Method and device for optimizing click effect of advertisement platform based on artificial intelligence

Abstract

The invention relates to a method and device for optimizing the click effect of an advertisement platform based on artificial intelligence, and belongs to the technical field of advertisement platform click effect optimization, and the method comprises the steps: carrying out the feature extraction of an image advertisement material and a text advertisement material through a feature extraction algorithm, and obtaining an image feature vector and a text feature vector; the classification algorithm classifies the image advertisement material and the text advertisement material according to the image feature vector and the text feature vector to obtain a classification result; acquiring historical behavior data of the user on the advertisement materials, and analyzing preferences of the user on different types of advertisement materials through a deep learning algorithm to generate a user preference model; acquiring real-time behavior data, and adjusting the display frequency and sequence of the advertisement materials according to the real-time behavior data and the user preference model; and adjusting the advertisement material display frequency, display position and material content according to the user preference model and the classification result. According to the method, the advertisement material click rate, the personalized recommendation effect and the advertisement material putting efficiency are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of optimizing click effects of advertising platforms, and in particular, relates to a method and device for optimizing click effects of advertising platforms based on artificial intelligence. Background Art

[0002] Ad click optimization is crucial in the field of digital marketing. The click-through rate of ads directly affects the effectiveness of ads and the advertiser's revenue. Currently, ad click optimization methods mainly improve click-through rates through user behavior analysis, ad delivery strategy optimization, and other means. Common technical means include recommendation systems based on user behavior, dynamic ad delivery strategies, A / B testing, and so on. However, these methods mainly focus on delivery strategies and user analysis, and pay less attention to the optimization of the ad creatives themselves.

[0003] In the prior art, some advertising platforms have begun to pay attention to the impact of advertising materials on click-through effects. For example, some platforms will adjust the display frequency of advertisements based on the historical performance data of advertising materials. However, this method has significant limitations, which are mainly manifested in the following aspects:

[0004] First, the homogeneity of advertising materials: Existing technologies classify advertising materials based on simple tags or keywords, resulting in a relatively rough classification result that cannot be refined to the specific features of the materials, such as color, composition, text content, etc. This rough classification method makes it impossible to accurately match different categories of materials to target users when displaying, resulting in a low click-through rate.

[0005] Second, lack of dynamic optimization mechanism: Current advertising material optimization is mostly based on static historical data, lacking a real-time dynamic adjustment mechanism. The performance of advertising materials is affected by many factors, such as time, season, hot events, etc. Static optimization methods cannot reflect these changes in time, resulting in poor optimization results.

[0006] Third, user preferences are not fully considered: Existing ad click optimization methods rely more on overall data analysis and lack in-depth mining and analysis of individual user preferences. Different users have different preferences for advertising materials, and a unified optimization strategy cannot meet the needs of individual users, affecting the personalized effect of advertising.

[0007] To sum up, the existing advertising material classification and click optimization methods have great deficiencies in classification precision, dynamic optimization capabilities and user preference considerations. There is an urgent need for a more refined, dynamic and personalized advertising material classification and click optimization method to improve the click effect of advertisements. Summary of the invention

[0008] In view of the above-mentioned deficiencies in the prior art, the purpose of the invention is to provide a method and device for optimizing click effects of an advertising platform based on artificial intelligence, so as to solve the problems that the advertising material classification and click optimization methods in the prior art have major deficiencies in classification precision, dynamic optimization capabilities and user preference considerations.

[0009] The first aspect of the present invention provides a method for optimizing click effects of an advertising platform based on artificial intelligence, comprising:

[0010] The feature extraction algorithm extracts features from the image advertising material and the text advertising material to obtain an image feature vector and a text feature vector;

[0011] The classification algorithm classifies the image advertising material and the text advertising material according to the image feature vector and the text feature vector to obtain a classification result;

[0012] Obtain historical user behavior data on advertising materials, and generate a user preference model by analyzing user preferences for different types of advertising materials through deep learning algorithms;

[0013] Obtain real-time user behavior data on advertising materials, and adjust the display frequency and order of advertising materials based on the real-time behavior data and user preference model;

[0014] Adjust the frequency, placement, and content of ad creatives based on user preference models and classification results.

[0015] Furthermore, in the above method for optimizing click effects of an advertising platform based on artificial intelligence, the feature extraction algorithm extracts features of the image advertising material to obtain an image feature vector, including:

[0016] Input the image creative into the convolutional neural network to extract high-level features of the image creative through convolutional layers, activation functions, and pooling layers;

[0017] The image feature vector is obtained by outputting a vector of high-level features of the image creative through a fully connected layer.

[0018] Furthermore, in the above-mentioned method for optimizing click effects of an advertising platform based on artificial intelligence, the feature extraction algorithm extracts features from text advertising materials to obtain text feature vectors, including:

[0019] Input the text advertising material into the recurrent neural network or the long short-term memory network and convert the text advertising material into a vector representation through the embedding layer;

[0020] Extracting text advertising creative features through multi-layer recurrent neural network units or long short-term memory network units;

[0021] The text feature vector is obtained by outputting the vector of text advertising material features through the fully connected layer.

[0022] Furthermore, in the above-mentioned method for optimizing click effects of an advertising platform based on artificial intelligence, the classification algorithm classifies the image advertising material and the text advertising material according to the image feature vector and the text feature vector to obtain the classification result, including:

[0023] Input the image feature vector and the text feature vector into the deep neural network;

[0024] The classification results are output through multiple layers of fully connected layers and non-linear activation functions.

[0025] Furthermore, in the above method for optimizing click effects of an advertising platform based on artificial intelligence, historical behavior data of users on advertising materials is obtained, and a user preference model is generated by analyzing the user's preference for different types of advertising materials through a deep learning algorithm, including:

[0026] Obtain the user's historical behavior data on the advertising material, and build a user profile based on the user's historical behavior data on the advertising material;

[0027] Analyze users’ preferences for different types of advertising materials through collaborative filtering algorithms and neural networks to generate user preference models;

[0028] The user's historical behavior data on the advertising material includes at least browsing history, click records and interactive behavior.

[0029] Furthermore, in the above method for optimizing click effects of an advertising platform based on artificial intelligence, the display frequency and order of advertising materials are adjusted according to real-time behavior data and user preference models, including:

[0030] The Q-learning algorithm continuously updates the state-action value based on real-time behavior data and user preference models;

[0031] Adjust the frequency and order of creative display based on the updated status-action value.

[0032] Furthermore, the above method for optimizing click effects of an advertising platform based on artificial intelligence also includes:

[0033] Periodically compare the click-through rate and conversion rate of the creatives before and after the adjustments;

[0034] Determine whether to improve the feature extraction algorithm, classification algorithm, or adjust the display frequency, sequence, placement, and content of the creatives based on the click-through rate and conversion rate.

[0035] The second aspect of the present invention further proposes a device for optimizing click effects of an advertising platform based on artificial intelligence, comprising:

[0036] Feature extraction module: used for feature extraction algorithm to extract features of image advertising materials and text advertising materials respectively to obtain image feature vectors and text feature vectors;

[0037] Classification module: used for the classification algorithm to classify the image advertising material and the text advertising material according to the image feature vector and the text feature vector to obtain the classification result;

[0038] Generation module: used to obtain historical behavior data of users on advertising materials, and generate a user preference model by analyzing user preferences for different types of advertising materials through deep learning algorithms;

[0039] The first adjustment module is used to obtain real-time behavior data of users on advertising materials, and adjust the display frequency and order of advertising materials according to the real-time behavior data and the user preference model;

[0040] The second adjustment module is used to adjust the display frequency, display position and material content of the advertising materials according to the user preference model and classification results.

[0041] According to a third aspect of the present invention, an electronic device is provided, comprising: a processor and a memory;

[0042] The processor calls the program or instruction stored in the memory to execute any of the above methods for optimizing click effects of an advertising platform based on artificial intelligence.

[0043] In a fourth aspect of the present invention, a computer-readable storage medium is proposed, which stores a program or instruction, and the program or instruction enables a computer to execute any one of the above methods for optimizing click effects of an artificial intelligence-based advertising platform.

[0044] The beneficial effects of the present invention are as follows: the present invention uses a feature extraction algorithm to extract features from image advertising materials and text advertising materials respectively to obtain image feature vectors and text feature vectors; the classification algorithm classifies image advertising materials and text advertising materials according to image feature vectors and text feature vectors to obtain classification results; obtains historical behavior data of users on advertising materials, and generates a user preference model by analyzing user preferences for different types of advertising materials through a deep learning algorithm; obtains real-time behavior data of users on advertising materials, and adjusts the display frequency and sequence of advertising materials according to the real-time behavior data and the user preference model; adjusts the display frequency, display position and material content of advertising materials according to the user preference model and classification results. The present invention improves the click-through effect of advertisements through refined material classification and dynamic optimization strategies, fully considers the personalized needs of users, and realizes the precision and dynamic optimization of advertising delivery, thereby improving the click-through effect and conversion rate of advertisements. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings are only used to illustrate specific embodiments and are not considered to limit the present invention. In the entire drawings, the same reference symbols represent the same components. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0046] Figure 1 A method diagram for optimizing click effects of an advertising platform based on artificial intelligence provided by an embodiment of the present invention;

[0047] Figure 2 A method diagram for extracting features to obtain image feature vectors provided by an embodiment of the present invention;

[0048] Figure 3 A method diagram for extracting features to obtain text feature vectors provided by an embodiment of the present invention;

[0049] Figure 4 A diagram of a method for classifying advertising materials to obtain classification results provided by an embodiment of the present invention;

[0050] Figure 5 A method diagram for generating a user preference model provided by an embodiment of the present invention;

[0051] Figure 6 A diagram of a device for optimizing click effects of an advertising platform based on artificial intelligence provided by an embodiment of the present invention;

[0052] Figure 7 A schematic block diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making creative work should fall within the scope of protection of the present invention.

[0054] Furthermore, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessarily obscuring the concepts disclosed in the present invention.

[0055] In the description of the present invention, the terms "first", "second", and "third" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. The terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be internal communication between two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0056] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of methods and systems consistent with some aspects of the present invention as detailed in the appended claims.

[0057] The present invention proposes a method and device for optimizing click effects of an advertising platform based on artificial intelligence, which significantly improves the click-through rate of advertising materials, personalized recommendation effects, and advertising material delivery efficiency.

[0058] Method Embodiment

[0059] Figure 1 A diagram of a method for optimizing click effects of an advertising platform based on artificial intelligence provided in an embodiment of the present invention.

[0060] In the first aspect of the present invention, a method for optimizing click effects of an advertising platform based on artificial intelligence is proposed, combining Figure 1 , including five steps from S1 to S5:

[0061] S1: The feature extraction algorithm extracts features from the image advertising material and the text advertising material to obtain an image feature vector and a text feature vector.

[0062] Specifically, in an embodiment of the present invention, the feature extraction algorithm uses a convolutional neural network to perform feature extraction on image advertising materials to obtain an image feature vector, and uses a recurrent neural network or a long short-term memory network to perform feature extraction on text advertising materials to obtain a text feature vector. The specific extraction method is described in detail below.

[0063] S2: The classification algorithm classifies the image advertising material and the text advertising material according to the image feature vector and the text feature vector to obtain a classification result.

[0064] Specifically, in the embodiment of the present invention, the classification algorithm uses a deep neural network to perform refined classification on image advertising materials and text advertising materials to obtain classification results. The specific classification method is described in detail below.

[0065] The classification results here are refined into specific styles, themes and types, making the advertisement display more accurate and significantly improving the user click-through rate. Compared with the traditional simple label classification method, the present invention has increased the click-through rate by about 20%-30% in actual tests.

[0066] S3: Obtain the user's historical behavior data on advertising materials, and generate a user preference model by analyzing the user's preferences for different types of advertising materials through a deep learning algorithm.

[0067] Specifically, in an embodiment of the present invention, historical behavior data such as the user's browsing history, click records, and interactive behavior on advertising materials are obtained to construct a detailed user portrait, which describes the user's interests and preferences. A collaborative filtering algorithm and an autoencoder are used to analyze the user's interests and preferences for advertising materials, and a dynamically updated user preference model is generated. The specific method for generating a user preference model is described in detail below.

[0068] Here, the user preference model can accurately predict user preferences for different types of advertising materials, thereby improving the personalization and precision of advertising material delivery.

[0069] S4: Obtain real-time user behavior data on advertising materials, and adjust the display frequency and sequence of advertising materials based on the real-time behavior data and the user preference model.

[0070] Specifically, in the embodiments of the present invention, the real-time behavior data of the advertising material refers to the browsing data, click data and interactive behavior data of the advertising material. It should be understood that the click effect of the advertising material can be dynamically monitored and analyzed through the real-time behavior data, and reinforcement learning algorithms such as Q-learning can be used to adjust the display frequency and sequence of the advertising material according to the real-time behavior data and user preference model, so as to achieve dynamic optimization of the advertising material delivery.

[0071] S5: Adjust the advertising creative display frequency, display position, and creative content based on the user preference model and classification results.

[0072] Specifically, in the embodiments of the present invention, on the one hand, the advertising material is personalized according to the classification results and the user preference model to ensure that the advertising material seen by different users meets their interests and needs; on the other hand, during the advertising material delivery process, the click effect of the advertising material is monitored in real time through real-time behavioral data, and the delivery strategy of the advertising material is dynamically adjusted, such as the advertising material display frequency, display position, and material content, to enhance the user's advertising experience and maximize the advertising click effect.

[0073] Figure 2 A diagram of a method for extracting features to obtain image feature vectors provided in an embodiment of the present invention.

[0074] Furthermore, in the above method for optimizing click effects of an advertising platform based on artificial intelligence, a feature extraction algorithm extracts features of image advertising materials to obtain image feature vectors, and combines Figure 2 , including two steps from S21 to S22:

[0075] S21: Input the image creative into the convolutional neural network to extract high-level features of the image creative through convolution layers, activation functions, and pooling layers.

[0076] Specifically, in an embodiment of the present invention, a convolutional neural network is used to extract high-level features such as color, composition, and texture of image advertisements.

[0077] S22: Outputting a vector of high-level features of the image creative through a fully connected layer to obtain an image feature vector.

[0078] Specifically, in the embodiment of the present invention, the formula for extracting features of the image advertising material using the convolutional neural network to obtain the image feature vector is as follows:

[0079] F i,j =Σ m,n I i+m,j+n ·K m,n

[0080] Among them, F i,j Represents the value on the feature map, I i+m,j+n represents the value on the input image, K m,n Represents the value of the convolution kernel.

[0081] Figure 3 A diagram of a method for extracting features to obtain text feature vectors provided in an embodiment of the present invention.

[0082] Furthermore, in the above method for optimizing click effects of an advertising platform based on artificial intelligence, a feature extraction algorithm extracts features of text advertising materials to obtain text feature vectors, and combines Figure 3 , including three steps from S31 to S33:

[0083] S31: Input the text advertising material into the recurrent neural network or the long short-term memory network, and convert the text advertising material into a vector representation through an embedding layer;

[0084] S32: extracting text advertising material features through a multi-layer recurrent neural network unit or a long short-term memory network unit;

[0085] S33: Outputting the vector of text advertising material features through the fully connected layer to obtain a text feature vector.

[0086] Specifically, in the embodiment of the present invention, a recurrent neural network or a long short-term memory network is used to extract text features such as keywords, sentence structure, and emotional tendency of text advertisements;

[0087] The formula for extracting text ad creative features through a recurrent neural network or a long short-term memory network is as follows:

[0088] h t =σ(W h ·h t-1 +W x ·x t +b)

[0089] Among them, h t represents the hidden state at the current moment, x t represents the input at the current moment, W h and W x represents the weight matrix, b represents the bias, and σ() represents the activation function.

[0090] Figure 4 A diagram of a method for classifying advertising materials to obtain classification results provided by an embodiment of the present invention.

[0091] Furthermore, in the above-mentioned method for optimizing click effects of an advertising platform based on artificial intelligence, the classification algorithm classifies the image advertising material and the text advertising material according to the image feature vector and the text feature vector to obtain the classification result, and combines Figure 4 , including two steps from S41 to S42:

[0092] S41: input the image feature vector and the text feature vector into the deep neural network;

[0093] S42: Output the classification results through multiple layers of fully connected layers and non-linear activation functions.

[0094] Specifically, in the embodiment of the present invention, the formula for obtaining the classification result by the deep neural network for classifying the image advertising material and the text advertising material is as follows:

[0095] y=softmax(W·x+b)

[0096] Among them, y represents the classification result, the softmax function indicates that the output is converted into a probability distribution, W represents the weight matrix, x represents the input feature vector, and b represents the bias.

[0097] The deep neural network classifies the extracted advertising material features. Through multiple layers of fully connected layers and nonlinear activation functions, the deep neural network can process high-dimensional feature vectors. The classification results are refined to specific styles, themes and types, which greatly improves the accuracy and practicality of classification. In addition, the deep neural network has powerful adaptive learning capabilities and can be trained on large-scale data sets to improve the accuracy and robustness of classification.

[0098] Figure 5 A diagram of a method for generating a user preference model provided by an embodiment of the present invention.

[0099] Furthermore, in the above-mentioned method for optimizing click effects of an advertising platform based on artificial intelligence, historical behavior data of users on advertising materials is obtained, and a user preference model is generated by analyzing the user's preference for different types of advertising materials through a deep learning algorithm. Figure 5 , including two steps from S51 to S52:

[0100] S51: Obtain historical behavior data of the user towards the advertising material, and construct a user profile based on the historical behavior data of the user towards the advertising material.

[0101] Specifically, in the embodiment of the present invention, historical behavior data such as the user's browsing history, click record, and interactive behavior on the advertising material are obtained to construct a detailed user portrait, which describes the user's interests and preferences.

[0102] S52: Analyze user preferences for different types of advertising materials through collaborative filtering algorithms and neural networks, and generate a user preference model.

[0103] Specifically, in an embodiment of the present invention, the neural network is an autoencoder, and a collaborative filtering algorithm and an autoencoder are used to analyze user preferences for advertising materials to generate a dynamically updated user preference model; before using the autoencoder to analyze user preferences for advertising materials, the user behavior data is matrix decomposed, and the autoencoder is used to analyze the user preferences for advertising materials as follows: the user behavior data after the matrix decomposition is encoded and decoded using the autoencoder.

[0104] The formula for encoding and decoding the user behavior data after matrix decomposition using the autoencoder is as follows:

[0105] in, represents the reconstructed rating matrix, E and D represent the encoding and decoding processes respectively, and R represents the user behavior data after matrix decomposition.

[0106] Furthermore, in the above method for optimizing click effects of an advertising platform based on artificial intelligence, the display frequency and order of advertising materials are adjusted according to real-time behavior data and user preference models, including:

[0107] The Q-learning algorithm continuously updates the state-action value based on real-time behavior data and user preference models;

[0108] Adjust the frequency and order of creative display based on the updated status-action value.

[0109] Specifically, in the embodiment of the present invention, the Q-learning formula is as follows:

[0110] Q(s,a)=Q(s,a)+α[r+γmax a' Q(s',a')-Q(s,a)]

[0111] Among them, Q(s,a) is the state-action value, which indicates the value of taking action a in state s, α indicates the learning rate, r indicates the immediate reward, γ indicates the discount factor, s' indicates the next state, and α' indicates the next action.

[0112] Here, the Q-learning algorithm can quickly adjust the advertising material delivery strategy, such as the display frequency and order of advertising materials, based on real-time feedback, to ensure real-time optimization of advertising display effects and adapt to market changes and user behavior changes.

[0113] Furthermore, the above method for optimizing click effects of an advertising platform based on artificial intelligence also includes:

[0114] Periodically compare the click-through rate and conversion rate of the creatives before and after the adjustments;

[0115] Determine whether to improve the feature extraction algorithm, classification algorithm, or adjust the display frequency, sequence, placement, and content of the creatives based on the click-through rate and conversion rate.

[0116] Specifically, in the embodiments of the present invention, the advertising optimization effect is periodically evaluated, and key indicators such as click-through rate and conversion rate before and after optimization are compared through data analysis. The feature extraction algorithm and classification algorithm are continuously improved, or the display frequency, sequence, display position and material content of the advertising materials are adjusted to ensure continuous improvement of the advertising effect.

[0117] Device Embodiment

[0118] Figure 6 A diagram of a device for optimizing click effects of an advertising platform based on artificial intelligence provided in an embodiment of the present invention.

[0119] In a second aspect of the present invention, a device for optimizing click effects of an advertising platform based on artificial intelligence is also proposed. Figure 6 ,include:

[0120] Feature extraction module 61: used for feature extraction algorithm to extract features from image advertising material and text advertising material respectively to obtain image feature vector and text feature vector.

[0121] Specifically, in the embodiment of the present invention, the feature extraction algorithm in the feature extraction module 61 uses a convolutional neural network to extract features of image advertising materials to obtain image feature vectors, and uses a recurrent neural network or a long short-term memory network to extract features of text advertising materials to obtain text feature vectors.

[0122] Classification module 62: used for a classification algorithm to classify image advertising materials and text advertising materials according to image feature vectors and text feature vectors to obtain classification results.

[0123] Specifically, in the embodiment of the present invention, the classification algorithm uses a deep neural network to perform fine classification on image advertising materials and text advertising materials to obtain classification results. The classification results here are refined to specific styles, themes and types, making the advertising display more accurate and significantly improving the user click-through rate. Compared with the traditional simple label classification method, the present invention has increased the click-through rate by about 20%-30% in actual tests.

[0124] Generation module 63: used to obtain historical behavior data of users on advertising materials, and to generate a user preference model by analyzing the user's preference for different types of advertising materials through a deep learning algorithm.

[0125] Specifically, in the embodiment of the present invention, the generation module 63 obtains historical behavior data such as the user's browsing history, click records, and interactive behaviors on the advertising materials, constructs a detailed user portrait, and the user portrait describes the user's interests and preferences. The collaborative filtering algorithm and autoencoder are used to analyze the user's interests and preferences for the advertising materials, and a dynamically updated user preference model is generated.

[0126] Here, the user preference model can accurately predict user preferences for different types of advertising materials, thereby improving the personalization and precision of advertising material delivery.

[0127] The first adjustment module 64 is used to obtain real-time behavior data of users on advertising materials, and adjust the display frequency and sequence of advertising materials according to the real-time behavior data and the user preference model.

[0128] Specifically, in the embodiment of the present invention, the real-time behavior data of the advertising material is the browsing data, click data and interactive behavior data of the advertising material. It should be understood that the click effect of the advertising material can be dynamically monitored and analyzed through the real-time behavior data. The first adjustment module 64 uses a reinforcement learning algorithm such as Q-learning to adjust the display frequency and sequence of the advertising material according to the real-time behavior data and the user preference model, so as to realize dynamic optimization of the advertising material delivery.

[0129] The second adjustment module 65 is used to adjust the display frequency, display position and material content of the advertisement material according to the user preference model and the classification result.

[0130] Specifically, in the embodiment of the present invention, on the one hand, the second adjustment module 65 performs personalized adjustment on the advertising material according to the classification results and the user preference model to ensure that the advertising materials seen by different users meet their interests and needs. On the other hand, during the advertising material delivery process, the click effect of the advertising material is monitored in real time through real-time behavioral data, and the delivery strategy of the advertising material, such as the display frequency, display location and material content of the advertising material, is dynamically adjusted to enhance the user's advertising experience and maximize the advertising click effect.

[0131] According to a third aspect of the present invention, an electronic device is provided, comprising: a processor and a memory;

[0132] The processor calls the program or instruction stored in the memory to execute any of the above methods for optimizing click effects of an advertising platform based on artificial intelligence.

[0133] In a fourth aspect of the present invention, a computer-readable storage medium is proposed, which stores a program or instruction, and the program or instruction enables a computer to execute any one of the above methods for optimizing click effects of an artificial intelligence-based advertising platform.

[0134] Figure 7 It is a schematic block diagram of an electronic device provided by an embodiment of the present invention.

[0135] like Figure 7 As shown, the electronic device includes: at least one processor 701, at least one memory 702 and at least one communication interface 703. The various components in the electronic device are coupled together through a bus system 704. The communication interface 703 is used to transmit information between external devices. It can be understood that the bus system 704 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 704 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, Figure 7 Various buses are labeled as bus system 704 .

[0136] It can be understood that the memory 702 in this embodiment can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories.

[0137] In some implementations, the memory 702 stores the following elements, executable units or data structures, or a subset thereof, or an extended set thereof: an operating system and application programs.

[0138] Among them, the operating system includes various system programs, such as the framework layer, the core library layer, the driver layer, etc., which are used to implement various basic services and process hardware-based tasks. The application program includes various application programs, such as a media player (Media Player), a browser (Browser), etc., which are used to implement various application services. A program that implements any method of the method for optimizing click effects of an advertising platform based on artificial intelligence provided in an embodiment of the present invention can be included in the application program.

[0139] In an embodiment of the present invention, the processor 701 calls a program or instruction stored in the memory 702, specifically, a program or instruction stored in an application, and the processor 701 is used to execute the steps of each embodiment of a method for optimizing click effects of an advertising platform based on artificial intelligence provided in an embodiment of the present invention.

[0140] The feature extraction algorithm extracts features from the image advertising material and the text advertising material to obtain an image feature vector and a text feature vector;

[0141] The classification algorithm classifies the image advertising material and the text advertising material according to the image feature vector and the text feature vector to obtain a classification result;

[0142] Obtain historical user behavior data on advertising materials, and generate a user preference model by analyzing user preferences for different types of advertising materials through deep learning algorithms;

[0143] Obtain real-time user behavior data on advertising materials, and adjust the display frequency and order of advertising materials based on the real-time behavior data and user preference model;

[0144] Adjust the frequency, placement, and content of ad creatives based on user preference models and classification results.

[0145] Any of the methods for optimizing click effects of an advertising platform based on artificial intelligence provided in an embodiment of the present invention can be applied to the processor 701, or implemented by the processor 701. The processor 701 can be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor 701 or instructions in the form of software. The above-mentioned processor 701 can be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0146] The steps of any method in the method for optimizing click effects of an advertising platform based on artificial intelligence provided by an embodiment of the present invention can be directly embodied as being executed by a hardware decoding processor, or being executed by a combination of hardware and software units in the decoding processor. The software unit can be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 702, and the processor 701 reads the information in the memory 702 and completes the steps of the method in combination with its hardware.

[0147] Those skilled in the art will appreciate that although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features from different embodiments is meant to be within the scope of the present invention and to form different embodiments.

[0148] Those skilled in the art will appreciate that the description of each embodiment has its own emphasis, and for parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0149] Although the embodiments of the present invention are described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations all fall within the scope defined by the appended claims. The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of various equivalent modifications or substitutions within the technical scope disclosed by the present invention, and these modifications or substitutions should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

[0150] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for optimizing click effects of an advertising platform based on artificial intelligence, characterized in that: include: The feature extraction algorithm extracts features from the image advertising material and the text advertising material to obtain an image feature vector and a text feature vector; The classification algorithm classifies the image advertising material and the text advertising material according to the image feature vector and the text feature vector to obtain a classification result; Obtain historical user behavior data on advertising materials, and generate a user preference model by analyzing user preferences for different types of advertising materials through deep learning algorithms; Acquire real-time behavior data of users on advertising materials, and adjust the display frequency and order of advertising materials according to the real-time behavior data and the user preference model; The advertisement material display frequency, display position and material content are adjusted according to the user preference model and the classification result.

2. The method for optimizing click effects of an advertising platform based on artificial intelligence according to claim 1, characterized in that: The feature extraction algorithm extracts features from the image ad material to obtain an image feature vector, including: Input the image advertising material into a convolutional neural network to extract high-level features of the image advertising material through a convolutional layer, an activation function, and a pooling layer; The image feature vector is obtained by outputting a vector of high-level features of the image creative through a fully connected layer.

3. The method for optimizing click effects of an advertising platform based on artificial intelligence according to claim 1, characterized in that: The feature extraction algorithm extracts features from the text advertising material to obtain a text feature vector, including: Input the text advertisement material into a recurrent neural network or a long short-term memory network to convert the text advertisement material into a vector representation through an embedding layer; Extracting text advertising creative features through multi-layer recurrent neural network units or long short-term memory network units; The text feature vector is obtained by outputting the vector of text advertising material features through the fully connected layer.

4. The method for optimizing click effects of an advertising platform based on artificial intelligence according to claim 1, characterized in that: The classification algorithm classifies the image advertising material and the text advertising material according to the image feature vector and the text feature vector to obtain a classification result, including: Inputting the image feature vector and the text feature vector into a deep neural network; The classification results are output through multiple layers of fully connected layers and non-linear activation functions.

5. The method for optimizing click effects of an advertising platform based on artificial intelligence according to claim 1, characterized in that: The step of obtaining the historical behavior data of the user on the advertising material and analyzing the user's preference on different types of advertising materials by a deep learning algorithm to generate a user preference model includes: Acquire historical behavior data of the user towards the advertising material, and construct a user profile based on the historical behavior data of the user towards the advertising material; Analyze users’ preferences for different types of advertising materials through collaborative filtering algorithms and neural networks to generate user preference models; The user's historical behavior data on the advertising material includes at least browsing history, click records and interactive behavior.

6. The method for optimizing click effects of an advertising platform based on artificial intelligence according to claim 1, characterized in that: The step of adjusting the display frequency and sequence of the advertisement materials according to the real-time behavior data and the user preference model includes: The Q-learning algorithm continuously updates the state-action value based on real-time behavior data and the user preference model; Adjust the frequency and order of creative display based on the updated status-action value.

7. The method for optimizing click effects of an advertising platform based on artificial intelligence according to claim 1, characterized in that: The method further comprises: Periodically compare the click-through rate and conversion rate of the creatives before and after the adjustments; Determine whether to improve the feature extraction algorithm, the classification algorithm, or adjust the display frequency, sequence, display position, and material content of the advertising material based on the click rate and the conversion rate.

8. A device for optimizing click effects of an advertising platform based on artificial intelligence, characterized in that: include: Feature extraction module: used for feature extraction algorithm to extract features of image advertising materials and text advertising materials respectively to obtain image feature vectors and text feature vectors; Classification module: used for a classification algorithm to classify the image advertising material and the text advertising material according to the image feature vector and the text feature vector to obtain a classification result; Generation module: used to obtain historical behavior data of users on advertising materials, and generate a user preference model by analyzing user preferences for different types of advertising materials through deep learning algorithms; A first adjustment module: used to obtain real-time behavior data of users on advertising materials, and adjust the display frequency and order of advertising materials according to the real-time behavior data and the user preference model; The second adjustment module is used to adjust the display frequency, display position and material content of the advertisement material according to the user preference model and the classification result.

9. An electronic device, characterized in that: include: Processor and memory; The processor is used to execute the method for optimizing click effects of an advertising platform based on artificial intelligence as described in any one of claims 1 to 7 by calling the program or instruction stored in the memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores programs or instructions, and the programs or instructions enable a computer to execute a method for optimizing click effects of an advertising platform based on artificial intelligence as described in any one of claims 1 to 7.

Images