Intelligent copywriting creation method and system based on AIGC

Abstract

The present invention discloses an intelligent copywriting method and system based on AIGC, which relates to the technical field of computing models, including a data acquisition unit, a model building unit, a copywriting generation unit, a copywriting evaluation unit and a copywriting optimization unit. The present invention obtains an AIGC model through recurrent neural network training, automatically generates a primary push copywriting according to a push purpose, a push display mode and preferred keywords, extracts comparison keywords based on the primary push copywriting, retrieves the copywriting content with the highest reading rate from the user's historical click records and historical reading records as the comparison copywriting, performs feature extraction on the comparison copywriting and the primary push copywriting to perform copywriting evaluation, corrects the AIGC model according to the copywriting evaluation result to obtain an optimized AIGC model, inputs the push purpose, the push display mode and the preferred keywords into the optimized AIGC model again to obtain an optimized push copywriting, provides a personalized copywriting generation process, and improves the accuracy of copywriting generation.

Description

Technical Field

[0001] The present invention relates to the field of computing model technology, and in particular to an AIGC-based intelligent copywriting method and system. Background Art

[0002] With the popularity of mobile smart terminal devices and the rapid development of mobile applications, users often browse news, videos and other information through smart terminals. To improve the user experience, various applications often generate and send push copy to push hot information to users for easy browsing.

[0003] Excellent push copy can improve user engagement with mobile apps and increase click-through rates. However, customizing push copy for different users and improving click-through rates remain significant challenges. Therefore, solutions with improved user experience and higher recall rates are needed.

[0004] In view of the above technical defects, a solution is now proposed. Summary of the Invention

[0005] The purpose of the present invention is to evaluate the copy by extracting features from the comparison copy and the primary push copy, and to modify the AIGC model according to the copy evaluation results to obtain an optimized AIGC model, thereby obtaining an optimized push copy, providing a personalized copy generation process, and improving the accuracy of copy generation.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an AIGC-based intelligent copywriting method, comprising the following steps:

[0007] Step 1: Obtain a push task, which includes the push target, push purpose, and push display method, and simultaneously obtain the push target's personal information, which includes historical click records, historical reading records, and preferred keywords;

[0008] Step 2: Obtain multiple sets of historical copywriting data, train the AIGC model through a recurrent neural network, and automatically generate primary push copywriting based on the push purpose, push display method, and preferred keywords;

[0009] Step 3: Extract comparison keywords based on the primary push copy, and search for comparison keywords from the user's historical click records and historical reading records to obtain a comparison database. Calculate the reading rate of each copy in the comparison database, and obtain the copy with the highest reading rate as the comparison copy;

[0010] Step 4: Extract features from the comparison copy and the primary push copy to obtain feature vectors F1 and F2, respectively. Calculate the vector similarity to obtain the copy evaluation results.

[0011] Step 5: Modify the AIGC model based on the copy evaluation results to obtain an optimized AIGC model. Input the push purpose, push display method, and preferred keywords into the optimized AIGC model again to obtain the optimized push copy.

[0012] Furthermore, the specific process of obtaining the AIGC model through recurrent neural network training is as follows:

[0013] S101: Acquire multiple sets of historical copywriting generation data, including historical push purposes, historical push display modes, historical push copywriting, and preferred keywords corresponding to the copywriting, split the historical copywriting generation data into words, assign a unique identifier to each word, and integrate them to obtain training samples;

[0014] S102, the training samples are divided into a training set and a test set in a ratio of 8:2, and all data are simultaneously completed, anomaly cleaned, and normalized;

[0015] S103. Construct a neural network model with an LSTM layer, set an embedding layer, and use a Dropout layer to randomly disconnect a certain proportion of network connections to avoid overfitting.

[0016] S104, connect the LSTM layer and specify a certain ratio of Dropout;

[0017] S105: Connect a fully connected layer and a softmax layer, and output the primary push copy;

[0018] S106, set the optimizer to adam stochastic gradient descent, set the loss function to multi-class cross entropy function, set the measurement indicator to accuracy, and iterate multiple times until the model converges;

[0019] S107. When the number of training times reaches the threshold and the loss function is greater than the historical minimum, the optimal model is selected as the AIGC model, and the training set data is used for training. The mean square error between the model result and the actual result is calculated using the test set as the model evaluation.

[0020] Furthermore, the specific process of obtaining the copywriting content with the highest reading rate as the comparison copywriting is as follows:

[0021] S201: Obtain the generated primary push copy, pre-process the text data of the primary push copy, perform word segmentation based on the domain dictionary, and calculate the comprehensive score of each word;

[0022] S202, sorting the comprehensive scores of each word from large to small, and selecting words with scores greater than a preset threshold as comparison keywords;

[0023] S203: Retrieve comparison keywords from the user's historical click records and historical reading records, use historical texts containing the keywords as a candidate database, perform a secondary search within the candidate database, and use historical texts with keyword frequencies greater than a preset frequency threshold as a comparison database;

[0024] S204: Obtain the historical click count Mi and historical reading time Ti corresponding to each copy content in the comparison database, and calculate the copy content reading rate Ui according to the following formula: , where α and β are preset weight coefficients, and the reading rate Ui is used to reflect the reading interest of the corresponding copy content of the push target. The higher the reading rate Ui;

[0025] S205. Sort the reading rates of each copy content from highest to lowest, and use the copy content ranked first as the comparison copy.

[0026] Furthermore, the specific process of calculating the comprehensive score of each word is as follows:

[0027] Obtain the word length e, word position d, and word frequency-inverse document frequency (IDF) of each word after word segmentation, and calculate the comprehensive score K of each word according to the following formula: , where the word length e is calculated based on a preset proportional coefficient, and different weights are assigned to words at different positions in the text content. The word position d is calculated based on the position of the word in the text content to obtain a position score. The comprehensive score K reflects the key proportion of each word in the text data. The larger the comprehensive score K, the higher the key importance of the word in the text data. Conversely, the smaller the comprehensive score K, the lower the key importance of the word in the text data.

[0028] Furthermore, the specific process of obtaining the copywriting evaluation results is as follows:

[0029] S301, using TfidfVectorizer in sklearn, extract features of the comparison copy and the primary push copy in turn to obtain feature vectors F1 and F2. , ;

[0030] S302. Calculate the vector similarity S(F1, F2) according to the following formula: ;

[0031] S303: Obtain a preset similarity threshold. If the similarity is greater than or equal to the similarity threshold, the copy evaluation result is obtained as follows: the similarity between the primary push copy and the comparison copy meets the push requirements;

[0032] If the similarity is less than the similarity threshold, the copy evaluation result is that the similarity between the primary push copy and the comparison copy does not meet the push requirements, and the primary push copy and the comparison copy are used as training samples.

[0033] Furthermore, the specific process of obtaining the optimized AIGC model is as follows:

[0034] S401: Obtain the primary push copy and the comparison copy as well as the corresponding feature vectors F1 and F2, and perform feature numerical processing on each element in the feature vectors F1 and F2;

[0035] S402, using LabelEncoder to perform numerical processing on the character data, mapping the label to the interval [0, n_classes-1] through LabelEncoder for encoding, and assigning a unique identifier to each label;

[0036] S403, after the features are digitized, the data is standardized and linearly mapped to the interval [0, 1]. Specifically, the StandardScale algorithm is used to standardize the data and integrate to obtain optimized training samples;

[0037] S404: Adjust the parameters of the softmax layer and the fully connected layer, and output the optimized push copy;

[0038] S405. Use the optimized training samples to perform multiple iterations again until the model converges. When the number of training times reaches a threshold and the loss function is greater than the historical minimum, the optimal model is selected as the optimized AIGC model.

[0039] The present invention also provides an AIGC-based intelligent copywriting system, comprising a data acquisition unit, a model building unit, a copywriting generation unit, a copywriting evaluation unit, and a copywriting optimization unit;

[0040] The data acquisition unit is used to acquire a push task, which includes a push target, a push purpose, and a push display mode, and simultaneously acquires the personal information of the push target, which includes historical click records, historical reading records, and preferred keywords;

[0041] The model building unit is used to obtain multiple sets of historical copy generation data, train the AIGC model through a recurrent neural network, and automatically generate primary push copy based on the push purpose, push display method, and preferred keywords;

[0042] The copy generation unit is used to extract comparison keywords from the primary push copy, and retrieve the comparison keywords from the user's historical click records and historical reading records to obtain a comparison database, calculate the reading rate of each copy content in the comparison database, and obtain the copy content with the highest reading rate as the comparison copy;

[0043] The copy evaluation unit is used to extract features from the comparison copy and the primary push copy, respectively, to obtain feature vectors F1 and F2, and to obtain the copy evaluation results by calculating vector similarity;

[0044] The copy optimization unit is used to modify the AIGC model according to the copy evaluation results to obtain an optimized AIGC model, and then input the push purpose, push display method and preferred keywords into the optimized AIGC model to obtain an optimized push copy.

[0045] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0046] The AIGC-based intelligent copywriting method and system obtain an AIGC model through recurrent neural network training, automatically generate primary push copywriting according to push purpose, push display mode and preferred keywords, extract comparison keywords based on the primary push copywriting, retrieve the copywriting content with the highest reading rate from the user's historical click records and historical reading records as comparison copywriting, perform feature extraction on the comparison copywriting and the primary push copywriting for copywriting evaluation, modify the AIGC model according to the copywriting evaluation results to obtain an optimized AIGC model, and again input the push purpose, push display mode and preferred keywords into the optimized AIGC model to obtain optimized push copywriting, thereby providing a personalized copywriting generation process and improving the accuracy of copywriting generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 Shown is a schematic diagram of the overall method flow of the present invention;

[0048] Figure 2 Shown is a schematic diagram of the overall structure of the present invention. DETAILED DESCRIPTION

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0050] Example 1:

[0051] like Figure 1As shown, the AIGC-based intelligent copywriting creation method includes the following steps:

[0052] Step 1: Get the push task, which includes the push target, push purpose, and push display method. At the same time, get the push target's personal information, which includes historical click records, historical reading records, and preferred keywords.

[0053] Step 2: Obtain multiple sets of historical copywriting data, train the AIGC model through a recurrent neural network, and automatically generate primary push copywriting based on the push purpose, push display method, and preferred keywords;

[0054] The specific process of obtaining the AIGC model through recurrent neural network training is as follows:

[0055] S101. Acquire multiple sets of historical copywriting generation data, including historical push purposes, historical push display methods, historical push copywriting, and preferred keywords corresponding to the copywriting, split the historical copywriting generation data into words, assign a unique identifier to each word, and integrate them to obtain training samples;

[0056] S102, the training samples are divided into a training set and a test set in a ratio of 8:2, and all data are simultaneously completed, anomaly cleaned, and normalized;

[0057] S103. Construct a neural network model with an LSTM layer, set an embedding layer, and use a Dropout layer to randomly disconnect a certain proportion of network connections to avoid overfitting.

[0058] S104, connect the LSTM layer and specify a certain ratio of Dropout;

[0059] S105: Connect a fully connected layer and a softmax layer, and output the primary push copy;

[0060] S106, set the optimizer to adam stochastic gradient descent, set the loss function to multi-class cross entropy function, set the measurement indicator to accuracy, and iterate multiple times until the model converges;

[0061] S107. When the number of training times reaches the threshold and the loss function is greater than the historical minimum, the optimal model is selected as the AIGC model, and the training set data is used for training. The mean square error between the model result and the actual result is calculated using the test set as the model evaluation.

[0062] Step 3: Extract comparison keywords based on the primary push copy, and search for comparison keywords from the user's historical click records and historical reading records to obtain a comparison database. Calculate the reading rate of each copy in the comparison database, and obtain the copy with the highest reading rate as the comparison copy;

[0063] The specific process of obtaining the copy with the highest reading rate as the comparison copy is as follows:

[0064] S201: Obtain the generated primary push copy, pre-process the text data of the primary push copy, perform word segmentation based on the domain dictionary, and calculate the comprehensive score of each word;

[0065] The specific process of calculating the comprehensive score of each word is as follows:

[0066] Obtain the word length e, word position d, and word frequency-inverse document frequency (IDF) of each word after word segmentation, and calculate the comprehensive score K of each word according to the following formula: , where the word length e is calculated based on a preset proportional coefficient, different weights are assigned to words at different positions in the text content, the word position d is calculated based on the position of the word in the text content to obtain a position score, and the comprehensive score K reflects the key proportion of each word in the text data. The larger the comprehensive score K, the higher the key importance of the word in the text data, and vice versa, the smaller the comprehensive score K, the lower the key importance of the word in the text data;

[0067] S202, sorting the comprehensive scores of each word from large to small, and selecting words with scores greater than a preset threshold as comparison keywords;

[0068] S203: Retrieve comparison keywords from the user's historical click records and historical reading records, use historical texts containing the keywords as a candidate database, perform a secondary search within the candidate database, and use historical texts with keyword frequencies greater than a preset frequency threshold as a comparison database;

[0069] S204: Obtain the historical click count Mi and historical reading time Ti corresponding to each copy content in the comparison database, and calculate the copy content reading rate Ui according to the following formula: , where α and β are preset weight coefficients, and the reading rate Ui is used to reflect the reading interest of the corresponding copy content of the push target. The higher the reading rate Ui;

[0070] S205. Sort the reading rates of each copy content from highest to lowest, and use the copy content ranked first as the comparison copy.

[0071] Step 4: Extract features from the comparison copy and the primary push copy to obtain feature vectors F1 and F2, respectively. Calculate the vector similarity to obtain the copy evaluation results.

[0072] The specific process of obtaining the copywriting evaluation results is as follows:

[0073] S301, using TfidfVectorizer in sklearn, extract features of the comparison copy and the primary push copy in turn to obtain feature vectors F1 and F2, , ;

[0074] S302. Calculate the vector similarity S(F1, F2) according to the following formula: ;

[0075] S303: Obtain a preset similarity threshold. If the similarity is greater than or equal to the similarity threshold, the copy evaluation result is obtained as follows: the similarity between the primary push copy and the comparison copy meets the push requirements;

[0076] If the similarity is less than the similarity threshold, the copy evaluation result is that the similarity between the primary push copy and the comparison copy does not meet the push requirements, and the primary push copy and the comparison copy are used as training samples.

[0077] Step 5: Modify the AIGC model based on the copy evaluation results to obtain an optimized AIGC model. Input the push purpose, push display method, and preferred keywords into the optimized AIGC model again to obtain the optimized push copy.

[0078] The specific process of obtaining the optimized AIGC model is as follows:

[0079] S401: Obtain the primary push copy and the comparison copy as well as the corresponding feature vectors F1 and F2, and perform feature numerical processing on each element in the feature vectors F1 and F2;

[0080] S402, using LabelEncoder to perform numerical processing on the character data, mapping the label to the interval [0, n_classes-1] through LabelEncoder for encoding, and assigning a unique identifier to each label;

[0081] S403, after the features are digitized, the data is standardized and linearly mapped to the interval [0, 1]. Specifically, the StandardScale algorithm is used to standardize the data and integrate to obtain optimized training samples;

[0082] S404: Adjust the parameters of the softmax layer and the fully connected layer, and output the optimized push copy;

[0083] S405. Use the optimized training samples to perform multiple iterations again until the model converges. When the number of training times reaches a threshold and the loss function is greater than the historical minimum, the optimal model is selected as the optimized AIGC model.

[0084] The present invention obtains an AIGC model through recurrent neural network training, automatically generates primary push copy according to push purpose, push display mode and preferred keywords, extracts comparison keywords based on the primary push copy, retrieves the copy content with the highest reading rate from the user's historical click records and historical reading records as the comparison copy, performs feature extraction on the comparison copy and the primary push copy to perform copy evaluation, corrects the AIGC model according to the copy evaluation result to obtain an optimized AIGC model, and again inputs the push purpose, push display mode and preferred keywords into the optimized AIGC model to obtain optimized push copy, thereby providing a personalized copy generation process and improving the accuracy of copy generation.

[0085] Example 2:

[0086] like Figure 2 As shown, the AIGC-based intelligent copywriting system includes a data acquisition unit, a model building unit, a copywriting generation unit, a copywriting evaluation unit, and a copywriting optimization unit;

[0087] The data acquisition unit is used to obtain push tasks, which include push targets, push purposes, and push display methods, and also obtain the personal information of the push targets, which includes historical click records, historical reading records, and preferred keywords;

[0088] The model building unit is used to obtain multiple sets of historical copy generation data, train the AIGC model through a recurrent neural network, and automatically generate primary push copy based on the push purpose, push display method, and preferred keywords;

[0089] The copy generation unit is used to extract comparison keywords from the primary push copy, and retrieve the comparison keywords from the user's historical click records and historical reading records to obtain a comparison database, calculate the reading rate of each copy content in the comparison database, and obtain the copy content with the highest reading rate as the comparison copy;

[0090] The copy evaluation unit is used to extract features from the comparison copy and the primary push copy, respectively, to obtain feature vectors F1 and F2, and to obtain the copy evaluation results by calculating vector similarity;

[0091] The copy optimization unit is used to modify the AIGC model according to the copy evaluation results to obtain an optimized AIGC model, and then input the push purpose, push display method and preferred keywords into the optimized AIGC model to obtain an optimized push copy.

[0092] The threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by technicians in this field for each set of sample data; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0093] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by those skilled in the art according to actual conditions.

[0094] In the two embodiments provided in this application, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the division of modules is merely a division of logical functions, and in actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not implemented. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces.

[0095] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. The AIGC-based intelligent copywriting method is characterized by: The following steps are involved: Step 1: Obtain a push task, which includes the push target, push purpose, and push display method, and simultaneously obtain the push target's personal information, which includes historical click records, historical reading records, and preferred keywords; Step 2: Obtain multiple sets of historical copywriting data, train the AIGC model through a recurrent neural network, and automatically generate primary push copywriting based on the push purpose, push display method, and preferred keywords; Step 3: Extract comparison keywords based on the primary push copy, and search for comparison keywords from the user's historical click records and historical reading records to obtain a comparison database. Calculate the reading rate of each copy in the comparison database, and obtain the copy with the highest reading rate as the comparison copy; The specific process of obtaining the copy with the highest reading rate as the comparison copy is as follows: S201: Obtain the generated primary push copy, pre-process the text data of the primary push copy, perform word segmentation based on the domain dictionary, and calculate the comprehensive score of each word; S202, sorting the comprehensive scores of each word from large to small, and selecting words with scores greater than a preset threshold as comparison keywords; S203: Retrieve comparison keywords from the user's historical click records and historical reading records, use historical texts containing the comparison keywords as a candidate database, perform a secondary search within the candidate database, and use historical texts whose comparison keyword frequency is greater than a preset frequency threshold as a comparison database; S204: Obtain the historical click count Mi and historical reading time Ti corresponding to each copy content in the comparison database, and calculate the copy content reading rate Ui according to the following formula: , where α and β are preset weight coefficients, and the reading rate Ui is used to reflect the reading interest of the push target in the copy content; S205. Sort the reading rates of each copy content from highest to lowest, and use the copy content ranked first as the comparison copy; Step 4: Extract features from the comparison copy and the primary push copy to obtain feature vectors F1 and F2, respectively. Calculate the vector similarity to obtain the copy evaluation results. Step 5: Modify the AIGC model based on the copy evaluation results to obtain an optimized AIGC model. Input the push purpose, push display method, and preferred keywords into the optimized AIGC model again to obtain the optimized push copy.

2. The AIGC-based intelligent copywriting method according to claim 1, characterized in that: The specific process of obtaining the AIGC model through recurrent neural network training is as follows: S101: Acquire multiple sets of historical copywriting generation data, wherein the historical copywriting generation data includes historical push purposes, historical push display modes, historical push copywriting, and preferred keywords corresponding to the copywriting, split the historical copywriting generation data into words, assign a unique identifier to each word, and integrate them to obtain training samples; S102, the training samples are divided into a training set and a test set in a ratio of 8:2, and all data are simultaneously completed, anomaly cleaned, and normalized; S103. Construct a neural network model with an LSTM layer, set an embedding layer, and use a Dropout layer to randomly disconnect a certain proportion of network connections to avoid overfitting. S104, connect the LSTM layer and specify a certain ratio of Dropout; S105: Connect a fully connected layer and a softmax layer, and output the primary push copy; S106, set the optimizer to adam stochastic gradient descent, set the loss function to multi-class cross entropy function, set the measurement indicator to accuracy, and iterate multiple times until the model converges; S107. When the number of training times reaches the threshold and the loss function is greater than the historical minimum, the optimal model is selected as the AIGC model, and the training set data is used for training. The mean square error between the model result and the actual result is calculated using the test set as the model evaluation.

3. The AIGC-based intelligent copywriting method according to claim 1, characterized in that: The specific process of calculating the comprehensive score of each word is as follows: Obtain the word length e, word position d, and inverse document frequency (IDF) of each word after word segmentation, and calculate the comprehensive score K of each word according to the following formula: , where the word length e is calculated based on a preset proportional coefficient, and different weights are assigned to words in different positions in the text content.

4. The AIGC-based intelligent copywriting method according to claim 1, characterized in that: The specific process of obtaining the copywriting evaluation results is as follows: S301, using TfidfVectorizer in sklearn, extract features of the comparison copy and the primary push copy in turn to obtain feature vectors F1 and F2. , ; S302. Calculate the vector similarity S(F1, F2) according to the following formula: ; S303: Obtain a preset similarity threshold. If the vector similarity is greater than or equal to the similarity threshold, the copy evaluation result is obtained as follows: the similarity between the primary push copy and the comparison copy meets the push requirements; If the vector similarity is less than the similarity threshold, the copy evaluation result is that the similarity between the primary push copy and the comparison copy does not meet the push requirements, and the primary push copy and the comparison copy are used as training samples.

5. The AIGC-based intelligent copywriting method according to claim 1, characterized in that: The specific process of obtaining the optimized AIGC model is as follows: S401: Obtain the primary push copy and the comparison copy as well as the corresponding feature vectors F1 and F2, and perform feature numerical processing on each element in the feature vectors F1 and F2; S402, using LabelEncoder to perform numerical processing on the character data, mapping the label to the interval [0, n_classes-1] through LabelEncoder for encoding, and assigning a unique identifier to each label; S403, after the features are digitized, the data is standardized and linearly mapped to the interval [0, 1]. Specifically, the StandardScale algorithm is used to standardize the data and integrate to obtain optimized training samples; S404: Adjust the parameters of the softmax layer and the fully connected layer, and output the optimized push copy; S405. Use the optimized training samples to perform multiple iterations again until the model converges. When the number of training times reaches a threshold and the loss function is greater than the historical minimum, the optimal model is selected as the optimized AIGC model.

6. The AIGC-based intelligent copywriting system is characterized by: The intelligent copywriting system is used to execute the AIGC-based intelligent copywriting method according to claim 1, and the intelligent copywriting system includes a data acquisition unit, a model building unit, a copywriting generation unit, a copywriting evaluation unit, and a copywriting optimization unit; The data acquisition unit is used to acquire a push task, which includes a push target, a push purpose, and a push display mode, and simultaneously acquires the personal information of the push target, which includes historical click records, historical reading records, and preferred keywords; The model building unit is used to obtain multiple sets of historical copy generation data, train the AIGC model through a recurrent neural network, and automatically generate primary push copy based on the push purpose, push display method, and preferred keywords; The copy generation unit is used to extract comparison keywords from the primary push copy, and retrieve the comparison keywords from the user's historical click records and historical reading records to obtain a comparison database, calculate the reading rate of each copy content in the comparison database, and obtain the copy content with the highest reading rate as the comparison copy; The copy evaluation unit is used to extract features from the comparison copy and the primary push copy, respectively, to obtain feature vectors F1 and F2, and to obtain the copy evaluation results by calculating vector similarity; The copy optimization unit is used to modify the AIGC model according to the copy evaluation results to obtain an optimized AIGC model, and then input the push purpose, push display method and preferred keywords into the optimized AIGC model to obtain an optimized push copy.

Images