display the similar creative materials in different preset display modes, for example, the server can sort the similar materials obtained by matching according to the sequence of similarity from large to small and then display the similar materials in the interactive interface in a list form. It should be understood that, in the embodiment of the present application, the display manner of the similar creative material is not specifically limited, and may be other display manners.

For example, the server obtains a pre-delivered creative material as a material a, and the delivery dimension corresponding to the material a is as follows: the input dimension is 1: putting conversion type into downloading APP (Application program, abbreviation of Application); the release dimension 2: the target coverage population is 18-22 years old, the server can query the database for similar creative stories matching story a based on the two impression dimensions. The material a may be a feature tag which is pre-tagged and carries a corresponding feature tag identifier 1004a, and all stock materials in the database are also pre-tagged and carry a corresponding feature tag identifier, the server may perform similarity matching in the database based on the feature tag identifier 1004a corresponding to the material a, that is, query the feature tag identifier matching with 1004a, and assume that the feature tag identifier matching with 1004a is 1003a in the database, the server may further obtain a material a1 corresponding to the feature tag identifier 1003a, and the server uses the material a1 as a similar creative material of the material a. It can be understood that the similarity matching method adopted in this embodiment may be to query the feature tag identifier meeting the similarity threshold condition based on a preset similarity threshold condition, for example, the preset similarity threshold is 75%, and the similarity threshold condition is greater than or equal to the similarity threshold 75%, then the server performs similarity matching in the database based on the feature tag identifier 1004a, and when it is detected that the similarity between a certain feature tag identifier and the feature tag identifier 1004a is greater than or equal to 75%, then the server may use the material corresponding to the queried feature tag identifier meeting the similarity threshold condition as the similar material of the material a. The similarity threshold condition may also be other preset threshold conditions, and is not specifically limited herein.

Step 206, acquiring a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result.

The delivery form refers to a delivery form of the creative promotion task, and the delivery form may include a plurality of different types, for example, at least one of a printed media form, an electronic media form, a network form, an outdoor form, and the like. The network form may further include at least one of a video form, a graphic form, and the like, and the delivery form in the embodiment of the present application is mainly a network delivery form for the creative promotion task. For example, the delivery modality of the creative promotion task can be set to a delivery form including text links and videos in a certain delivery platform. The delivery form can also be other creative forms, such as: 1000 × 1000 multi-picture carousel, 228 × 100 multi-picture (text), and 540 × 540 video.

The audience group types refer to audience group types oriented by the creative promotion task, and the audience group types can comprise various types of audience groups and can also understand different industry fields. For example, audience segment types may include audience segments facing the educational industry, audience segments facing the travel industry, and so forth.

The creative guidance information generation mode is used for describing a mode of generating creative guidance information, and the creative guidance information generation mode may include a creative generation mode and a data processing mode.

The creative generation mode is used for describing a creative generation mode, and the creative generation mode can be a configuration file or a code which is stored in advance and is used for describing the creative generation. The creative generation mode in this embodiment may also be implemented by different types of templates obtained through templating in advance, for example, the server may obtain a template corresponding to the creative guidance information according to the delivery form of the creative material and the audience group type, and calculate all data required in the creative guidance information according to a data processing mode obtained by encoding the creative dimension, the creative index, and the delivery result included in the template in a calculation mode.

The historical creative guidance information refers to creative guidance information corresponding to the delivered creative promotion task, for example, the server may obtain all delivered creative guidance information from the creative guidance information base according to the historical delivery data, and perform templating processing on the historical creative guidance information.

The creative dimension refers to a dimension in the creative guidance information, and the creative dimension may be a dimension that a publisher cares about in the creative guidance information, for example, the creative dimension may include the number of delivered creative promotion tasks, the number of creatives, the type of creatives, and the like.

The creative index is an index for measuring the effect of an internet creative promotion task, for example, the creative index may include rigid indexes such as CTR and CVR, where CTR (Click-Through-Rate) is a Click Through Rate, and CTR refers to a Click reach Rate of a web advertisement (a picture advertisement, a text advertisement, a keyword advertisement, a ranking advertisement, a video advertisement, etc.), i.e., an actual Click number of the advertisement is divided by a display amount of the advertisement. Cvr (conversion rate), is an index for measuring the effectiveness of CPA advertisements, and in short, is the conversion rate from the time when the user clicks on an advertisement to the time when the user becomes an active user or even a paid user.

The calculation mode of the delivery result refers to a conclusion calculation method, and may include, for example, a data aggregation calculation method, a normalization method, or other statistical methods.

The data processing method is used to describe a data processing method, and the data processing method may be obtained by extracting and integrating dimensional data in the historical creative guidance information in advance, for example, the data processing method may include multiple types of processing functions.

Specifically, the server can obtain creative guide information generation modes according to the delivery forms of creative materials and audience population types, and the creative guide information generation modes comprise a creative generation mode and a data processing mode obtained by coding a calculation mode based on the creative dimension, the creative index and the delivery result in the historical creative guide information. All the delivered creative guidance information in the database can be obtained in advance, and the delivered creative guidance information is classified into different types of creative guidance information sets according to industries and delivery forms of delivered targets. Furthermore, the dimension, the index, the conclusion calculation and other data of each guide information in each creative guide information set can be analyzed, namely, each component in the guide information is disassembled, the calculation method of each index and conclusion is extracted from the creative guide information, the calculation process and method are coded to form a function process of the data processing method, and the data processing mode corresponding to the components such as the index and the conclusion in the type of guide information can be obtained. The creative generation mode can be a methodology for generating various guidance information templates formed in advance according to different industries and object forms. The target object mainly refers to a specific object to be delivered, such as an application program, a certain software product and the like.

For example, assume that the delivery form of the creative material is a video form; if the audience group type is an education software product oriented to the education industry, the server can query the corresponding creative guidance information template as the template A2 and query the corresponding data processing mode according to the delivery form of the creative materials and the audience group type. For example, template a2 contains an indicator of bid strategy, and the corresponding settings may be for manual pay-per-click and may also include promotional purposes, such as application downloads, landing pages, etc.

And step 208, generating creative guide information which contains similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guide information generation mode.

The creative guidance information is creative delivery guidance information, that is, information for guiding a user how to deliver a creative promotion task, for example, the creative guidance information may be a creative report, the basic composition of the creative report mainly includes creative indicators, descriptive parsing words and icon diagrams, various indicator calculation function rules and the like may be summarized and extracted according to different indicators used by different creative reports, where offline calculation and real-time calculation are involved, so that it is necessary to encode the basic split minimum calculation process as a function and then combine and combine the functions to form various different calculation processes, so as to realize calculation of different indicators.

Specifically, the server may generate creative guidance information that includes similar creative materials and matches the creative dimensions, creative indicators, and delivery results based on a creative guidance information generation manner. Wherein similar creative materials can be added to the creative guidance information as creative examples for display. The delivery results can include a creative effect, which can be a predicted creative effect calculated from the queried plurality of similar creative materials. The creative guidance information generation mode can be realized by different types of guidance information templates which are obtained through template processing in advance, for example, the server can obtain a template corresponding to the creative guidance information according to the delivery form of the creative material and the audience group type, and calculate all data required in the creative guidance information according to a data processing mode which is obtained by encoding the creative dimension, the creative index and the delivery result included in the template in a calculation mode.

For example, assume that the delivery form of the creative material is a video form; if the audience group type is an education software product oriented to the education industry, the server can query the corresponding creative guidance information template as the template A2 and query the corresponding data processing mode according to the delivery form of the creative materials and the audience group type. Further, the server can process the queried multiple similar creative materials by using the acquired data processing mode to obtain all the required guidance information data in the creative guidance information template. Further, the server may fill the obtained guidance information data into the template a2 according to the template a2, so as to obtain the initial creative guidance information. The server can add the similar materials as creative samples to the initial creative guidance information, and finally formed creative delivery guidance information B with luxuriant pictures and texts can be obtained.

In the method for generating the creative guidance information, the creative materials to be delivered in advance and the delivery dimensions of the creative materials are obtained; the creative material is multimedia information displayed to the user; inquiring similar creative materials matched with the creative materials based on the delivery dimension; acquiring a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result; and generating creative guide information which contains similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guide information generation mode. Because similar creative materials matched with the creative materials can be inquired through similar identification, new edition creative guide information which is not delivered before can be automatically generated according to related delivery data of similar creatives and creative guide information generation modes corresponding to delivery forms of pre-delivered creative materials and audience groups, and delivery suggestions which are delivered before can be carried out on users. From this, the human effect of operation output intention guide information has been improved, a large amount of work that originally manual handling changed into systematic automated processing to the creation efficiency of intention guide information has effectively been improved.

In one embodiment, as shown in FIG. 3, the step of obtaining pre-posted creative material and posting dimensions for the creative material includes:

and step 302, receiving the uploaded pre-cast creative material.

And 304, displaying different candidate delivery dimensions in the creative guidance information generation interface.

Step 306, responding to dimension screening operation, and screening the delivery dimensions of the creative materials from the candidate delivery dimensions displayed on the creative guidance information generation interface; the delivery dimension comprises at least one of delivery flow position information, delivery conversion types and target coverage crowd.

The candidate delivery dimension refers to a delivery dimension for screening shown in the interactive interface, and for example, the candidate delivery dimension includes at least one of delivery flow position information, delivery conversion type, and target coverage crowd.

The release traffic position information refers to release position information selected when a creative promotion task is released, for example, the release traffic position information may include position information or traffic information in different types of release platforms, such as a traffic position in a mobile platform or a traffic position in a pc (personal computer) platform.

The delivery conversion type refers to a promotion purpose of delivering the creative promotion task, that is, a type of conversion is obtained by delivering the creative promotion task, for example, the delivery conversion type may include application downloading, user payment, and the like, that is, a promotion purpose of delivering the creative promotion task is to enable a user to download an application or pay for purchase, and the like.

The target coverage crowd refers to a target coverage crowd for delivering the creative promotion task, namely the basic crowd orientation for determining the creative promotion task is determined, and the target coverage crowd mainly comprises specific orientation data such as a delivery area and crowd attributes.

Specifically, as shown in fig. 4, an interface diagram for uploading creative materials and screening delivery dimensions for a user is shown. The server can receive pre-delivered creative materials uploaded by the user, the user can upload various types of materials such as video materials, picture materials, text materials and voice materials in the interactive interface, and the user can search materials in the database in the interactive interface to serve as the pre-delivered creative materials.

As shown in fig. 4, the server may display different candidate delivery dimensions in the interface, assuming that the creative material uploaded by the user is a file 1 and a file 2, and selecting a dimension of target coverage crowd from the plurality of candidate delivery dimensions, the server responds to the dimension screening operation of the user, and selects the delivery dimension of the creative material as the target coverage crowd in the candidate delivery dimensions displayed on the creative guidance information generation interface. As shown in fig. 4, the delivery condition setting includes a plurality of candidate delivery dimensions, for example, delivery traffic slot information, delivery conversion types, and target coverage crowd. In addition, the user may also set a custom delivery dimension in the delivery condition setting interface, for example, the user may add a new delivery dimension as the crowd information, such as: the crowd clustering information is set to be active at 8-9 points earlier or active on weekends, and the like, and can be specifically set in association with specific delivery time zone information that a user wants to deliver creative materials. Therefore, the creative material pre-delivered by the user and the corresponding delivery dimension can be accurately acquired through the interactive interface, the corresponding creative delivery guidance information can be automatically generated subsequently according to the creative material pre-delivered by the user and the corresponding delivery dimension, a large amount of manual processing flows in the traditional mode are avoided, and accurate information pre-delivered by the user is provided for the subsequent automatic generation of the creative delivery guidance information.

In one embodiment, the step of querying similar creative materials that match the creative materials based on the dimensions of the impression includes:

when the delivery dimension is delivery traffic position information, a delivery conversion type and a target coverage crowd, querying a first creative material corresponding to the delivery traffic position information, a second creative material corresponding to the delivery conversion type and a third creative material corresponding to the target coverage crowd;

and merging the first creative material, the second creative material and the third creative material to obtain similar creative materials matched with the creative materials.

Specifically, when the server acquires the creative materials to be delivered in advance and the delivery dimensions are delivery flow position information, delivery conversion types and target coverage groups, the server can query a first creative material corresponding to the delivery flow position information, a second creative material corresponding to the delivery conversion types and a third creative material corresponding to the target coverage groups in the database, namely the server can query according to the delivery conditions of different dimensions respectively to obtain a data set corresponding to each dimension. Further, the server can merge the first creative material, the second creative material and the third creative material to obtain similar creative materials matched with the creative materials. In the present embodiment, a join method may be used for merging, where the join method is a character string method used to connect a plurality of character strings included in a list (or a tuple) into one character string. When merging strings using the join method, multiple strings in a list (or tuple) can be connected together with a fixed separator. For example, the string "c. bianching. net" can be viewed as the result of "merging [ 'c', 'bianching', 'net' ] into one string by a separator".

For example, assuming that the server queries the database for a first data set corresponding to placement traffic placement information as a1, a second data set corresponding to a placement conversion type as a2, and a third data set corresponding to a target coverage crowd as A3, the server may join the first data set a1, the second data set a2, and the third data set A3 to obtain a similar creative material set "a 1.a2. A3" matching the creative material. And the obtained similar creative material set comprises similar data sets corresponding to the three dimensions. Therefore, the dimensionalities selected by the user are respectively queried, the different dimensionalities of the similar data corresponding to the pre-delivery originality can be obtained, the queried different dimensionalities of the similar data are combined to obtain the different dimensionalities of the similar data, the data in the subsequently generated creative guidance information can reflect the effect of the pre-delivery originality more closely, and the accuracy of the creative guidance information is improved.

In one embodiment, as shown in FIG. 5, the step of merging the first creative material, the second creative material, and the third creative material to obtain similar creative materials matching the creative materials includes:

and 502, respectively performing characteristic labeling on the first creative material, the second creative material and the third creative material.

Step 504, a creative material intersection is determined among the labeled first creative material, the labeled second creative material, and the labeled third creative material.

Step 506, the creative material intersection is used as a similar creative material matched with the creative material.

The characteristic marking refers to marking the obtained characteristics after characteristic extraction is carried out on the materials, and then the characteristic sequence after characteristic marking can be obtained. For example, the trained neural network model can be used in advance to perform feature extraction and labeling processing on the creative materials in the database, so as to obtain a feature sequence after feature labeling.

Specifically, when the server inquires the first creative material that corresponds with input flow position information in the database, the second creative material that corresponds with input conversion type, and after the third creative material that corresponds with target coverage crowd, the server can carry out the feature marking to first creative material, second creative material and third creative material respectively, and confirm the creative material intersection in the first creative material after the marking, second creative material and third creative material, the server regards the creative material intersection as the similar creative material that matches with the creative material. The creative material may be a plurality of types of materials, for example, the creative material may be at least one of a video material, a picture material, a voice material, and the like.

For example, if the server queries in the database that a first creative material set corresponding to the placement traffic placement information is a1, a second creative material set corresponding to the placement conversion type is a2, and a third creative material set corresponding to the target coverage group is A3, the server may perform feature extraction on the three creative material sets a1, a2, and A3, respectively, to obtain features B1, B2, and B3 corresponding to the three creative material sets a1, a2, and A3, respectively, and further, the server performs tagging processing according to the features corresponding to each creative material set, so as to obtain feature tags C1, C2, and C3 corresponding to the three creative material sets a1, a2, and A3, respectively. The server can obtain the intersection C4 of the three creative material sets after being characterized according to the feature tag aggregation dimension, and the intersection C4 is used as a sample reference creative set of the subsequent creative guidance information.

In this embodiment, through carrying out characteristic tagging respectively to three intention material sets, mark the characteristic of every intention material set promptly, and according to characteristic tag polymerization dimension, the sample that guides information is regarded as to the intersection of the three intention material sets that acquire refers to the intention set, compare in traditional mode, the plate position flow that relies on physical input, directional crowd, the mode that the conversion type directly merges, carry out the mode that intersects the processing after characterizing respectively in this embodiment, can obtain the sample reference intention set of more accurate intention guide information, thereby the accuracy that the intention guided information generated has been improved.

In one embodiment, prior to obtaining pre-delivered creative materials and delivery dimensions for the creative materials, the method further includes the step of storing creative delivery data in a creative library, including:

acquiring a release dimension corresponding to a released creative promotion task within a preset release time period;

acquiring creative data which has an association relation with the released creative promotion task;

forming creative delivery data according to the delivered creative promotion tasks and the creative data;

storing the creative delivery data in a creative library;

based on the dimensions of the input, similar creative materials matched with the creative materials are queried, including:

and based on the delivery dimension, inquiring similar creative materials matched with the creative materials in creative delivery data of the creative library.

The released creative promotion tasks refer to released creative promotion tasks, such as promotion tasks of old version software products in different industries.

The creative data refers to data capable of improving the resource value, and for example, the creative data may include various methodologies or personalized popularization modes designed.

Specifically, the server may obtain delivery data of an intention promotion task in the large-scale, and may include delivery dimensions corresponding to a delivered intention promotion task within a preset delivery time period, for example, a user wants to know the intention promotion task data of a software product within approximately three months, and may set the preset delivery time period to 6-8 months, and the server may obtain all the intention promotion task data of the software product within this time period of 6-8 months, where the delivery dimensions corresponding to the delivered intention promotion task at least include at least one of delivery orientation and crowd orientation data, traffic position data, and conversion type data. Further, the server can acquire creative data associated with the delivered creative promotion task, and form creative delivery data according to the delivered creative promotion task and the creative data, and store the obtained creative delivery data in a creative library. Therefore, in the data processing process of generating the creative guidance information, the server can inquire similar creative materials matched with the pre-posted creative materials in creative putting data of the creative library based on putting dimensions corresponding to the pre-posted creative materials.

For example, the creative promotion task is taken as an example of promoting the advertisement. Assuming that the preset delivery time interval is set to be within nearly 3 days, the server can acquire large-disk advertisement delivery data within nearly three days, including delivery orientation and crowd orientation data, traffic positions, conversion types and other data, and meanwhile, the server can extract the association relationship between the advertisement and the creative idea to form an association table of the advertisement creative idea. Furthermore, the server associates the advertisement delivery data with the advertisement creative association table to form creative delivery data, and because the data volume of the creative delivery data is large, the obtained creative delivery data can be stored in a hive table in the embodiment, and the hive table has no upper limit on the stored data, wherein the hive table is a data warehouse tool for extracting, converting and loading data. In addition, the server can write the creative release data into the HBASE database corresponding to each dimension segment, although the warehousing process of HBASE is slow, the query speed is high, HBASE is a distributed column-oriented open source database, and meanwhile, the server can provide a query interface of the creative release data for the HBASE database. It should be understood that the storage manner of the creative launch data in this embodiment may be other storage manners besides the storage manner of the hive library table and the storage manner of the fragmented write-in HBASE database, and the storage manner of the creative launch data are not particularly limited herein.

In this embodiment, the server can inquire similar creative materials matched with the pre-posted creative materials in the creative posting data of the creative library based on the posting dimensions corresponding to the pre-posted creative materials, that is, similar data used for generating creative guidance information is determined in a similar matching manner, so that subsequent processing can be automatically performed according to the inquired similar materials, corresponding creative posting guidance information is generated, and the generation efficiency of the creative guidance information is effectively improved.

In one embodiment, the method further includes a step of querying the impression effect data, specifically including:

inquiring launching effect data corresponding to the first creative material, the second creative material and the third creative material;

generating creative guidance information which contains similar creative materials and is matched with creative dimensions, creative indexes and delivery results based on a creative guidance information generation mode, wherein the creative guidance information generation method comprises the following steps:

generating initial creative guide information matched with the creative dimension, the creative index and the delivery result based on a creative guide information generation mode;

and adding the similar creative materials serving as creative samples into the initial creative guide information to obtain the creative guide information.

The delivery effect data refers to predicted effect data, and the initial creative guidance information refers to guidance information including a creative dimension, a creative index, and delivery result data, for example, the initial creative guidance information may be an initial creative report. The creative sample refers to a sample, namely, the sample serves as a reference sample to enable a user to know the creative putting effect more intuitively. For example, a creative sample may include a specific example picture, an example video, and the like.

Specifically, when the delivery dimensions corresponding to the pre-delivered creative material are delivery traffic position information, delivery conversion type and target coverage crowd, the server may query a first creative material corresponding to the delivery traffic position information, a second creative material corresponding to the delivery conversion type and a third creative material corresponding to the target coverage crowd, further, the server may query delivery effect data corresponding to the first creative material, the second creative material and the third creative material in a database, for example, by taking creative guidance information as an example, the server may obtain a creative report generation manner according to the delivery form and the industry type to which the creative material belongs, the creative report generation manner includes a creative generation manner and a data processing manner encoded by a calculation manner based on the creative dimensions, creative indexes and delivery results in the historical creative report, the server can generate an initial creative report matched with the creative dimension, the creative index and the delivery result based on a creative report generation mode, and adds similar creative materials into the initial creative report to obtain the creative report as creative samples. Therefore, the corresponding creative delivery guidance report is automatically generated according to the creative materials, a large amount of original manual processing work is converted into systematic automatic processing, and the generation efficiency of the creative report is effectively improved.

In one embodiment, the step of generating initial creative guidance information matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generation mode comprises the following steps:

processing the similar creative materials by utilizing a creative guide information generation mode to generate creative guide information data;

inquiring a corresponding creative guidance information template based on the delivery form and audience group type of the creative material;

and adding the creative guide information data to the creative guide information template to obtain initial creative guide information.

The creative guidance information data refers to data to be filled in generating the creative guidance information, for example, the creative guidance information data may be creative report data, and the creative report data may include specific index data, delivery effect data, and dimension data obtained through calculation.

The creative guide information template refers to a guide information template generated in advance according to historical creative guide information, for example, the creative guide information template may be a creative report template, and the creative report template may be divided into different types of creative report templates according to different industries.

Specifically, the explanation will be given by taking the creative guidance information template as a creative report template. The server can acquire the industry of the user and the form of the delivered object, and inquire a proper report template in the database according to the industry and the form of the object. Furthermore, the server can extract a corresponding creative method argument calculation function from the index of the database, process the similar creative materials in the queried similar creative list by using the creative method argument calculation function, calculate and generate all data required in the creative report, namely calculate and obtain creative report data, and finally fill the calculated creative report data into the report template according to creative dimensions, creative indexes and delivery results in the creative report template to generate the content required by the report, namely obtain the initial creative report. Therefore, a large amount of original manual processing work is converted into automatic processing of the system, and the generation efficiency of the creative report is effectively improved.

In one embodiment, the creative promotion task is taken as a promotion task of a certain software product. As shown in fig. 6, a schematic diagram of the creative promotion task in an image-text form is put. Wherein, in FIG. 6

Representing the advertiser head portrait and name, for example, may be a specific business icon and business name. In FIG. 6

Representing a case;

representing a picture material;

representing a conversation window, namely social interaction, for example, when a user sees the creative promotion task in FIG. 6, the user can click the conversation window to establish real-time conversation interaction to know related information;

representing the text chain, namely, the user can jump to a specific product introduction detail page through the text chain;

the representative icon identification may be, for example, an application icon for a particular software product download. Assuming that the server obtains the target object in the form of image and text from the advertiser industry as the tourism industry and the target object is released, and a suitable report template A is inquired in the database according to the industries and the object forms, the report template A comprises creative dimensions, creative indexes, a generation mode of a delivery result and a corresponding data processing mode, the server can extract the corresponding creative method argument computation function from the index of the database, process the inquired similar creative materials by using the creative method argument computation function, compute and generate all the data required in the creative report, and finally, the server fills the calculated creative report data into the report template according to the creative dimension, the creative index and the delivery result in the creative report template A to generate the content required by the report, so that the initial creative report can be obtained. Further, the server can also take similar creative materials as creative samples and add the creative samples to the initial creative report to obtain the creative report. I.e. as in fig. 6

Similar picture materials of pre-projected picture materials can be used as creative examples for displaying, so that an advertiser can know the effect of creative material projection more visually.

As shown in fig. 7, a schematic diagram of a video format is put on the creative promotion task. Wherein, in FIG. 7

Representing an advertiser's avatar and name;

representing a case;

representing video material;

a representative session window;

represents a chain of words;

representing an application icon identification. Assuming that the server can acquire that the advertiser industry is the tourism industry and the delivered target object is in a video form, inquiring a proper report template B in a database according to the industry and the target object form, and acquiring an initial creative report according to the same data processing flow as the above. Further, the server can also take similar creative materials as creative samples and add the creative samples to the initial creative report to obtain the creative report. I.e. as in fig. 7

Similar video materials of the pre-released video materials can be used as creative examples for displaying, so that an advertiser can know the releasing effect of the creative materials more visually.

As shown in fig. 8, the project form for the creative promotion task is a schematic diagram in the form of a voting card. Wherein, in FIG. 8

Representing an advertiser's avatar and name;

representing video material;

a representative case title;

representative documentation details;

an icon identification representing an application;

the representative selection button is that the user can vote to select the APP (application program) which the user likes to support by clicking the button. The server can acquire the form of the target object which is launched in the tourism industry as the advertiser industry, inquire a proper report template C in a database according to the industry and the target form, and acquire an initial creative report according to the same data processing flow as the above. Further, the server can also take similar creative materials as creative samples and add the creative samples to the initial creative report to obtain the creative report. I.e. as in fig. 8

Similar video materials of the pre-released video materials can be used as creative examples for displaying, so that an advertiser can know the releasing effect of the creative materials more visually. In FIG. 8

The specific data in the creative report may be 6-15 seconds in duration, 640 x 360 pixels in resolution, skip: and the click jump system automatically generates a simple native page and the like. It should be understood that the delivery form of the creative promotion task in the embodiment of the present application may be other delivery forms besides the above-mentioned form, and is not limited specifically here. From this, the human effect of operation output intention report has been improved, converts a large amount of work of original manual handling into systematic automated processing to the generation efficiency of intention report has effectively been improved.

In one embodiment, the method further includes a step of forming a creative generation manner, specifically including:

in the creative guidance information base, acquiring at least one piece of relevant historical creative guidance information based on the delivery form of creative materials and the audience group type;

analyzing the creative dimension, the creative index and the delivery result of at least one piece of historical creative guidance information to obtain a calculation mode of the creative dimension, the creative index and the delivery result of the historical creative guidance information;

coding the obtained calculation mode to obtain a data processing mode;

storing the obtained data processing mode according to the configuration type and the function type;

and forming a creative generation mode according to the delivery form of the creative materials and the audience group types in the historical creative guidance information.

The historical creative guidance information refers to creative guidance information corresponding to a delivered creative promotion task, and for example, the historical creative guidance information may be a historical creative report.

Specifically, the description will be given by taking the historical creative guidance information as an example of a historical creative report. The server can obtain all operated creative reports in the creative report library according to historical delivery data, and further, the server can select a related creative report list according to the industry and the forms of different delivery objects, namely, the server obtains at least one related historical creative report based on the delivery forms and audience group types of creative materials. The delivery target object mainly refers to a delivered specific object in the creative promotion task, such as an app, a commodity and other objects. Furthermore, the server analyzes the dimension, the index, the conclusion calculation and the like in the template, namely analyzes the creative dimension, the creative index and the delivery result of at least one historical creative report to obtain the calculation mode of the creative dimension, the creative index and the delivery result of the historical creative report. After the server obtains the calculation methods of the indexes and the conclusions from the historical creative report, the server can encode the calculation processes and methods to form a data processing method function process, and stores the dimensions, the index calculation methods and the conclusion calculation methods in a configuration type and a function type to form an associated index; and finally, the server forms various report template generation methodologies according to the target object forms set by the industry and the users, namely, the server forms corresponding creative generation modes according to the delivery forms and audience group types of creative materials in the historical creative reports.

In the embodiment, the creative reports which are operated and output are templated, the sharing methodology is extracted, and classification and division are performed according to the industry of the advertiser and the object form. Before the advertiser puts, creative materials can be put in through providing a sample, namely similar identification is carried out in historical putting full creative materials, meanwhile, according to the industry to which the advertiser belongs and the expected putting target object form, a new edition of creative guide report before putting is formed according to related putting data of similar creative and a set methodology in a creative guide report template, and therefore putting suggestions before putting can be carried out on the advertiser.

In one embodiment, the creative material includes video material; based on the delivery dimensions, the step of querying similar creative materials matched with the creative materials comprises the following steps:

extracting a characteristic vector corresponding to each frame of image in the video material;

performing dimensionality reduction processing on the feature vectors to obtain dimensionality reduction feature vectors of each frame of image;

performing feature conversion on the dimensionality reduction feature vector to obtain a feature vector of the video material;

and inquiring similar creative materials matched with the feature vectors of the video materials based on the delivery dimensions.

The user can upload various types of pre-delivered creative materials, and the creative materials comprise at least one of video materials, picture materials, voice materials, text materials and the like. Specifically, the creative material is taken as a video material for example. The server can acquire the video material which is uploaded by the user and is pre-launched and the launching dimensionality of the video material, and further can extract the feature vectors corresponding to all the frames of images in the video material and perform dimensionality reduction processing on the feature vectors to obtain dimensionality reduction feature vectors of all the frames of images; the server performs feature conversion on the dimensionality reduction feature vector to obtain a feature vector of the video material, and the server can inquire similar creative materials matched with the feature vector of the video material based on the corresponding release dimensionality of the video material.

In this embodiment of the present application, the server may perform a Pooling (i.e., pooling) method to perform a feature extraction operation on each frame of image in the video to obtain a corresponding feature vector, and perform max Pooling (i.e., maximum pooling) or average Pooling (i.e., average pooling) processing on a time dimension, so as to obtain a feature obtained by performing pooling operation on features of multiple frames of images, and further fuse the features of the pooled multiple frames of images into a feature of the entire video. It can be understood that, in the data preprocessing, the server may also perform feature extraction on the video material in the history library by using a pooling method, and store the features corresponding to the entire video material obtained by the final fusion in a cache such as redis or mysql.

The pooling operation has the effect that dimension reduction processing can be performed on the image, namely, the image is compressed. For example, a pooling layer may be placed in the neural network model among successive convolutional layers, the pooling layer being used to compress the data and reduce the amount of parameters, reducing overfitting. The pooling operation used in the embodiments of the present application may be a maximum pooling operation, which retains texture features, or an average pooling operation, which retains overall data features. Pooling may be understood to be essentially a "down-sampling" operation.

Therefore, the characteristic tagging processing is carried out on the historical stock materials, the characteristic tagging processing is also carried out on the pre-delivered materials uploaded by the user, the system can automatically carry out similar matching on the pre-delivered materials and the off-line stock materials, the similar creatives searched in the creative library can be used as matching reference objects to generate creative guidance information, and a large amount of original manual processing work is converted into the automatic processing of the system, so that the generation efficiency of the creative guidance information is effectively improved.

In one embodiment, the feature vectors corresponding to each frame of image in the video material include frame rate feature vectors and definition feature vectors, and the step of performing feature conversion on the reduced-dimension feature vectors to obtain the feature vectors of the video material includes:

respectively carrying out dimensionality reduction processing on the frame rate feature vector and the definition feature vector to obtain a dimensionality reduction frame rate feature vector and a dimensionality reduction definition feature vector;

performing feature conversion processing on the dimensionality-reduced frame rate feature vector and the dimensionality-reduced definition feature vector to obtain a frame rate feature vector and a definition feature vector of the video material;

and merging the frame rate characteristic vector and the definition characteristic vector of the video material to obtain the characteristic vector of the video material.

Specifically, the feature vectors corresponding to each frame of image in the video material may include a frame rate feature vector and a definition feature vector, and the server may perform dimension reduction processing on the frame rate feature vector and the definition feature vector by using a posing method, respectively, to obtain a dimension-reduced frame rate feature vector and a dimension-reduced definition feature vector, and further, the server may perform feature conversion processing on the dimension-reduced frame rate feature vector and the dimension-reduced definition feature vector, to obtain a frame rate feature vector and a definition feature vector of the video material, and merge the frame rate feature vector and the definition feature vector of the video material, to obtain the feature vector of the video material. It can be understood that the pooling method in this embodiment mainly performs dimension reduction on feature maps, that is, a data set for marking creatives through feature groups may have thousands of enumerated values in original features, and dimension reduction needs to be performed to less than one hundred to perform faster calculation. For example, assume that there is a creative feature set called the creative basic attributes, which includes two features: frame rate and definition, it is clear that the enumerated values contained in these two fields overflow if they are exhausted, so that the average posing (i.e. average pooling) process is required for the dimension reduction process. Therefore, by performing dimension reduction processing on the characteristics of the video material, the efficiency of calculation processing can be effectively improved and faster calculation can be performed under the condition that the accuracy of characteristic extraction is ensured, and the generation efficiency of creative guidance information is effectively improved.

In one embodiment, the step of querying for similar creative material that matches feature vectors of video material comprises:

coding the feature vectors of the video material to obtain video feature codes corresponding to the video material;

and inquiring the creative materials according to the video characteristic codes to obtain similar creative materials matched with the video materials.

The server combines the creative characteristics of all dimensions of the video material to obtain the characteristic vector of the whole video material, and then the server can encode the characteristic vector of the video material to obtain the video characteristic code corresponding to the video material and inquire the creative material according to the video characteristic code to obtain the similar creative material matched with the video material. The server may perform hash (i.e., hash function or hash code) encoding on the corresponding features of the entire video material, and store the encoded features in the cache. It should be understood that, in addition to the hash coding method, the coding method in this embodiment may be other coding methods, and a specific coding method is not limited herein. Therefore, the system can automatically perform similar matching with the offline stock materials in the creative library according to the feature codes corresponding to the pre-placed video materials, the searched similar creatives are used as matching reference objects to generate creative guidance information, and a large amount of original manual processing work is converted into automatic processing of the system, so that the generation efficiency of the creative guidance information is effectively improved.

In one embodiment, the method further includes the step of storing the material feature codes in the creative library, specifically including:

performing feature extraction on stock creative materials in a creative library to obtain feature vectors of the stock creative materials;

coding the feature vectors of the stock creative materials to obtain material feature codes corresponding to the stock creative materials;

storing the characteristic codes of all the materials in a creative library;

and inquiring creative materials according to the video characteristic coding to obtain similar creative materials matched with the video materials, wherein the creative materials comprise the following steps:

searching a target material characteristic code matched with the video characteristic code in the creative library;

and acquiring similar creative materials corresponding to the target material feature codes in a creative library.

The stock creative materials refer to materials stored in a database and can be processed in an off-line mode.

Specifically, the server can extract the features of the stock creative materials in the creative library by using a pooling method to obtain the feature vectors of the stock creative materials, perform hash coding on the feature vectors of the stock creative materials to obtain the material feature codes corresponding to the stock creative materials, store the material feature codes in the creative library, and further perform similar matching query according to the feature codes of the video materials to obtain similar creative materials matched with the video materials in a data processing flow of generating creative guidance information in a subsequent server, namely the server can search the target material feature codes matched with the video feature codes in the creative library and obtain the similar creative materials corresponding to the target material feature codes in the creative library. Therefore, the system can automatically perform similar matching with the offline stock materials in the creative library according to the feature codes corresponding to the pre-delivered video materials, so that the found similar creatives are used as matching reference objects to generate creative guidance information, and a large amount of original manual processing work is converted into automatic processing of the system, so that the generation efficiency of the creative guidance information is effectively improved.

In one embodiment, the method further comprises the step of performing creative material query based on the verification code and the feature code, and the step of performing creative material query comprises the following steps:

acquiring a check code corresponding to the video material;

and inquiring the creative materials based on the check codes and the characteristic codes to obtain similar creative materials matched with the video materials.

The check code is usually the last digit of a group of digits, and is derived from the previous digit by some operation to check the correctness of the group of digits. The algorithms for different check codes are often different, for example, the check code algorithm has: code distance, parity check, hamming check, cyclic redundancy check, etc., the check code in this embodiment may be MD5 value, and MD5 refers to MD5 information Digest Algorithm (english: MD5 Message-Digest Algorithm).

Specifically, after the user uploads the pre-delivered video materials to the server, the server can acquire the pre-delivered video materials, store the pre-delivered video materials in the database, and simultaneously generate the corresponding verification video md5, so that the subsequent server can perform creative material query based on the verification code, namely the md5 value, and the feature code to obtain similar creative materials matched with the video materials. Therefore, the system can be matched with the off-line stock materials in a similar manner in the creative library automatically according to the feature codes and the check codes corresponding to the pre-placed video materials, the found similar creatives are used as matching reference objects to generate creative guidance information, and a large amount of original manual processing work is converted into automatic processing of the system, so that the generation efficiency of the creative guidance information is effectively improved.

The application also provides an application scene, and the application scene applies the creative guidance information generation method. Specifically, the creative guidance information generation method is applied to the application scenario as follows:

when a user needs to launch an creative promotion task, the method for generating the creative guidance information can be adopted, namely, the user logs in the cloud platform, the user can upload creative materials launched in advance to the cloud platform in a customized manner, the creative materials launched in advance and corresponding launching dimensions can be directly selected from the cloud platform, and then the cloud platform automatically processes according to the creative materials launched in advance and the corresponding launching dimensions to generate a corresponding creative guidance report. The cloud platform can be implemented by an independent server or a server cluster composed of a plurality of servers. For example, when a user needs to perform creative promotion task delivery for a certain education software product, the cloud platform may automatically generate a creative report function by starting the system, for example, an application program for generating the creative report, so that the system may automatically generate a corresponding creative guidance report according to a pre-delivered creative material and delivery dimensions uploaded by the user, and return the generated creative guidance report to the user interface, so that the user may directly view the creative guidance report related to the pre-delivered creative material in the display interface.

The method provided by the embodiment of the application can be applied to scenes for automatically generating the creative guidance information of the internet advertisement. The following description will be made by taking the automatic generation of creative guide information of internet advertisements as an example, and the method for generating creative guide information provided by the embodiment of the present application includes the following steps:

in the current internet advertisement, when an advertiser puts an advertisement, the advertiser usually knows the traffic condition corresponding to each putting platform in advance and the creative related data put on the advertisement positions such as each traffic and position, if the advertiser cannot completely know the information, the advertiser can only output multiple creatives by uploading multiple materials or carry out 2-3 days of putting experiments in different advertisement accounts to determine the creative idea and putting plan of the advertisement putting, the period of the putting experiments can be called as a learning period, if the creative can not be put normally in the learning period, the advertiser needs to do more initial experiments to adapt to an advertisement system, and because the processing process needs manual participation to process a large amount of information, the generation efficiency of creative guidance information is low. Therefore, in order to effectively improve the generation efficiency of creative guide information and enable the server to guide the advertiser to deliver the advertisement with an entity capable of realizing creativity, such as creative, the embodiment provides an automatic generation method of a creative material delivery guide report, which specifically includes:

on the product side, as shown in fig. 9, a schematic processing flow diagram after logging in an advertisement platform for an advertiser is shown. The applicable product is an input guidance report for automatically generating creative materials, and the whole process is as follows: after an advertiser logs in an advertisement platform, pre-delivery creative materials (including video materials, picture materials and the like) are uploaded, delivery conditions are screened, and screening dimensions include ad slot information such as a, flow, position and the like; b. transformation types such as oca (i.e., Optimized CPA), etc.; c. and basic crowd orientation including the gender of the crowd, the oriented province and city region, the type of the mobile phone operating system and the like. oca provides advertisers with automatic optimization of effectiveness based on placement goals and bids, continuously improving the efficiency and the input-output ratio of the advertisers' advertisements. Furthermore, the advertisement system positions the proper creative effect and example creative according to the pre-delivery creative uploaded by the advertiser and the screened delivery conditions, and forms a corresponding creative guidance report, wherein the creative guidance report is fed back to the advertiser in a mode of combining pictures and texts and digitally enhancing the description, so that creative delivery guidance of the advertisement is given when the advertiser does not start delivery, namely, under the background without delivery history data.

The main product realization process comprises: s1, an advertiser logs in an advertisement platform and uploads a pre-cast creative material which can be a sample video and a picture; s2, the advertiser determines the conditions to be delivered in the interactive interface, such as: advertisement position information such as flow and position information, conversion type information such as app activation, app downloading, form filling and the like, and basic crowd targeting information such as conditions of a delivery area, crowd attributes and the like are determined; s3, matching the pre-released materials and releasing conditions in a creative library by the system; s4, returning the delivered prediction effect and creative example by the system; and S5, generating a creative delivery guidance report for the advertiser according to the result delivery prediction effect and the creative example. In addition, if the user needs to add customized delivery conditions, the user can also set in the interactive interface, for example, set crowd clustering information such as: the early 8-9 points are active, the weekend is active and the like, and the method can be specifically associated with the time period of advertiser delivery and the like. The system mainly processes through similar matching when matching is conducted in the creative library according to the pre-delivery materials and delivery conditions, on one hand, all stock materials are subjected to labeling processing in an off-line marking mode, on the other hand, pre-delivery material sets uploaded by an advertiser are also subjected to labeling processing to be similarly matched with the off-line stock materials, so that the system can find similar creative materials corresponding to the pre-delivery materials in the creative library according to labeling marks and generate reports by taking the similar creative materials as matching reference objects.

On the technical side, in the traditional technology, the guidance of the delivery platform to the advertiser is mainly in the advertisement granularity, that is, a delivery methodology is formed to guide the advertiser simply by delivering the targeted suggestions, bid suggestions, traffic slots and other delivery conditions. The granularity is lowered to the originality in the embodiment, and two main reasons exist, namely for the internet advertisement which is developed increasingly at present, as the engine tends to understand the content, the advertisement does not have more analysis dimensions as a carrier, and the originality has more and more complete attributes as the granularity of a deeper layer of the advertisement; secondly, because the creative behavior relationship is closer to the behavior relationship of the user, the creative portrait can be better described by adding various attributes to the quality and the creative, and further, the finer portrait of the advertisement can be described. Therefore, the embodiment provides an automatic generation method of an creative material delivery guidance report, which includes the following specific steps:

1) acquiring a pre-launched creative material and launching dimensions of the creative material; the creative material is multimedia information displayed to the user; and inquiring similar creative materials matched with the creative materials based on the delivery dimension.

Specifically, the system can perform similarity retrieval processing on the pre-projected material, including: s1, through uploading by a user, the system acquires the pre-cast creative material, stores the creative material in the data storage and simultaneously generates a check code, namely an md5 value; and S2, for the creative materials uploaded by the user, assuming that the creative materials uploaded by the user are video materials, the system can extract the global characteristics of the video materials and calculate the similarity with each material in the database to judge whether the materials are similar. The method mainly used in the step of extracting the features by the system is a pooling method, namely, a server performs max pooling or average pooling processing on the feature vector of each frame of image in a video in a time dimension to obtain the feature vector of a multi-frame image subjected to the dimension reduction processing, so that the features of the multi-frame image are combined into the features of the whole video, in addition, when data is preprocessed, the system can also uniformly perform feature extraction on the video in a historical library by using the pooling method, perform hash coding on the extracted features, and store the features in a cache such as redis or mysql; and S3, inquiring in an inquiry library by calculating the hash code value and the md5 value of the creative material uploaded by the user, and returning a similar video list if the creative material is hit.

Furthermore, the system can screen the throwing conditions of the pre-thrown materials, and mainly comprises two parts of data offline processing and retrieval. Firstly, a data offline processing flow is as follows:

the method comprises the following steps: s1, the system acquires large-disk advertisement delivery data including delivery orientation and crowd orientation data, flow format positions, conversion types and the like, and all the data are extracted according to advertisement granularity; s2, the system acquires a release time section, generally an interval of nearly three months, acquires the association relationship between the advertisement and the creative, and forms an association table of the advertisement > the creative; s3, associating the advertisement delivery data with the associated advertisement > creative association table to form creative delivery data; s4, suggesting to store a hive library table because of large data volume of creative delivery data, wherein the hive library table has no upper limit on data storage; s5, writing the creative delivery data into HBASE corresponding to each dimension segment, wherein the storage process of HBASE is slow, but the query is fast; and S6, providing a query interface of creative delivery data for HBASE by the system. The association relation between the advertisement and the creative is generated autonomously in the advertising process of the advertiser, the advertiser can update and replace the creative in each period in the advertising process, and the behavior operations can be recorded by the system to form an association table of the advertisement and the creative. The hive library table and HBASE database are tools used in big data storage and analysis, and can be used for processing large-batch data.

Secondly, a processing flow of searching according to the throwing condition:

the method comprises the following steps: s1, the system acquires the releasing conditions selected by the user; s2, dividing dimensions by the system according to the delivery flow position, the directional crowd and the conversion type; s3, the system queries in the database according to different dimensions to obtain a returned data set; s4, the system carries out join merging on the data sets (flow format, oriented crowd and conversion type) inquired by the three dimensions; and S5, returning a final query result, namely an intersection of the three data sets after merging, wherein the intersection comprises sample originality and putting effect data. The method mainly comprises the steps that a system finds released creatives meeting requirements in a corresponding creative set range according to releasing conditions screened by users, three dimensions respectively find corresponding three data sets, the system respectively carries out characteristic labeling processing on the three data sets, join merging is carried out according to characteristic label aggregation dimensions to obtain intersection of the three data sets, and the obtained intersection is used as a sample reference creative set of a subsequent report. Compared with the traditional mode, the method has the advantages that the screening and the combination are directly carried out depending on the conditions of physically input layout flow, oriented population, conversion types and the like, and the processing steps in the example are characterized and then intersected, so that the obtained data is more accurate, and the generated report has a guiding value.

2) Acquiring a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result; and generating creative guide information which contains similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guide information generation mode.

Specifically, the system can automatically generate the creative guidance report according to the pre-delivery materials and the delivery conditions, and the generation of the creative guidance report mainly comprises two parts:

the first part is the generation of a guide report template, and the main steps are as follows: s1, the system acquires all operation creative reports in the creative report library according to historical delivery, and selects a related creative report list according to the industry and the form of different delivery targets; s2, calculating the dimension, the index and the conclusion in the system analysis template, and the like, and splitting all the components in the report; s3, the system obtains the index and the conclusion calculation method, and encodes the calculation process and method to form the data processing method function process; s4, the system stores the dimension, the index calculation method and the conclusion calculation method in a configuration type and a function type, and forms an associated index; and S5, forming various report template generation methodologies according to the object forms of the industry and the users by the system. The system can summarize and refine various index calculation function rules and the like from historical reports in advance according to different indexes used by different reports, wherein offline calculation and real-time calculation are involved, so that the minimum calculation process after basic splitting is required to be used as a function to be coded, combined and combined to form various different calculation processes to calculate different indexes. The specific dimensions in the report may be dimensions of interest in the advertiser report, generally dimensions of interest in the industry, such as number of delivered advertisements, number of creatives, type of creatives, and the like; the specific indexes in the report can be hard indexes such as ctr, cvr and the like; the calculation method of the conclusion in the report can be a data aggregation calculation method, such as a statistical method like a normalization method, and the dimension and index methods can be divided into two types in development: configuration and function. The formed associated index refers to a report, and various configurations and function methods corresponding to the report can be retrieved from a repository, so that the report can be queried and retrieved quickly. Compared with a hard coding mode adopted in a traditional mode, for example, one report corresponds to one calculation process, namely, the report and the calculation function are in one-to-one association, and the universality is lacked, the method adopted in the embodiment stores and retrieves the general configuration and the general function as resources, and in the processing flow of automatically generating the report by the system, the report can be generated, analyzed and calculated in real time by establishing indexes, so that the method is more flexible and suitable for generating creative reports of different types.

The second part is guiding report automatic filling, and the main steps are as follows: s1, the system acquires the form of the user industry and the released object; s2, the system searches and inquires a proper template according to the industry and the target form; s3, the system acquires the queried similar creative list, including the delivery data of the creative sample and the creative; s4, the system extracts creative method argument calculation functions from the index, analyzes and calculates all data needed in the generated report; and S5, generating the content required by the report according to the template, and adding the creative sample.

In the embodiment, the creative reports which are operated and output are templated, the sharing methodology is extracted, and classification and division are performed according to the industry of the advertiser and the object form. Before the advertiser puts, creative materials can be put in through providing a sample, namely similar identification is carried out in historical putting full creative materials, meanwhile, according to the industry to which the advertiser belongs and the expected putting target object form, a new edition of creative guide report before putting is formed according to related putting data of similar creative and a set methodology in a creative guide report template, and therefore putting suggestions before putting can be carried out on the advertiser. Therefore, the human efficiency of the operation output creative guidance report is improved, a large amount of original manual work is converted into systematic automatic processing, and the generation efficiency of the creative report is effectively improved.

It should be understood that although the various steps in the flow charts of fig. 1-9 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 1-9 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least some of the other steps.

In one embodiment, as shown in fig. 10, there is provided an apparatus for generating creative guidance information, which may be a part of a computer device using a software module or a hardware module, or a combination of the two, the apparatus specifically includes: a creative material acquisition module 1002, a similar material query module 1004, a generation mode acquisition module 1006, and a generation module 1008, wherein:

a creative material acquisition module 1002, configured to acquire pre-delivered creative materials and delivery dimensions of the creative materials; creative material is multimedia information that is presented to a user.

And the similar material query module 1004 is used for querying similar creative materials matched with the creative materials based on the delivery dimensions.

A generation mode acquisition module 1006, configured to acquire a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result.

And a generating module 1008, configured to generate creative guidance information that includes similar creative materials and is matched with the creative dimension, the creative index, and the delivery result based on the creative guidance information generating manner.

In one embodiment, the creative material acquisition module includes: the creative material receiving unit is used for receiving the uploaded pre-posted creative material; the delivery dimension display unit is used for displaying different candidate delivery dimensions in the creative guidance information generation interface; the delivery dimension screening unit is used for responding to dimension screening operation and screening out delivery dimensions of the creative materials from the candidate delivery dimensions displayed on the creative guidance information generation interface; the delivery dimension comprises at least one of delivery flow position information, delivery conversion types and target coverage crowd.

In one embodiment, the similar material query module is further configured to query a first creative material corresponding to the delivery traffic placement information, a second creative material corresponding to the delivery conversion type, and a third creative material corresponding to the target coverage group when the delivery dimensions are the delivery traffic placement information, the delivery conversion type, and the target coverage group; the similar material inquiry module comprises: and the creative material merging unit is used for merging the first creative material, the second creative material and the third creative material to obtain similar creative materials matched with the creative materials.

In one embodiment, the apparatus further comprises: the release dimension acquisition module is used for acquiring release dimensions corresponding to released creative promotion tasks within a preset release time period; the creative data acquisition module is used for acquiring creative data which has an association relation with the released creative promotion task; the storage module is used for forming creative delivery data according to the delivered creative promotion tasks and the creative data; storing the creative delivery data in a creative library; the similar material query module is also used for querying similar creative materials matched with the creative materials in creative delivery data of the creative library based on delivery dimensions.

In one embodiment, the apparatus further comprises: the delivery data query module is used for querying delivery effect data corresponding to the first creative material, the second creative material and the third creative material; the generating module is also used for generating initial creative guidance information matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generating mode; and the adding module is used for adding the similar creative materials serving as creative samples and putting effect data into the initial creative guidance information to obtain the creative guidance information.

In one embodiment, the generating module comprises: the guidance information data generation unit is used for processing the similar creative materials by utilizing a creative guidance information generation mode to generate creative guidance information data; the creative guide information template inquiry unit is used for inquiring a corresponding creative guide information template based on the delivery form and audience group type of the creative material; and the adding unit is used for adding the creative guidance information data to the creative guidance information template to obtain initial creative guidance information.

In one embodiment, the apparatus further comprises: the historical guidance information acquisition module is used for acquiring at least one piece of relevant historical creative guidance information in a creative guidance information base based on the delivery form of creative materials and the audience group type; the analysis module is used for analyzing the creative dimension, the creative index and the delivery result of at least one piece of historical creative guidance information to obtain the calculation mode of the creative dimension, the creative index and the delivery result of the historical creative guidance information; the coding module is used for coding the obtained calculation mode to obtain a data processing mode; the storage module is also used for storing the obtained data processing mode according to the configuration type and the function type; the generation module is also used for forming a creative generation mode according to the delivery form and audience group types of the creative materials in the historical creative guidance information.

In one embodiment, the apparatus further comprises: the characteristic marking module is used for respectively marking the characteristics of the first creative material, the second creative material and the third creative material; the determining module is used for determining creative material intersection in the marked first creative material, the marked second creative material and the marked third creative material; and taking the intersection of the creative materials as similar creative materials matched with the creative materials.

In one embodiment, the similar material query module comprises: the characteristic vector acquisition unit is used for extracting characteristic vectors corresponding to all frames of images in the video material; performing dimensionality reduction processing on the feature vectors to obtain dimensionality reduction feature vectors of each frame of image; performing feature conversion on the dimensionality reduction feature vector to obtain a feature vector of the video material; and inquiring similar creative materials matched with the feature vectors of the video materials based on the delivery dimensions.

In one embodiment, the feature vector obtaining unit is further configured to perform dimension reduction processing on the frame rate feature vector and the definition feature vector, respectively, to obtain a dimension-reduced frame rate feature vector and a dimension-reduced definition feature vector; performing feature conversion processing on the dimensionality-reduced frame rate feature vector and the dimensionality-reduced definition feature vector to obtain a frame rate feature vector and a definition feature vector of the video material; and merging the frame rate characteristic vector and the definition characteristic vector of the video material to obtain the characteristic vector of the video material.

In one embodiment, the encoding module is further configured to encode the feature vectors of the video material to obtain video feature codes corresponding to the video material; the similar material inquiry module is also used for inquiring the creative materials according to the video characteristic codes to obtain similar creative materials matched with the video materials.

In one embodiment, the apparatus further comprises: the characteristic extraction module is used for extracting the characteristics of stock creative materials in the creative library to obtain the characteristic vector of each stock creative material; the coding module is also used for coding the characteristic vectors of the stock creative materials to obtain material characteristic codes corresponding to the stock creative materials; the storage module is also used for storing the characteristic codes of all the materials in the creative library; the characteristic code searching module is used for searching a target material characteristic code matched with the video characteristic code in the creative library; and the similar material acquisition module is used for acquiring similar creative materials corresponding to the target material feature codes in the creative library.

In one embodiment, the apparatus further comprises: the verification code acquisition module is used for acquiring a verification code corresponding to the video material; the similar material query module is also used for querying creative materials based on the check codes and the characteristic codes to obtain similar creative materials matched with the video materials.

Specific limitations regarding the creative guidance information generation device can be referred to the above limitations regarding the creative guidance information generation method, and are not described herein again. The modules in the creative guidance information generation device can be wholly or partially implemented by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 11. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing the generation data of the creative guide information. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of generating creative guidance information.

Those skilled in the art will appreciate that the architecture shown in fig. 11 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is further provided, which includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method embodiments when executing the computer program.

In an embodiment, a computer-readable storage medium is provided, in which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.

In one embodiment, a computer program product or computer program is provided that includes computer instructions stored in a computer-readable storage medium. The computer instructions are read by a processor of a computer device from a computer-readable storage medium, and the computer instructions are executed by the processor to cause the computer device to perform the steps in the above-mentioned method embodiments.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

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

1.A method for generating creative guidance information, the method comprising:

acquiring a pre-delivered creative material and a delivery dimension of the creative material; the creative material is multimedia information displayed to a user;

based on the delivery dimension, inquiring similar creative materials matched with the creative materials;

acquiring a creative guidance information generation mode according to the delivery form of the creative material and the audience group type; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result;

and generating creative guidance information which contains the similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generation mode.

2. The method of claim 1, wherein obtaining pre-posted creative material and posting dimensions for the creative material comprises:

receiving uploaded pre-released creative materials;

displaying different candidate delivery dimensions in an creative guidance information generation interface;

responding to dimension screening operation, and screening the delivery dimensions of the creative materials from the candidate delivery dimensions displayed on the creative guidance information generation interface; the delivery dimension comprises at least one of delivery flow position information, delivery conversion types and target coverage crowd.

3. The method of claim 1, wherein the querying for similar creative material that matches the creative material based on the impression dimensions comprises:

4. The method of claim 3, wherein prior to obtaining pre-posted creative material and a posting dimension for the creative material, the method further comprises:

storing the creative delivery data in a creative library;

based on the delivery dimensions, similar creative materials matched with the creative materials are queried, including:

5. The method of claim 3, further comprising:

querying delivery effect data corresponding to the first creative material, the second creative material and the third creative material;

generating creative guidance information which contains the similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generating mode, wherein the creative guidance information generating mode comprises the following steps:

generating initial creative guide information matched with the creative dimension, the creative index and the delivery result based on the creative guide information generation mode;

and adding the similar creative materials serving as creative examples and the putting effect data into the initial creative guidance information to obtain the creative guidance information.

6. The method of claim 5, wherein generating initial creative guidance information that matches the creative dimension, the creative index, and the delivery result based on the creative guidance information generation manner comprises:

processing the similar creative materials by utilizing the creative guide information generation mode to generate creative guide information data;

and adding the creative guidance information data to the creative guidance information template to obtain initial creative guidance information.

7. The method of claim 6, further comprising:

in a creative guide information base, acquiring at least one piece of relevant historical creative guide information based on the delivery form and audience group type of the creative material;

coding the obtained calculation mode to obtain a data processing mode;

and forming the creative generation mode according to the delivery form and audience group type of the creative materials in the historical creative guidance information.

8. The method of claim 3, wherein merging the first creative material, the second creative material, and the third creative material resulting in similar creative materials that match the creative materials comprises:

respectively performing feature labeling on the first creative material, the second creative material and the third creative material;

determining creative material intersection among the marked first creative material, the marked second creative material and the marked third creative material;

and taking the intersection of the creative materials as similar creative materials matched with the creative materials.

9. The method of claim 1, wherein the creative material comprises video material; based on the delivery dimensions, similar creative materials matched with the creative materials are queried, including:

extracting a feature vector corresponding to each frame of image in the video material;

performing dimension reduction processing on the feature vector to obtain a dimension reduction feature vector of each frame of image;

10. The method of claim 9, wherein the feature vectors corresponding to each frame of image in the video material comprise a frame rate feature vector and a sharpness feature vector;

the performing feature transformation on the dimensionality reduction feature vector to obtain the feature vector of the video material comprises:

performing feature conversion processing on the dimensionality reduction frame rate feature vector and the dimensionality reduction definition feature vector to obtain a frame rate feature vector and a definition feature vector of the video material;

11. The method of claim 9 or 10, wherein the querying similar creative material that matches feature vectors of the video material comprises:

and inquiring creative materials according to the video characteristic codes to obtain similar creative materials matched with the video materials.

12. The method of claim 11, further comprising:

storing each material characteristic code in the creative library;

and inquiring creative material according to the video feature coding to obtain similar creative material matched with the video material, wherein the method comprises the following steps:

searching a target material feature code matched with the video feature code in the creative library;

and acquiring similar creative materials corresponding to the target material feature codes in the creative library.

13. The method of claim 11, further comprising:

acquiring a check code corresponding to the video material;

and inquiring creative materials based on the check codes and the feature codes to obtain similar creative materials matched with the video materials.

14. An apparatus for generating creative guide information, the apparatus comprising:

the creative material acquisition module is used for acquiring pre-delivered creative materials and delivery dimensions of the creative materials; the creative material is multimedia information displayed to a user;

the similar material query module is used for querying similar creative materials matched with the creative materials based on the release dimension;

the generation mode acquisition module is used for acquiring a creative guidance information generation mode according to the delivery form and audience group type of the creative material; the creative guidance information generation mode comprises a creative generation mode and a data processing mode obtained by coding a calculation mode based on the historical creative guidance information, wherein the calculation mode comprises a creative dimension, a creative index and a delivery result;

and the generating module is used for generating creative guidance information which contains the similar creative materials and is matched with the creative dimension, the creative index and the delivery result based on the creative guidance information generating mode.

15. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method of any one of claims 1 to 13 when executing the computer program.

16. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 13.

17. A computer program product comprising a computer program, characterized in that the computer program realizes the steps of the method of any one of claims 1 to 13 when executed by a processor.