Advertisement delivery strategy determination method, device, computer equipment and storage medium

Abstract

The present application relates to a method, apparatus, computer equipment and storage medium for determining an advertising delivery strategy. The method comprises: receiving an advertising delivery strategy determination message, wherein the advertising delivery strategy determination message carries data of products to be delivered; obtaining a target topic and characteristic data of the target topic based on the data of the products to be delivered; constructing a feature vector based on the characteristic data of the target topic; inputting the characteristic vector into a trained topic trend prediction model to obtain a trend prediction result of the target topic; determining an advertising delivery strategy based on the topic trend prediction result of the target topic; wherein the topic trend prediction model is trained based on historical topic data and historical topic characteristic data carrying topic classification labels, and the historical topic data corresponds to the same product field as the data of the products to be delivered. The use of this method can get rid of the practice of relying on human experience to predict hot topics, ensure the accuracy of topic trend prediction, and improve the effect of advertising delivery.

Description

Technical Field

[0001] The present application relates to the field of machine learning technology, and in particular to a method, apparatus, computer device, and storage medium for determining an advertising delivery strategy. Background Art

[0002] Social media is a tool and platform used by people to share opinions, insights, experiences and views with each other. Currently, it mainly includes social networking sites, Weibo, WeChat, blogs, forums, podcasts, etc.

[0003] Social media platforms often categorize user-posted information and identify trending topics. For example, on Weibo, a trending topic refers to a topic that has attracted widespread attention and discussion among users on the platform, with the number of topic-related Weibo posts reaching a certain threshold within a given period of time. The trending search list provided by Weibo can be used as a reference.

[0004] In advertising applications, there's a trend of preparing creatives and determining advertising strategies based on trending topics to improve conversion rates. However, current methods for determining advertising strategies based on trending topics primarily rely on manual experience to determine topic trends, which can't accurately predict them, leading to poor advertising effectiveness. Summary of the Invention

[0005] Based on this, it is necessary to provide a method, device, computer equipment and storage medium for determining an advertising delivery strategy that can improve the effectiveness of advertising delivery in response to the above technical problems.

[0006] A method for determining an advertisement delivery strategy, the method comprising:

[0007] Receiving an advertisement delivery strategy determination message, wherein the advertisement delivery strategy determination message carries product data to be delivered;

[0008] Obtain target topics and their characteristic data based on the product data to be launched;

[0009] Construct a feature vector based on the feature data of the target topic;

[0010] Input the feature vector into the trained topic trend prediction model to obtain the trend prediction result of the target topic;

[0011] Determine advertising strategies based on the topic trend prediction results of the target topic;

[0012] Among them, the topic trend prediction model is trained based on historical topic data and historical topic feature data carrying topic classification labels. The historical topic data and the data of the products to be launched correspond to the same product field.

[0013] In one embodiment, obtaining a target topic based on the product data to be launched includes:

[0014] Get topic keywords based on the product data to be launched;

[0015] Calling a first preset interface to access a third-party social networking platform, and performing a search on the third-party social networking platform based on the topic keyword to obtain a search result;

[0016] Get the target topic based on the query results and the preset comment count threshold.

[0017] In one embodiment, topic keywords obtained based on the product data to be launched include:

[0018] Calling a second preset interface to access a third-party data search platform, and performing a data search on the third-party data search platform according to the product field corresponding to the product data to be released, and selecting keywords in the data search results based on the data search results to obtain first topic keywords;

[0019] And / or, based on the product data to be launched, dynamic data related to the product field corresponding to the product data to be launched, posted by target social users, is obtained through a crawler mechanism, and keywords in the dynamic data are extracted to obtain second topic keywords.

[0020] In one embodiment, obtaining characteristic data of a target topic includes:

[0021] Search the third-party social networking platform based on the target topic to obtain user behavior data on the target topic;

[0022] Determine rules based on user behavior data and preset feature data to obtain feature data of the target topic.

[0023] In one embodiment, the characteristic data of the target topic includes the target group participation rate, comment rate, forwarding rate and / or topic discussion growth rate of the target topic within a preset unit time.

[0024] In one embodiment, before inputting the feature vector into a trained topic trend prediction model to obtain a trend prediction result for the target topic, the process further includes:

[0025] Obtaining historical topic data and feature information of historical topic data with topic classification labels;

[0026] Perform data standardization on the feature information of historical topic data and historical topic data with topic classification labels to obtain a model training set;

[0027] Obtaining an initial topic trend prediction model built based on a DNN (Deep Neural Networks) algorithm, the initial topic trend prediction model comprising a multi-layer neural network;

[0028] Based on the model training set, regression analysis is performed on the initial topic trend prediction model to determine the weights and intercepts of each layer of the neural network to obtain the trained topic trend prediction model.

[0029] In one embodiment, determining an advertisement delivery strategy based on a topic trend prediction result of a target topic includes:

[0030] If the topic trend prediction result of the target topic carries a hot topic classification label, the product to be launched will be marked as a launchable product and an advertising launch reminder message will be pushed.

[0031] An advertisement delivery strategy determination device includes a data acquisition unit and a data processing unit, wherein:

[0032] A data acquisition unit, configured to receive an advertisement delivery strategy determination message carrying product data to be delivered, and acquire a target topic and characteristic data of the target topic based on the product data to be delivered;

[0033] A data processing unit is used to construct a feature vector based on the feature data of the target topic, input the feature vector into a trained topic trend prediction model to obtain a trend prediction result of the target topic, and determine an advertising delivery strategy based on the topic trend prediction result of the target topic;

[0034] Among them, the topic trend prediction model is trained based on historical topic data and historical topic feature data carrying topic classification labels. The historical topic data and the data of the products to be launched correspond to the same product field.

[0035] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0036] Receiving an advertisement delivery strategy determination message, wherein the advertisement delivery strategy determination message carries product data to be delivered;

[0037] Obtain target topics and their characteristic data based on the product data to be launched;

[0038] Construct a feature vector based on the feature data of the target topic;

[0039] Input the feature vector into the trained topic trend prediction model to obtain the trend prediction result of the target topic;

[0040] Determine advertising strategies based on the topic trend prediction results of the target topic;

[0041] Among them, the topic trend prediction model is trained based on historical topic data and historical topic feature data carrying topic classification labels. The historical topic data and the data of the products to be launched correspond to the same product field.

[0042] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0043] Receiving an advertisement delivery strategy determination message, wherein the advertisement delivery strategy determination message carries product data to be delivered;

[0044] Obtain target topics and their characteristic data based on the product data to be launched;

[0045] Construct a feature vector based on the feature data of the target topic;

[0046] Input the feature vector into the trained topic trend prediction model to obtain the trend prediction result of the target topic;

[0047] Determine advertising strategies based on the topic trend prediction results of the target topic;

[0048] Among them, the topic trend prediction model is trained based on historical topic data and historical topic feature data carrying topic classification labels. The historical topic data and the data of the products to be launched correspond to the same product field.

[0049] The above-mentioned advertising delivery strategy determination method, apparatus, computer device, and storage medium, upon receiving an advertising delivery strategy determination message carrying product data to be delivered, can automatically obtain the target topic and its characteristic data based on the product data to be delivered, construct a feature vector, and input the feature vector into a trained topic trend prediction model to obtain a trend prediction result for the target topic. The above-mentioned scheme can obtain the target topic through an automated process based on the product data to be delivered, and obtain an objective and accurate trend prediction result for the target topic. This fundamentally breaks away from the practice of relying on human experience to predict topic trends, ensures the accuracy of topic trend predictions, and thus can improve the effectiveness of advertising delivery. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A diagram illustrating an application environment of a method for determining an advertisement delivery strategy in one embodiment;

[0051] Figure 2 A flowchart of a method for determining an advertisement delivery strategy in one embodiment is shown;

[0052] Figure 3A flowchart of the steps of obtaining a target topic in one embodiment;

[0053] Figure 4 A flowchart illustrating steps for training a topic trend prediction model in another embodiment;

[0054] Figure 5 This is a structural block diagram of an apparatus for determining an advertisement delivery strategy in one embodiment;

[0055] Figure 6 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0057] The advertising delivery strategy determination method provided in this application can be applied to Figure 1 In the application environment shown. The terminal 102 communicates with the server 104 via a network. Specifically, the user can operate at the terminal 102 and send an advertisement delivery strategy determination message carrying the product data to be delivered to the server 104. The server 104 responds to the advertisement delivery strategy determination message, obtains the target topic and the feature data of the target topic based on the product data to be delivered, and constructs a feature vector based on the feature data of the target topic. Then, the feature vector is input into the trained topic trend prediction model to obtain the trend prediction result of the target topic. Finally, based on the topic trend prediction result of the target topic, the advertisement delivery strategy is determined, wherein the topic trend prediction model is trained based on historical topic data and historical topic feature data carrying topic classification labels, and the historical topic data corresponds to the same product field as the product data to be delivered. The terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers and portable wearable devices, and the server 104 can be implemented as an independent server or a server cluster consisting of multiple servers.

[0058] In one embodiment, Figure 2 As shown, a method for determining an advertisement delivery strategy is provided, and the method is applied to Figure 1 The following steps are used as an example to illustrate the server in the example:

[0059] Step 202: Receive an advertisement delivery strategy determination message, which carries product data to be delivered.

[0060] Product data refers to the products advertisers wish to promote through advertising. This data includes the product name, product type, industry, and product. In practice, a user can enter this data through a terminal interface, such as a mobile phone, and then the terminal sends an advertising strategy confirmation message containing this data.

[0061] Step 204: Obtain target topics and feature data of the target topics based on the data of the product to be launched.

[0062] Target topics are key discussion topics related to the product area of ​​the product data to be launched. Feature data for target topics includes data such as topic discussion growth rate, comment rate, and forwarding rate. Once the product data to be launched is obtained, the product area of ​​the product data to be launched can be analyzed. Using third-party data platforms and keyword extraction technology, key topics related to the product area can be identified, and then feature data for the target topics can be obtained.

[0063] Step 206: construct a feature vector based on the feature data of the target topic.

[0064] After obtaining the feature data of the target topic, the feature data may be standardized to construct a feature vector.

[0065] In step 208, the feature vector is input into the trained topic trend prediction model to obtain the trend prediction result of the target topic, wherein the topic trend prediction model is trained based on historical topic data and historical topic feature data carrying topic classification labels, and the historical topic data and the product data to be launched correspond to the same product field.

[0066] Topic classification labels include hot topic labels and non-hot topic labels. In specific implementation, developers can build a training sample set based on historical topic data and historical topic feature data with topic classification labels, and train a topic trend prediction model based on the training sample set, where the historical topic data and the product data to be launched correspond to the same product field. The purpose of this model is to monitor the changing trend of the target topic's attention within one week of its release, and to predict whether the topic will become a hot topic within a preset time period based on the changing trend. The topic is then divided into hot topics and non-hot topics to obtain the topic trend prediction result of the target topic.

[0067] Step 210: Determine an advertisement delivery strategy based on the topic trend prediction result of the target topic.

[0068] The advertising delivery strategy refers to determining whether to deliver advertisements based on the topic trend prediction results of the target topic. In specific implementation, the release time of the topic can be used as the starting time, and the characteristic data of the topic within one week of the starting time can be counted to monitor the changing trend of the topic's attention (number of followers, discussion growth rate, etc.), and predict whether the topic will become a hot topic one week later based on the changing trend. The topic trend prediction results of the target topic include whether it will become a hot topic or not in the future preset time period. If the topic trend prediction results indicate that the product to be launched can become a hot topic in the preset time period, the product to be launched can be screened out or marked as a product that can be launched, so as to prepare advertising materials for the product that can be launched and conduct customized advertising delivery.

[0069] In the above-mentioned method for determining an advertising delivery strategy, upon receiving an advertising delivery strategy determination message carrying the product data to be delivered, the method can automatically obtain the target topic and its characteristic data based on the product data to be delivered, construct a feature vector, and input the feature vector into a trained topic trend prediction model to obtain a trend prediction result for the target topic. Furthermore, the advertising delivery strategy can be determined based on the topic trend prediction result for the target topic. The above-mentioned scheme can obtain the target topic through an automated process based on the product data to be delivered, and obtain objective and accurate topic trend prediction results. This fundamentally breaks away from the practice of relying on human experience to predict topic trends, ensures the accuracy of topic trend predictions, and thus improves the effectiveness of advertising delivery.

[0070] like Figure 3 As shown, in one embodiment, obtaining a target topic based on the product data to be launched includes:

[0071] Step 222: Obtain topic keywords based on the product data to be launched;

[0072] Step 242: calling a first preset interface to access a third-party social networking platform, and performing a search on the third-party social networking platform based on the topic keyword to obtain a search result;

[0073] Step 262: Obtain the target topic based on the query result and a preset comment number threshold.

[0074] Topic keywords refer to representative keywords in the topic. As described in the above embodiment, the product data to be released includes data such as the product field and product name to which the product belongs. In this embodiment, it can be based on the product field to obtain the key discussion topics in the field. Specifically, it can be based on the product field to obtain topic keywords, and then call a preset interface to access a third-party social networking platform such as Weibo. Then, based on the topic keywords, search and query on Weibo to obtain social dynamic data related to the product field, and then select the dynamic data with the number of comments exceeding a preset comment threshold such as 1,000 and the highest ranking, and obtain the target topic based on the selected dynamic data. In this embodiment, obtaining the target topic based on the third-party social networking platform is low-cost and highly timely, and the screened target topic is representative.

[0075] In one embodiment, step 222 includes: calling a second preset interface to access a third-party data search platform, and performing a data search on the third-party data search platform according to the product field corresponding to the product data to be launched, obtaining data search results, selecting keywords in the data search results, and obtaining first topic keywords; and / or, based on the product data to be launched, obtaining dynamic data related to the product field corresponding to the product data to be launched posted by target social users through a crawler mechanism, extracting keywords in the dynamic data, and obtaining second topic keywords.

[0076] During specific implementation, the determination of topic keywords can be based on Baidu's search index and / or domain big V (big V refers to users who obtain personal certification on microblog platforms such as Sina, Tencent, NetEase, and have many fans) speech or dynamic data to extract keywords. Specifically, it can be to call the second preset interface to access a third-party data search platform such as Baidu Index, search for keywords in the product field in the search index disclosed daily by Baidu Index, obtain data search results, select the top 5 keywords in the data search results, and determine them as the first topic keywords. And / or, it can be based on the scrapy crawler mechanism, on major social network platforms such as microblog, search for big V's microblog in the product field, and then combine keyword extraction technology to extract keywords of microblog related content with the highest comment volume on the same day. For example, a big V microblog in a certain mobile phone field publishes a performance test microblog for a certain unlisted mobile phone, and the number of comments is high. The threshold value of the number of comments can be set, and the microblog dynamics whose number of comments exceeds the preset comment number threshold value are selected to extract the key topic keywords in the field. In the present embodiment, the third-party platform is accessed by calling the preset interface, and then combined with the keyword extraction technology, the topic keywords of the target topic are obtained, which also has the characteristics of low cost, and the determined topic keywords are representative.

[0077] In one embodiment, obtaining characteristic data of the target topic includes: querying a third-party social network platform according to the target topic to obtain user behavior data of the target topic; and determining rules according to the user behavior data and preset characteristic data to obtain the characteristic data of the target topic.

[0078] User behaviors include forwarding with comments, forwarding, commenting, and original creation. In specific implementations, the characteristic data for a target topic can be obtained by querying Weibo based on the target topic, obtaining user behavior data related to the target topic, including the number of user comments, comment time, number of user forwarding, forwarding time, and original dynamic posting time and original creation time. Then, combined with preset characteristic data determination rules, the characteristic data for the target topic is statistically obtained. In this embodiment, the preset characteristic data determination rules and the third-party social network platform can be used to quickly obtain the characteristic data for the target topic.

[0079] In another embodiment, the characteristic data of the target topic includes the target group participation rate, comment rate, forwarding rate and / or topic discussion growth rate of the target topic within a preset unit time.

[0080] The target group participation rate may include the opinion leader participation rate and the public opinion leader participation rate. Opinion leaders play a key role in two-step communication. They are the first or most exposed to mass media information in the crowd and disseminate this processed information to others. Opinion leaders are generally considered to be charismatic individuals with strong comprehensive abilities and a high social status or sense of identity. They are active in social situations, belong to the same group and share common interests with those influenced by them, are knowledgeable about specific issues, and are willing to accept and disseminate relevant information. In this embodiment, the characteristic data may be the opinion leader participation rate, Weibo comment rate, Weibo forwarding rate, and / or topic discussion growth rate for the target topic within a week.

[0081] (1) The opinion leader participation rate refers to the percentage of opinion leaders participating in hot topic discussions per unit time (hour) compared to the total number of opinion leaders. The participation of opinion leaders directly expands the scope of hot topic dissemination. The higher the participation rate, the larger the user group affected, and the number of ordinary users and Weibo accounts participating in the discussion will also increase, which in turn affects the development trend of the topic.

[0082] (2) The Weibo comment rate refers to the percentage of the total number of comments on Weibo related to a topic to the total number of comments on all Weibo posts within a unit of time (per hour). Among the four types of Weibo user behaviors (retweet with comment, retweet, comment, and original post), the cost of user participation in commenting is the lowest. The number of comments directly indicates the degree of attention paid to the Weibo post and the intensity of user discussions. The more comments a Weibo post has, the more intense the user discussion, the greater the user attention, and the wider the scope of dissemination. Therefore, the Weibo comment rate can also be considered as one of the factors affecting the development trend of a topic.

[0083] (3) The Weibo forwarding rate refers to the percentage of the sum of the number of forwardings of Weibo posts related to a topic published within a unit of time (per hour) to the sum of the number of forwardings of all Weibo posts. The Weibo forwarding rate reflects the level of attention a topic receives on the Weibo platform. The higher the forwarding rate of an opinion leader’s Weibo post, the more users are following the topic. Conversely, the lower the forwarding rate of a hot topic, the more users are beginning to shift their attention to the topic.

[0084] (4) The topic discussion growth rate is to first calculate the growth rate of the number of times the topic is discussed per hour compared with the previous hour, and then take the average growth rate of the day.

[0085] In this embodiment, the target group participation rate, comment rate, forwarding rate and / or topic discussion growth rate of the target topic within a preset unit time are selected as representative feature data, which can make the topic trend prediction results obtained based on the feature numbers more accurate.

[0086] like Figure 4 As shown, in one embodiment, before inputting the feature vector into the trained topic trend prediction model to obtain the trend prediction result of the target topic, the following steps are further included:

[0087] Step 207: Acquire historical topic data and feature information of the historical topic data carrying topic classification tags;

[0088] Step 227: performing data normalization processing on the historical topic data and the feature information of the historical topic data carrying topic classification labels to obtain a model training set;

[0089] Step 247: Obtain an initial topic trend prediction model constructed based on a DNN algorithm, where the initial topic trend prediction model includes a multi-layer neural network.

[0090] Step 267: Based on the model training set, perform regression analysis on the initial topic trend prediction model to determine the weights and intercepts of each layer of the neural network to obtain a trained topic trend prediction model.

[0091] Specifically, a corresponding topic trend prediction model can be trained for each product category (such as electronic products, cosmetics, and daily necessities). In this embodiment, taking electronic products as an example, hot topics and non-hot topics related to the electronic products field can be obtained from the Weibo platform, and the feature data of hot topics and non-hot topics can be counted. Then, a "hot topic" classification label is added to the feature data of hot topics, and a "non-hot topic" classification label is added to the feature data of non-hot topics. The classification basis for hot topics and non-hot topics is whether the topic enters the top 100 Weibo hot search list one week after the topic is released. If it enters the top 100, it is a hot topic; otherwise, it is a non-hot topic. It is understandable that the specific criteria for determining hot topics can be set according to actual conditions. Based on the labeled feature data and topic data, a training sample of about 100 is constructed. Then, the training sample is subjected to data normalization processing to obtain a model training set. Then, the DNN algorithm is used to construct an initial topic trend prediction model. The model is a multi-layer neural network consisting of an input layer and an output layer, with each layer having a weight value and an intercept term parameter. Based on the standardized historical topic data and feature information carrying topic classification labels, a historical feature vector is constructed. With the historical feature vector as input and the topic classification label as the prediction variable, i.e., the output, the DNNRegressor provided by the TensorFlow software is used for regression analysis. The model is fitted to determine the weights and intercepts of each layer of the neural network that minimizes the error. The weights and intercepts of the model are saved to complete the model training. In this embodiment, model training is performed based on historical topic data and feature data from the same product field, which can make the application of the model more in line with actual needs and improve the accuracy of topic trends.

[0092] In one embodiment, determining an advertising delivery strategy based on the topic trend prediction result of the target topic includes: if the topic trend prediction result of the target topic carries a hot topic classification tag, marking the product to be delivered as a deliverable product and pushing an advertising delivery prompt message.

[0093] A product that can be put on the market refers to a product that can be advertised. In specific implementation, if the trend prediction result of the target topic carries a "hot topic" classification label, indicating that the target topic can become a hot topic within a preset time period in the future, such as within a week, then the product to be put on the target topic corresponding to the target topic is marked as a product that can be put on the market, and an advertising placement prompt message is pushed. The advertising placement prompt message can be sent to the user terminal or the terminal or mailbox of a designated person in the form of a pop-up window or a text message. If the trend prediction result of the target topic carries a "non-hot topic" classification label, indicating that the target topic cannot become a hot topic within a preset time period in the future, such as within a week, then the product to be put on the target topic corresponding to the target topic is skipped, and advertising is not placed for the product to be put on the market. The topic trend of the next product to be put on the market is predicted. Furthermore, a prompt message that the product to be put on the market does not meet the advertising placement requirements can be output. In this embodiment, by pushing an advertising placement prompt message, relevant personnel can be notified in a timely manner to prepare advertising materials and adjust marketing strategies.

[0094] It should be understood that, although the various steps in the various flow charts that the above-described embodiments relate to are shown in sequence according to the indications of the arrows, these steps are not necessarily performed in sequence according to the order indicated by the arrows. Unless clearly stated herein, the execution of these steps is not strictly limited in order, and these steps can be performed in other orders. Moreover, at least a portion of the steps in the various flow charts that the above-described embodiments relate to can include multiple steps or multiple stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of the steps or stages in other steps or other steps.

[0095] In one embodiment, Figure 5 As shown, an advertisement delivery strategy determination device is provided, including a data acquisition unit 510 and a data processing unit 520, wherein:

[0096] The data acquisition unit 510 is configured to receive an advertisement delivery strategy determination message carrying product data to be delivered, and acquire a target topic and characteristic data of the target topic based on the product data to be delivered.

[0097] The data processing unit 520 is used to construct a feature vector based on the feature data of the target topic, input the feature vector into the trained topic trend prediction model to obtain the trend prediction result of the target topic, and determine the advertising delivery strategy based on the topic trend prediction result of the target topic;

[0098] Among them, the topic trend prediction model is trained based on historical topic data and historical topic feature data carrying topic classification labels. The historical topic data and the data of the products to be launched correspond to the same product field.

[0099] Upon receiving an advertisement placement strategy determination message carrying product data to be placed, the aforementioned advertisement placement strategy determination device can automatically acquire the target topic and its characteristic data based on the product data to be placed, construct a feature vector, and input the feature vector into a trained topic trend prediction model to obtain a trend prediction result for the target topic. This device can then determine the advertisement placement strategy based on the topic trend prediction result for the target topic. This solution, based on the product data to be placed, can acquire the target topic through an automated process and obtain an objective and accurate trend prediction result for the target topic. This fundamentally eliminates the practice of relying on human experience to predict topic trends, ensures the accuracy of topic trend predictions, and thus improves the effectiveness of advertisement placement.

[0100] The above solution can obtain target topics through an automated process based on the data of the products to be launched, and obtain objective and accurate trend prediction results of the topics. It fundamentally gets rid of the practice of relying on human experience to predict topic trends, ensures the accuracy of topic trend prediction, and thus can improve the effectiveness of advertising.

[0101] In one embodiment, the data acquisition unit 510 is also used to obtain topic keywords based on the product data to be launched, call the first preset interface to access the third-party social networking platform, and query the third-party social networking platform based on the topic keywords to obtain query results, and obtain the target topic based on the query results and a preset comment number threshold.

[0102] In one embodiment, the data acquisition unit 510 is also used to call a second preset interface to access a third-party data search platform, and perform a data search on the third-party data search platform according to the product field corresponding to the product data to be launched, obtain data search results, select keywords in the data search results, and obtain first topic keywords; and / or, based on the product data to be launched, through a crawler mechanism, obtain dynamic data related to the product field corresponding to the product data to be launched posted by target social users, extract keywords in the dynamic data, and obtain second topic keywords.

[0103] In one embodiment, the data acquisition unit 510 is further configured to query a third-party social networking platform based on the target topic, obtain user behavior data of the target topic, determine rules based on the user behavior data and preset feature data, and obtain feature data of the target topic.

[0104] In one embodiment, the data processing unit 520 is also used to obtain feature information of historical topic data and historical topic data carrying topic classification labels, perform data standardization on the feature information of historical topic data and historical topic data carrying topic classification labels, obtain a model training set, and obtain an initial topic trend prediction model constructed based on the DNN algorithm. The initial topic trend prediction model includes a multi-layer neural network. Based on the model training set, the initial topic trend prediction model is subjected to regression analysis to determine the weights and intercepts of each layer of the neural network to obtain a trained topic trend prediction model.

[0105] In one embodiment, the data processing unit 520 is further configured to mark the product to be launched as a launchable product and push an advertisement launch prompt message if the topic trend prediction result of the target topic carries a hot topic classification tag.

[0106] Specific embodiments of the advertising delivery strategy determination apparatus can be found in the embodiments of the advertising delivery strategy determination method described above and will not be further elaborated here. Each module in the aforementioned advertising delivery strategy determination apparatus may be implemented in whole or in part via software, hardware, or a combination thereof. Each of these modules may be embedded in or independent of a processor within a computer device in hardware form, or may be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0107] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, a memory and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile 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 the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store product data to be launched and feature data of target topics, etc. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for determining an advertising delivery strategy is implemented.

[0108] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0109] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above-mentioned method for determining an advertisement delivery strategy when executing the computer program.

[0110] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method for determining an advertisement delivery strategy are implemented.

[0111] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may 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 or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0112] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0113] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A method for determining an advertisement delivery strategy, characterized in that: The method comprises: Receiving an advertisement delivery strategy determination message, wherein the advertisement delivery strategy determination message carries product data to be delivered; Obtaining a target topic and characteristic data of the target topic based on the product data to be launched, wherein the characteristic data of the target topic includes a target group participation rate, comment rate, forwarding rate and / or topic discussion growth rate of the target topic within a preset unit time; Constructing a feature vector based on the feature data of the target topic; Inputting the feature vector into a trained topic trend prediction model to obtain a trend prediction result of the target topic; If the topic trend prediction result of the target topic carries a hot topic classification tag, the product to be launched is marked as a product that can be launched, and an advertisement launch prompt message is pushed; A topic trend prediction model is trained for each product field. The topic trend prediction model is trained based on the following method: Acquire historical topic data and feature information of the historical topic data carrying topic classification tags, wherein the historical topic data and the product data to be launched correspond to the same product field, and the historical topic data includes historical hot topic data and historical non-hot topic data; Performing data standardization on the historical topic data and the feature information of the historical topic data carrying the topic classification labels to obtain a model training set; Obtaining an initial topic trend prediction model constructed based on a DNN algorithm, wherein the initial topic trend prediction model includes a multi-layer neural network; Based on the model training set, regression analysis is performed on the initial topic trend prediction model to determine the weights and intercepts of each layer of the neural network to obtain a trained topic trend prediction model.

2. The method for determining an advertisement delivery strategy according to claim 1, wherein: According to the product data to be launched, obtaining target topics includes: Obtain topic keywords based on the product data to be launched; Calling a first preset interface to access a third-party social networking platform, and performing a query on the third-party social networking platform based on the topic keyword to obtain a query result; The target topic is obtained according to the query result and a preset comment number threshold.

3. The method for determining an advertisement delivery strategy according to claim 2, wherein: The topic keywords obtained based on the product data to be launched include: Calling a second preset interface to access a third-party data search platform, and performing a data search on the third-party data search platform according to the product field corresponding to the product data to be launched, obtaining data search results, and selecting keywords from the data search results to obtain first topic keywords; And / or, based on the product data to be launched, dynamic data related to the product field corresponding to the product data to be launched, published by target social users, is obtained through a crawler mechanism, and keywords in the dynamic data are extracted to obtain second topic keywords.

4. The method for determining an advertisement delivery strategy according to claim 2, wherein: Acquiring the characteristic data of the target topic includes: Performing a query on the third-party social networking platform based on the target topic to obtain user behavior data on the target topic; A rule is determined based on the user behavior data and preset feature data to obtain feature data of the target topic.

5. An advertisement delivery strategy determination device, characterized in that: The device comprises a data acquisition unit and a data processing unit, wherein: The data acquisition unit is configured to receive an advertisement delivery strategy determination message carrying product data to be delivered, and acquire a target topic and characteristic data of the target topic based on the product data to be delivered, wherein the characteristic data of the target topic includes a target group participation rate, comment rate, forwarding rate, and / or topic discussion growth rate of the target topic within a preset unit time; The data processing unit is configured to construct a feature vector based on the feature data of the target topic, input the feature vector into a trained topic trend prediction model, obtain a trend prediction result of the target topic, and if the topic trend prediction result of the target topic carries a hot topic classification label, mark the product to be launched as a product that can be launched, and push an advertisement launch prompt message; The data processing unit is further used to obtain historical topic data and characteristic information of the historical topic data carrying topic classification labels, wherein the historical topic data and the data of the product to be launched correspond to the same product field, and the historical topic data include historical hot topic data and historical non-hot topic data; perform data standardization processing on the characteristic information of the historical topic data and the historical topic data carrying topic classification labels to obtain a model training set; obtain an initial topic trend prediction model constructed based on the DNN algorithm, wherein the initial topic trend prediction model includes a multi-layer neural network; based on the model training set, perform regression analysis on the initial topic trend prediction model to determine the weights and intercepts of each layer of the neural network to obtain a trained topic trend prediction model.

6. The device according to claim 5, characterized in that The data acquisition unit is further configured to obtain topic keywords based on the product data to be launched, call a first preset interface to access a third-party social networking platform, perform a query on the third-party social networking platform based on the topic keywords, obtain a query result, and obtain the target topic based on the query result and a preset comment number threshold.

7. The device according to claim 6, characterized in that The data acquisition unit is further configured to call a second preset interface to access a third-party data search platform, perform a data search on the third-party data search platform based on the product field corresponding to the data of the product to be launched, obtain data search results, select keywords in the data search results, and obtain first topic keywords; and / or, based on the data of the product to be launched, obtain dynamic data related to the product field corresponding to the data of the product to be launched posted by target social users through a crawler mechanism, extract keywords in the dynamic data, and obtain second topic keywords.

8. The device according to claim 5, characterized in that The data acquisition unit is further configured to query a third-party social networking platform based on the target topic to obtain user behavior data of the target topic, determine rules based on the user behavior data and preset feature data, and obtain feature data of the target topic.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images