A method for generating Chinese model essays based on deep learning

Abstract

The present invention discloses a method for generating Chinese model essays based on deep learning, which is used to solve the problem that there is currently no Chinese model essay generation method that can quickly and accurately generate Chinese model essays expected by users. The method includes: classifying a number of model essays according to the model essay type to obtain model essay collections of different types; constructing a model essay model for each model essay in each model essay collection; for each model essay collection, fusing the model essay models corresponding to all the model essays therein into a class fusion model corresponding to the model essay collection; receiving Chinese writing information input by a user; calculating the matching value of the Chinese writing information and various fusion models according to a preset matching value algorithm; and outputting the model essay in the model essay collection corresponding to the class fusion model corresponding to the calculated maximum matching value. The method displays the model essay in the model essay collection corresponding to the class fusion model with the largest matching value to the user, ensuring that the displayed model essay is the model essay expected by the user.

Description

Technical Field

[0001] The present invention relates to the technical field of model essay generation, and in particular to a Chinese model essay generation method based on deep learning. Background Art

[0002] With the rapid development of the Internet, there are more and more Internet users, and the relationship between users is becoming closer and closer through information sharing. Among them, information sharing not only makes resource allocation more reasonable, but also saves social costs and creates more wealth. At present, there are many resources for information sharing, including model essays, papers, software programs, movies, music, etc. Among them, in terms of model essay resource sharing, users can obtain model essay resources on certain target websites through the Internet. However, in the process of model essay retrieval and query, after the user enters the model essay keywords, there are often many model essays retrieved with different formats. The user can only spend a lot of time to exclude irrelevant model essays. Therefore, there is an urgent need for a model essay generation method that can quickly and accurately generate the model essays required by users. Summary of the invention

[0003] The present invention provides a method for generating Chinese model essays based on deep learning, which is used to solve the problem that there is currently no Chinese model essay generation method that can quickly and accurately generate Chinese model essays expected by users. The present invention provides a method for generating Chinese model essays based on deep learning, which calculates the matching value between Chinese writing information input by the user and various fusion models, and then displays the model essays in the model essay collection corresponding to the fusion model with the largest matching value to the user, thereby ensuring that the displayed model essays are the model essays expected by the user.

[0004] The present invention provides a method for generating Chinese model essays based on deep learning, comprising:

[0005] Classify several sample articles crawled from the target website according to the sample article type to obtain sample article collections of different types;

[0006] Construct a model of each essay in each essay collection;

[0007] For each essay set, the essay models corresponding to all the essays in it are fused into a class fusion model corresponding to the essay set;

[0008] Receiving Chinese writing information input by a user; the Chinese writing information at least includes a writing topic and a writing type;

[0009] Calculate the matching value between the Chinese writing information and various fusion models according to a preset matching value algorithm;

[0010] The class fusion model corresponding to the calculated maximum matching value is output to correspond to the example in the example set.

[0011] In one embodiment, constructing a model of each model essay in each model essay collection includes:

[0012] Extract the key items and their corresponding item attributes of each essay in each essay collection;

[0013] According to all the entry attributes of each sample article, a sample article model is constructed for each sample article.

[0014] In one embodiment, for each model essay set, fusing the model essay models corresponding to all the model essays therein into a class fusion model corresponding to the model essay set includes:

[0015] For each essay collection, according to the essay model of each essay in the current essay collection, the item attributes of each essay and the preset essay fusion value algorithm, the fusion value of the essay models of all essays in the current essay collection is calculated;

[0016] Determine whether the fusion value of the model corresponding to the current model set is less than the preset fusion value;

[0017] If the fusion value of the essay model corresponding to the current essay set is less than the preset fusion value, generating a corresponding class fusion model based on all essays in the current essay set and the preset essay generation model;

[0018] If the fusion value of the model essay model corresponding to the current model essay collection is not less than the preset fusion value, a number of model essays are crawled from the target website, and the steps of classifying the model essays crawled from the target website according to the model essay type to obtain model essay collections of different types are re-executed.

[0019] In one embodiment, the preset example fusion value algorithm is:

[0020]

[0021] Among them, R is the fusion value of the essay model of all essays in the current essay set, and M is the total number of essays in the current essay set; Represents the model fusion factor of the j-th model in the current model set, and its value range is [0.5, 0.8]; Represents the model fusion factor of the j+1th model in the current model set, and its value range is [0.5, 0.8]; Represents the standard fusion factor of all model essays in the current model essay set, is a preset value; exp() represents an exponential function; H is the total number of entry attributes corresponding to the jth model in the current model set; η his the attribute value of the hth item attribute in the jth example model in the current example set; η represents the average attribute value of all item attributes in the jth example model in the current example set.

[0022] In one embodiment, the preset matching value algorithm is:

[0023]

[0024] Among them, P i represents the matching value between the Chinese writing information input by the user and the i-th class fusion model; N represents the total number of pre-obtained class fusion models; β i is the correlation value between the writing topic and the i-th class fusion model, β i The value range of is [0,1]; i is the association value between the writing type and the i-th class fusion model, δ i The value range of F is [0.8,1]; i is the model trust value of the i-th class fusion model, and its value range is [3,5].

[0025] In one embodiment, the target website is a website related to the article.

[0026] In one embodiment, the model essay key entries may include: model essay topic, model essay key words, model essay technical content, model essay language modification, and model essay description fluency; the corresponding entry attributes may include: subject attributes, keyword attributes, technical attributes, language modification attributes, and description fluency attributes.

[0027] The present invention provides a method for generating Chinese model essays based on deep learning, which crawls model essays from a target website and classifies them to obtain model essay collections of different types, and determines the model essay key items of each model essay and their corresponding item attributes, thereby constructing a model essay model for each model essay, and then calculating the fusion value of the model essay models of all the model essays in the model essay collection according to a preset model essay fusion value algorithm, and when the fusion value of the model essay model is less than the preset fusion value, a corresponding class fusion model is generated according to all the model essays in the model essay collection and the preset model essay generation model, thereby ensuring the reliability and accuracy of the constructed model essay model. Finally, after obtaining Chinese writing information such as the Chinese subject and writing type input by the user, the matching value of the Chinese writing information and various fusion models is calculated, and finally the model essay in the model essay collection corresponding to the class fusion model with the highest matching value is displayed to the user, thereby ensuring that the model essay displayed to the user is the model essay expected by the user.

[0028] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0029] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0031] Figure 1 This is a flow chart of a first embodiment of a method for generating a Chinese model essay based on deep learning in an embodiment of the present invention;

[0032] Figure 2 for Figure 1 The method flow chart of step S102 in FIG.

[0033] Figure 3 This is a flowchart of Example 2 of a method for generating Chinese model essays based on deep learning in an embodiment of the present invention. DETAILED DESCRIPTION

[0034] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0035] Figure 1 FIG. 1 is a flow chart of a method for generating a Chinese model essay based on deep learning in an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0036] S101: classifying a number of sample articles crawled from a target website according to the sample article type to obtain sample article collections of different types;

[0037] In this embodiment, the types of model essays may include: application model essays, emotional model essays, scientific research model essays, etc., and the target website is a website related to the article, such as HowNet, Wanfang, VIP, Baidu, etc.

[0038] S102: constructing a model of each essay in each essay collection;

[0039] In this embodiment, step S102 may include:

[0040] S201: extracting key items and corresponding item attributes of each sample article in each sample article collection;

[0041] In this step, the key items may include: essay theme, essay key words, essay technical content, essay language modification, essay description fluency, etc.; the corresponding item attributes may include: subject attributes, keyword attributes, technical attributes, language modification attributes, description fluency attributes, etc. When extracting all essay key items corresponding to each essay in the same essay collection, based on the essay attribute calibration model, corresponding item attributes may be assigned to each essay key item in each essay.

[0042] S202: Construct a model of each sample article based on all item attributes of each sample article.

[0043] In this step, the sample document model includes entry attributes and their attribute values.

[0044] S103: For each essay set, the essay models corresponding to all the essays therein are fused into a class fusion model corresponding to the essay set;

[0045] In this embodiment, the class fusion model can be a single fusion model of application class, emotion class, scientific research class, etc., or it can be a combined fusion model, such as an application class and emotion class fusion model. For example: there is a model A for generating application class essays, and there is a model B for generating emotion class essays. By fusing model A with model B, model C is generated, and model C can generate both application class essays and emotion class essays.

[0046] S104: receiving Chinese writing information input by a user; the Chinese writing information at least includes a writing topic and a writing type;

[0047] S105: Calculating the matching value between the Chinese writing information and various fusion models according to a preset matching value algorithm;

[0048] S106: Output the class fusion model corresponding to the calculated maximum matching value and the example in the example set.

[0049] In this embodiment, the matching values ​​between the Chinese writing information and various fusion models are calculated, and then a model essay in the model essay set corresponding to the fusion model with the largest matching value is obtained. This model essay is the model essay expected by the user.

[0050] The embodiment of the present invention provides a method for generating Chinese model essays based on deep learning, which crawls model essays from a target website and classifies them to obtain different types of model essay collections, and determines the corresponding item attributes by determining the model key items of each model essay, thereby constructing a model essay model for each model essay, and then generating a corresponding class fusion model from the model essay model in the model essay collection. Finally, after obtaining the Chinese writing information such as the language subject and writing type input by the user, the matching value between the Chinese writing information and various fusion models is calculated, and finally the model essay in the model essay collection corresponding to the class fusion model with the highest matching value is displayed to the user, ensuring that the model essay displayed to the user is the model essay expected by the user.

[0051] Figure 3 FIG. 2 is a flow chart of a second embodiment of a method for generating a Chinese model essay based on deep learning in an embodiment of the present invention. Figure 3 As shown, the method comprises the following steps:

[0052] S301: Classifying a number of sample articles crawled from a target website according to the sample article type to obtain sample article collections of different types;

[0053] S302: extracting key items and corresponding item attributes of each sample essay in each sample essay collection;

[0054] S303: constructing a model of each sample article according to all item attributes of each sample article;

[0055] S304: For each essay collection, according to the essay model of each essay in the current essay collection, the item attributes of each essay, and the preset essay fusion value algorithm, the fusion value of the essay models of all essays in the current essay collection is calculated;

[0056] In this embodiment, the preset model essay fusion value algorithm is:

[0057]

[0058] Among them, R is the fusion value of the essay model of all essays in the current essay set, and M is the total number of essays in the current essay set; Indicates the model fusion factor of the j-th model in the current model set, and the value range is [0.5, 0.8]. The larger the value, the better the fusion, and the smaller the value, the worse the fusion; Indicates the model fusion factor of the j+1th model in the current model set, and the value range is [0.5, 0.8]. The larger the value, the better the fusion, and the smaller the value, the worse the fusion; Represents the standard fusion factor of all model essays in the current model essay set, is a preset value; exp() represents an exponential function; H is the total number of entry attributes corresponding to the jth model in the current model set; η h is the attribute value of the hth item attribute in the jth example model in the current example set; η represents the average attribute value of all item attributes in the jth example model in the current example set.

[0059] S305: Determine whether the fusion value of the model corresponding to the current model set is less than the preset fusion value, if yes, execute step S307, otherwise execute step S306;

[0060] S306: crawling a number of sample articles from the target website and executing step S301;

[0061] S307: Generate a corresponding class fusion model based on all the essays in the current essay collection and the preset essay generation model;

[0062] S308: receiving Chinese writing information input by the user; the Chinese writing information at least includes a writing topic and a writing type;

[0063] S309: Calculating the matching value between the Chinese writing information and various fusion models according to a preset matching value algorithm;

[0064] In this embodiment, the preset matching value algorithm is:

[0065]

[0066] Among them, P i represents the matching value between the Chinese writing information input by the user and the i-th class fusion model; N represents the total number of pre-obtained class fusion models; β i is the correlation value between the writing topic and the i-th class fusion model, β i The value range of is [0,1], where the larger the correlation value, the higher the correlation between the writing topic and the i-th class fusion model; i is the association value between the writing type and the i-th class fusion model, δ i The value range of F is [0.8, 1], where the larger the correlation value, the higher the correlation between the writing type and the i-th class fusion model; i is the model trust value of the i-th class fusion model, and its value range is [3,5]. The larger the model trust value is, the higher the model trust of the i-th class fusion model is.

[0067] S310: Output the class fusion model corresponding to the calculated maximum matching value and the example in the example set.

[0068] The embodiment of the present invention provides a method for generating Chinese model essays based on deep learning. The method crawls model essays from a target website and classifies them to obtain model essay collections of different types. The method determines the corresponding item attributes by determining the model key items of each model essay, and then constructs a model essay model for each model essay. Then, the fusion value of the model essay models of all the model essays in the model essay collection is calculated according to a preset model essay fusion value algorithm. When the fusion value of the model essay model is less than the preset fusion value, the model essay models in the model essay collection are generated into a corresponding class fusion model according to all the model essays in the model essay collection and the preset model essay generation model. When the fusion value of the model essay model is not less than the preset fusion value, the model essays are crawled from the target website, and the model essay collection is reconstructed to obtain the class fusion model, thereby ensuring the reliability and accuracy of the constructed model essay model. Finally, after obtaining the Chinese writing information such as the language subject and writing type input by the user, the matching value of the Chinese writing information and the various fusion models is calculated, and finally the model essay in the model essay collection corresponding to the class fusion model with the highest matching value is displayed to the user, thereby ensuring that the model essay displayed to the user is the model essay expected by the user.

[0069] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0070] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0071] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0072] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0073] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for generating Chinese model essays based on deep learning, characterized in that: include: Classify several sample articles crawled from the target website according to the sample article type to obtain sample article collections of different types; Construct a model of each essay in each essay collection; For each essay set, the essay models corresponding to all the essays in it are fused into a class fusion model corresponding to the essay set; Receiving Chinese writing information input by a user; the Chinese writing information at least includes a writing topic and a writing type; Calculate the matching value between the Chinese writing information and various fusion models according to a preset matching value algorithm; Output the class fusion model corresponding to the calculated maximum matching value and the example in the example set; The step of constructing a model for each essay in each essay collection includes: Extract the key items and their corresponding item attributes of each essay in each essay collection; According to all the item attributes of each sample article, a sample article model is constructed for each sample article; The step of fusing the model models corresponding to all the model essays in each model essay set into a class fusion model corresponding to the model essay set includes: For each essay collection, according to the essay model of each essay in the current essay collection, the item attributes of each essay and the preset essay fusion value algorithm, the fusion value of the essay models of all essays in the current essay collection is calculated; Determine whether the fusion value of the model corresponding to the current model set is less than the preset fusion value; If the fusion value of the essay model corresponding to the current essay set is less than the preset fusion value, generating a corresponding class fusion model based on all essays in the current essay set and the preset essay generation model; If the fusion value of the model essay model corresponding to the current model essay set is not less than the preset fusion value, crawling a number of model essays from the target website, and re-classifying the model essays crawled from the target website according to the model essay type to obtain model essay sets of different types; Wherein, the preset matching value algorithm is: Among them, P i represents the matching value between the Chinese writing information input by the user and the i-th class fusion model; N represents the total number of pre-obtained class fusion models; β i is the correlation value between the writing topic and the i-th class fusion model, β i The value range of is [0,1]; i is the association value between the writing type and the i-th class fusion model, δ i The value range of F is [0.8,1]; i is the model trust value of the i-th class fusion model, and its value range is [3,5].

2. The method for generating Chinese model essays based on deep learning according to claim 1, characterized in that: The preset model fusion value algorithm is: Among them, R is the fusion value of the essay model of all essays in the current essay set, and M is the total number of essays in the current essay set; Represents the model fusion factor of the j-th model in the current model set, and its value range is [0.5, 0.8]; Represents the model fusion factor of the j+1th model in the current model set, and its value range is [0.5, 0.8]; Represents the standard fusion factor of all model essays in the current model essay set, is a preset value; exp() represents an exponential function; H is the total number of entry attributes corresponding to the jth model in the current model set; η h is the attribute value of the hth item attribute in the jth example model in the current example set; η represents the average attribute value of all item attributes in the jth example model in the current example set.

3. The method for generating Chinese model essays based on deep learning according to claim 1, characterized in that: The target website is a website related to the article.

4. The method for generating Chinese model essays based on deep learning according to claim 1, characterized in that: The key items of the model essay include: model essay theme, model essay key words, model essay technical content, model essay language modification, and model essay description fluency; the corresponding item attributes include: subject attributes, keyword attributes, technical attributes, language modification attributes, and description fluency attributes.

Images