Abnormal script rewriting method and device, electronic equipment and storage medium

Abstract

The embodiment of the application provides an abnormal text rewriting method and device, an electronic device and a storage medium, and belongs to the technical field of text processing and is suitable for the field of financial technology. The method comprises the following steps: obtaining text evaluation guide data; performing feature extraction on preset target text data to obtain text features; performing text classification on the target text data based on the text features to obtain a text category; performing screening on a preset review instruction template library based on the text category and the text features to obtain a text review instruction text; performing information extraction on the text evaluation guide data based on the text review instruction text to obtain a text evaluation index; performing text content anomaly detection on the target text data based on the text evaluation index to obtain text content anomaly information; and performing text rewriting on the target text data based on the text content anomaly information. The embodiment of the application can improve the accuracy of abnormal text rewriting.

Description

Technical Field

[0001] The present application relates to the field of text processing technology, and is applicable to the field of financial technology, and in particular to a method and device for rewriting abnormal text, an electronic device, and a storage medium. Background Art

[0002] Copy review is the inspection and evaluation of copy content to ensure its accuracy, compliance, clarity and appropriateness. For example, in the financial field, by reviewing financial product promotional copy, anomalies in the financial product promotional copy can be discovered, making it easier for financial industry staff to rewrite the financial product promotional copy to improve its accuracy, compliance, clarity and appropriateness.

[0003] At present, the common method of rewriting abnormal copywriting is to modify the abnormal copywriting found by manual review. This method is easily affected by subjective factors and it is difficult to ensure the accuracy of the copywriting, which may lead to errors in the rewritten copywriting. Therefore, how to improve the accuracy of abnormal copywriting has become a technical problem that needs to be solved urgently. Summary of the Invention

[0004] The main purpose of the embodiments of the present application is to provide a method and device for rewriting abnormal text, an electronic device and a storage medium, aiming to improve the accuracy of abnormal text rewriting.

[0005] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application provides a method for rewriting abnormal text, the method comprising:

[0006] Obtain copywriting evaluation guide data;

[0007] Perform feature extraction on the preset target copy data to obtain copy features;

[0008] Classifying the target copy data based on the copy features to obtain a copy category;

[0009] Performing a copy analysis on the copy features to obtain basic copy information, and screening a preset review instruction template library based on the copy category and the basic copy information to obtain a copy review instruction text;

[0010] Based on the copy review instruction text, extract information from the copy evaluation guide data to obtain copy evaluation indicators;

[0011] Based on the copy evaluation index, the target copy data is subjected to copy content anomaly detection to obtain copy content anomaly information;

[0012] Based on the abnormal information of the copy content, the target copy data is rewritten.

[0013] In some embodiments, performing copy content anomaly detection on the target copy data based on the copy evaluation index to obtain copy content anomaly information includes:

[0014] Performing sentiment analysis on the target copywriting data to obtain the copywriting sentiment;

[0015] Based on the copy evaluation index, the copy emotion is evaluated to obtain abnormal copy emotion information;

[0016] Performing word segmentation processing on the target copywriting data to obtain copywriting phrases;

[0017] Based on the copy evaluation index, performing phrase anomaly detection on the copy phrase to obtain copy abnormal phrase information;

[0018] Based on the copy phrases and the copy evaluation indicators, the target copy data is subjected to grammatical anomaly detection to obtain copy abnormal grammatical information;

[0019] The abnormal sentiment information of the copy, the abnormal phrase information of the copy and the abnormal grammatical information of the copy are merged to obtain the abnormal content information of the copy.

[0020] In some embodiments, performing phrase anomaly detection on the copy phrases based on the copy evaluation index to obtain copy abnormal phrase information includes:

[0021] Performing phrase interpretation on the copywriting phrase to obtain the phrase meaning;

[0022] Based on the meaning of the phrase, the copywriting phrase is screened to obtain a target phrase;

[0023] Based on the text evaluation index, the target phrase is detected for phrase normativity to obtain the abnormal phrase information of the text.

[0024] In some embodiments, the step of performing grammatical anomaly detection on the target copy data based on the copy phrases and the copy evaluation index to obtain copy grammatical anomaly information includes:

[0025] Performing phrase analysis on the copywriting phrase to obtain the phrase part of speech;

[0026] Based on the part of speech of the phrase, the target copywriting data is parsed to obtain the copywriting grammar;

[0027] Based on the text evaluation index, the text grammar is checked for grammatical standardization to obtain the text abnormal grammatical information.

[0028] In some embodiments, the script classification of the target script data based on the script features comprises:

[0029] The script features are mapped to obtain feature categories and category confidence, wherein the category confidence is used to represent the credibility of the script features being the feature categories;

[0030] The script categories are selected based on the category confidence script features to obtain the script categories.

[0031] In some embodiments, the information extraction of the script evaluation guideline data based on the script review instruction text comprises:

[0032] The abnormal script rewriting key points are obtained by performing review information extraction on the script review instruction text;

[0033] The script evaluation indicators are obtained by performing information screening on the script evaluation guideline data based on the abnormal script rewriting key points.

[0034] In some embodiments, the script rewriting of the target script data based on the script content abnormal information comprises:

[0035] The script context information is obtained by performing natural language processing on the target script data;

[0036] The abnormal information position is obtained by performing abnormal positioning on the target script data based on the script content abnormal information;

[0037] The standard script is obtained by performing script modification on the abnormal information position based on the script context information.

[0038] To achieve the above-mentioned purpose, a second aspect of the embodiments of the present application proposes an abnormal script rewriting device, the device comprises:

[0039] The evaluation guideline acquisition module is used to acquire script evaluation guideline data;

[0040] The script category prediction module is used to perform script classification on the preset target script data to obtain script categories;

[0041] The instruction text generation module is used to generate script review instruction text based on the script categories;

[0042] The evaluation index extraction module is used to perform information extraction on the script evaluation guideline data based on the script review instruction text to obtain script evaluation indicators;

[0043] A copy anomaly detection module, configured to perform copy content anomaly detection on the target copy data based on the copy evaluation index to obtain copy content anomaly information;

[0044] The abnormal copy rewriting module is used to rewrite the target copy data based on the abnormal copy content information.

[0045] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

[0046] To achieve the above-mentioned purpose, the fourth aspect of the embodiments of the present application proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the method described in the first aspect.

[0047] The abnormal copy rewriting method and device, electronic device and storage medium proposed in this application establish the review standard of the copy by obtaining copy evaluation guide data, then perform feature extraction on the preset target copy data to obtain copy features, and classify the target copy data according to the copy features to identify the field to which the target copy data belongs, then perform copy analysis on the copy features to obtain basic copy information, and based on the copy category and basic copy information, screen the preset review instruction template library to obtain the copy review instruction text, and provide targeted instructions for the copy review system, then extract key copy evaluation indicators from the copy evaluation guide data, use the copy evaluation indicators to perform anomaly detection on the target copy data, identify non-compliant or inaccurate content, and finally, perform necessary rewriting on the copy based on the detection results to ensure its accuracy, compliance, clarity and appropriateness, thereby reducing the subjectivity of manual review and improving the accuracy of abnormal copy rewriting. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a flowchart of the abnormal text rewriting method provided by an embodiment of the present application;

[0049] Figure 2 yes Figure 1 Flowchart of step S102 in FIG.

[0050] Figure 3 yes Figure 1 Flowchart of step S105 in FIG.

[0051] Figure 4 yes Figure 1 Flowchart of step S106 in FIG.

[0052] Figure 5 yes Figure 4 Flowchart of step S404 in FIG.

[0053] Figure 6 yes Figure 4 Flowchart of step S405 in FIG.

[0054] Figure 7 yes Figure 1 Flowchart of step S107 in FIG.

[0055] Figure 8 Schematic diagram of the structure of the abnormal text rewriting device provided in an embodiment of the present application;

[0056] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0058] It should be noted that although the device schematics illustrate functional module divisions and the flowcharts illustrate logical sequences, in certain circumstances, the steps shown or described may be performed in a sequence that differs from the module divisions in the device or the sequence in the flowcharts. The terms "first," "second," and so on, in the specification, claims, and drawings, are used to distinguish similar items and are not necessarily used to describe a specific sequence or precedence.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0060] First, let’s analyze some of the terms used in this application:

[0061] The Copy Review System: The Copy Review System utilizes advanced natural language processing technology specifically designed for the financial sector, aiming to automate and accurately verify the compliance of copywriting. The system first acquires and analyzes data from the financial industry's copywriting evaluation guidelines. It then intelligently categorizes the target copywriting data to determine its category, such as investment products or insurance services. Based on the copywriting category, the system generates customized copywriting review instructions, guiding the system to extract key copywriting evaluation metrics from the evaluation guidelines. The system then applies these metrics to conduct in-depth anomaly detection on the copywriting, including sentiment analysis, word segmentation, phrase anomaly detection, and grammatical anomaly detection, to identify and locate potential issues within the copywriting. If anomalies are detected, such as overly optimistic sentiment or non-compliant phrases, the system automatically or recommends rewriting to ensure that the copy meets industry standards in terms of sentiment expression, word accuracy, and grammatical standardization. Through this process, the Copy Review System not only improves the compliance and professionalism of the copywriting, but also reduces the subjectivity of manual review, improving efficiency and consistency, thereby effectively safeguarding the quality of financial copywriting and protecting consumer rights.

[0062] Natural language processing technology: Natural language processing technology is an important branch of artificial intelligence. It focuses on enabling computers to understand, interpret, generate and respond to human language, including text and voice. This technology involves a variety of algorithms and models, and can perform tasks such as text analysis, sentiment analysis, machine translation, speech recognition and generation. The core challenge of natural language processing technology lies in the complexity of language, including grammar, semantics, context and cultural differences. Through deep learning, machine learning and other advanced computing methods, natural language processing systems are able to learn language patterns from large amounts of data and continuously optimize their performance to achieve more natural and accurate language interactions. In many fields such as finance, healthcare, and customer service, natural language processing technology is widely used to improve efficiency, enhance user experience and promote automated decision-making processes.

[0063] Copy review is the inspection and evaluation of copy content to ensure its accuracy, compliance, clarity and appropriateness. For example, in the financial field, by reviewing financial product promotional copy, anomalies in the financial product promotional copy can be discovered, making it easier for financial industry staff to rewrite the financial product promotional copy to improve its accuracy, compliance, clarity and appropriateness.

[0064] At present, the common method of rewriting abnormal copywriting is to modify the abnormal copywriting found by manual review. This method is easily affected by subjective factors and it is difficult to ensure the accuracy of the copywriting, which may lead to errors in the rewritten copywriting. Therefore, how to improve the accuracy of abnormal copywriting has become a technical problem that needs to be solved urgently.

[0065] Based on this, the embodiments of the present application provide a method and device for rewriting abnormal text, an electronic device, and a storage medium, aiming to improve the accuracy of abnormal text rewriting.

[0066] The abnormal text rewriting method and device, electronic device and storage medium provided in the embodiments of the present application are specifically illustrated through the following embodiments. First, the abnormal text rewriting method in the embodiments of the present application is described.

[0067] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to achieve optimal results.

[0068] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0069] The abnormal text rewriting method provided in the embodiment of the present application relates to the field of text processing technology and is applicable to the field of financial technology. The abnormal text rewriting method provided in the embodiment of the present application can be applied in a terminal, can be applied in a server side, or can be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or as a server cluster or distributed system composed of multiple physical servers, or as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the abnormal text rewriting method, etc., but is not limited to the above forms.

[0070] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0071] Figure 1 This is an optional flow chart of the abnormal copy rewriting method provided in the embodiment of the present application, which is applied to the copy review system. Figure 1 The method may include but is not limited to steps S101 to S107.

[0072] Step S101, obtaining copywriting evaluation guide data;

[0073] Step S102: extracting features from the preset target copy data to obtain copy features;

[0074] Step S103, classifying the target copy data based on the copy features to obtain a copy category;

[0075] Step S104: Analyze the document features to obtain basic document information, and based on the document category and basic document information, filter the preset review instruction template library to obtain the document review instruction text;

[0076] Step S105: extract information from the copywriting evaluation guide data based on the copywriting review instruction text to obtain copywriting evaluation indicators;

[0077] Step S106: Based on the copy evaluation index, the target copy data is subjected to copy content anomaly detection to obtain copy content anomaly information;

[0078] Step S107: rewrite the target copy data based on the copy content abnormality information.

[0079] In steps S101 to S107 shown in the embodiment of the present application, the copy review system obtains copy evaluation guide data, performs feature extraction on preset target copy data to obtain copy features, and classifies the target copy data according to the copy features to obtain copy categories. Furthermore, the copy features are analyzed to obtain basic copy information, and based on the copy categories and basic copy information, a preset review instruction template library is screened to obtain a copy review instruction text. Secondly, based on the copy review instruction text, information is extracted from the copy evaluation guide data to obtain copy evaluation indicators. Then, based on the copy evaluation indicators, the target copy data is detected for copy content anomaly to obtain copy content anomaly information. Finally, based on the copy content anomaly information, the target copy data is rewritten. Therefore, this application establishes the review standards for the copy by obtaining the copy evaluation guide data, then performs feature extraction on the preset target copy data to obtain the copy features, and classifies the target copy data according to the copy features to identify the field to which the target copy data belongs, then performs copy analysis on the copy features to obtain the basic copy information, and based on the copy category and the basic copy information, screens the preset review instruction template library to obtain the copy review instruction text to provide targeted instructions for the copy review system, then extracts key copy evaluation indicators from the copy evaluation guide data, uses the copy evaluation indicator indicators to perform anomaly detection on the target copy data, identifies non-compliant or inaccurate content, and finally, performs necessary rewriting on the copy based on the detection results to ensure its accuracy, compliance, clarity and appropriateness, thereby reducing the subjectivity of manual review and improving the accuracy of abnormal copy rewriting.

[0080] In step S101 of some embodiments, the copy evaluation guide data refers to the rules and standards used to guide the copy review process. For example, in the financial copy scenario, the copy evaluation guide data can be insurance industry regulatory rules, advertising laws, internal company legal compliance requirements, etc.

[0081] This application can crawl the laws and regulations on the terminology evaluation documents on the Internet through a pre-set web crawler script to obtain document evaluation information, and then use the document review system to organize the document evaluation information to obtain document evaluation guide data.

[0082] In step S102 of some embodiments, the target copy data refers to a newly written copy awaiting review, for example, a health insurance promotion copy just written by an insurance salesperson in an insurance promotion scenario. The copy category can be determined based on the content of the target copy data. For example, if the target copy data primarily describes the advantages of health insurance, the target copy data may be classified as a health insurance promotion copy.

[0083] In an embodiment of the present application, the copy review system can use natural language processing technology to perform content analysis on the target copy data, thereby extracting the copy features of the target copy data. Among them, natural language processing technology includes text segmentation, named entity recognition, grammatical parsing, semantic parsing and other processes. For a target copy data about "high-yield investment plan", feature extraction may identify keywords such as "high yield", "investment plan", "return rate", and named entities such as "financial products" and "market trends".

[0084] It should be noted that the copy features extracted by the above-mentioned natural language processing technology can be keywords, phrases, named entities, semantic roles, etc. in the target copy data, as well as the distribution and frequency of keywords, phrases, named entities, semantic roles, etc. in the target copy data.

[0085] In step S103 of some embodiments, a category confidence can be assigned to each copy feature through a deep learning model. This confidence reflects the degree of credibility of the copy feature belonging to a certain feature category. Subsequently, the copy review system screens the copy features according to these category confidences, selects feature categories with higher confidence, and finally determines the copy category of the target copy data.

[0086] For details, see Figure 2 In some embodiments, step S103 may include but is not limited to steps S201 to S202:

[0087] Step S201: feature mapping is performed on the document features to obtain feature categories and category confidences, wherein the category confidences are used to indicate the degree of confidence that the document features are feature categories;

[0088] Step S202: Based on the category confidence copy feature, a feature category is selected to obtain a copy category.

[0089] In steps S201 and S202 of some embodiments, after obtaining the copy features in the target copy data, the copy review system can predict the category of the target copy data through a machine learning model or a deep learning model. In detail, the machine learning model or the deep learning model can analyze the copy features and, in combination with the text information of the target copy data, predict the probability that the target copy data belongs to a certain feature category, that is, the category confidence, and then use the feature category with the highest category confidence as the copy category of the target copy data. For example, in the financial field, a trained deep learning model may classify the above-mentioned "high-yield investment plan" target copy data into the "investment product promotion" category.

[0090] In steps S201 to S202 shown in this embodiment, feature mapping is performed on the copy features to obtain feature categories and category confidences, wherein the category confidence is used to indicate the degree of credibility of the copy features as feature categories. Furthermore, based on the category confidence copy features, feature categories are selected to obtain copy categories, thereby reducing errors and subjective biases in manual copy classification, and by ensuring the objectivity and consistency of copy classification, providing a solid foundation for generating customized copy review instruction texts. In addition, the automated copy classification of the copy review system improves the accuracy of copy content anomaly detection, thereby facilitating timely discovery and correction of non-compliant content in the copy, and ensuring that the copy complies with industry standards and regulatory requirements before publication.

[0091] In step S104 of some embodiments, the review instruction template library is a pre-built database for storing review instruction templates, and the review instruction template library contains review instruction templates for different document categories, wherein the review instruction template is used to guide the document review system to perform document review to ensure the content compliance and accuracy of the target document.

[0092] This application can use natural language processing technology to conduct in-depth analysis of the target copy and extract key basic copy information, such as core keywords, phrases and named entities. Then, based on the above copy category, the copy review system can filter out the review instruction text template that is most relevant to the copy category from the pre-built review instruction template library, where the review instruction text template contains the review points and standards for the copy category. Then, combined with the basic copy information of the target copy, the screened review instruction template is customized and adjusted to generate the final copy review instruction text. The copy review instruction text will guide the review procedure in the copy review system to conduct a detailed review of the target copy to ensure the content compliance, accuracy and appropriateness of the target copy.

[0093] In step S105 of some embodiments, the subsequent program body in the copy review system can obtain the key points of abnormal copy rewriting by extracting key information from the copy review instruction text, and then filter the copy evaluation information in the above-mentioned copy evaluation guide data based on the key points of abnormal copy rewriting to obtain the copy evaluation indicators.

[0094] For details, see Figure 3 In some embodiments, step S105 may include but is not limited to steps S301 to S302:

[0095] Step S301: extracting audit information from the document review instruction text to obtain key points for rewriting abnormal documents;

[0096] Step S302 : Based on the abnormal copy rewriting key points, the copy evaluation guide data is screened to obtain copy evaluation indicators.

[0097] In step S301 of some embodiments, by performing text parsing and context understanding on the copy review instruction text, potential problems of the target copy data described in the copy review instruction text can be identified. Furthermore, the potential problems of the target copy data are sorted out to obtain key points for rewriting abnormal copy.

[0098] It is important to note that abnormal key points in copywriting can be non-compliant statements, misleading information, or content that requires further clarification in the target copywriting data. For example, when the copywriting review instruction text points out that attention should be paid to the statement of "return on investment", then the "guaranteed high returns" in the target copywriting data may be considered as abnormal copywriting rewriting key points that mislead investors into believing that the investment is risk-free.

[0099] In step S302 of some embodiments, after extracting the abnormal copy rewriting key points in the copy review instruction text, the copy review system can filter out copy evaluation information related to the abnormal copy rewriting key points from the copy evaluation guide data, thereby obtaining copy evaluation indicators. For example, for the above-mentioned abnormal copy rewriting key point "guarantee high returns", the copy evaluation guide data may provide an indicator of "avoiding the use of absolute or misleading expressions". At this time, the copy evaluation indicator may include "ensuring that all investment-related expressions comply with the principle of authenticity and do not use words such as "guarantee" and "absolute"."

[0100] In steps S301 to S302 shown in this embodiment, by extracting audit information from the copy review instruction text, key points for rewriting abnormal copies are obtained. Furthermore, based on the key points for rewriting abnormal copies, information on the copy evaluation guide data is screened to obtain copy evaluation indicators, which greatly improves the pertinence and efficiency of copy review, reduces subjective judgment errors in the manual review process, and enhances the compliance and transparency of the copy.

[0101] In step S106 of some embodiments, by performing natural language processing on the target copy data, copy information such as the copy sentiment, copy phrases and copy grammar of the target copy data can be obtained. Then, based on the copy evaluation indicators, the copy anomaly evaluation is performed on the above copy information such as the copy sentiment, copy phrases and copy grammar to determine whether the target copy data has copy sentiment anomalies, copy phrase anomalies and copy grammar anomalies. Furthermore, the copy review system summarizes the above judgment results to obtain copy content anomaly information.

[0102] For details, see Figure 4In some embodiments, step S106 may include but is not limited to steps S401 to S406:

[0103] Step S401, performing sentiment analysis on the target copy data to obtain the copy sentiment;

[0104] Step S402: Evaluate the sentiment of the copy based on the copy evaluation index to obtain abnormal sentiment information of the copy;

[0105] Step S403: performing word segmentation processing on the target copywriting data to obtain copywriting phrases;

[0106] Step S404: Based on the copy evaluation index, perform phrase anomaly detection on the copy phrases to obtain copy abnormal phrase information;

[0107] Step S405: Based on the copy phrases and copy evaluation indicators, perform grammatical anomaly detection on the target copy data to obtain copy abnormal grammatical information;

[0108] Step S406: Merge the abnormal sentiment information, abnormal phrase information, and abnormal grammatical information of the copy to obtain abnormal content information of the copy.

[0109] In step S401 of some embodiments, sentiment analysis is to identify and extract the sentiment tendency in the text through natural language processing technology.

[0110] This application can use sentiment recognition tools or algorithms to perform phrase analysis on words or phrases in the target copy data, so as to determine the sentiment tendency of the target copy data, that is, the copy sentiment, where the sentiment tendency includes positive, negative or neutral. For example, for the sentence "Invest in our fund and enjoy unprecedented growth" in the target copy data of financial product promotion, the sentiment recognition tool may identify "enjoy" and "growth" as positive sentiment words.

[0111] In step S402 of some embodiments, the text sentiment evaluation refers to further determining whether the emotional expression of the target text data is appropriate and whether it meets the compliance requirements of the text evaluation indicators based on the results of the sentiment analysis.

[0112] This application can determine the compliant copywriting sentiment tendency based on the copywriting evaluation indicators, and then compare the above copywriting sentiment with the compliant copywriting sentiment tendency to determine whether the copywriting sentiment is abnormal, and generate copywriting abnormal sentiment information when the copywriting sentiment is abnormal. For example, according to the copywriting evaluation indicators of the financial industry, the target copywriting data must not be overly optimistic or pessimistic. If words or phrases such as "enjoy" and "unprecedented growth" that indicate overly optimistic emotions appear in the target copywriting data, it can be determined that the copywriting sentiment of the target copywriting data is abnormal, thereby generating copywriting abnormal sentiment information.

[0113] It is important to know that the abnormal emotional information of the copy includes the words or phrases with incorrect emotional content, and the reasons for the errors are marked.

[0114] In step S403 of some embodiments, word segmentation processing is to decompose the target text data into separate words or phrases so as to perform a more detailed analysis of the target text data.

[0115] This application can use a word segmentation algorithm to decompose the target copy data to obtain separate words or phrases, i.e., copy phrases. For example, by performing word segmentation processing on the target copy data "Our investment services provide personalized consulting", the target copy data can be decomposed into copy phrases such as "our", "investment services", "provide", and "personalized consulting".

[0116] In step S404 of some embodiments, phrase anomaly detection is to identify whether a copywriting phrase in the target copywriting data meets the copywriting evaluation index.

[0117] This application can determine the meaning of the copywriting phrases by performing a detailed interpretation of the copywriting phrases, and then screen out target phrases related to the copywriting evaluation indicators based on the phrase meanings. Finally, the target phrases are subjected to a normative test to identify abnormal phrases that do not meet the copywriting evaluation indicators, and copywriting abnormal phrase information is generated based on the abnormal phrases.

[0118] For details, see Figure 5 In some embodiments, step S404 may include but is not limited to steps S501 to S503:

[0119] Step S501, interpreting the copywriting phrases to obtain the meaning of the phrases;

[0120] Step S502: Screen the copywriting phrases based on the phrase meanings to obtain target phrases;

[0121] Step S503: Based on the copywriting evaluation index, the target phrase is checked for phrase normativity to obtain abnormal copywriting phrase information.

[0122] In step S501 of some embodiments, the present application may utilize semantic recognition tools to interpret each copywriting phrase in the target copywriting data to determine the meaning of the copywriting phrase in a specific context, i.e., the phrase meaning. For example, in financial target copywriting data, the phrase meaning of the phrase "high-yield investment" may be an investment opportunity that promises high returns.

[0123] In step S502 of some embodiments, based on the phrase meanings of the above-mentioned copywriting phrases, the copywriting review system can filter out target phrases related to the copywriting evaluation indicators. For example, in the financial target copywriting data, if there are copywriting phrases such as "risk-free", "high-yield investment", and "guaranteed return", then target phrases such as "risk-free" and "guaranteed return" that violate the compliance requirements of financial advertisements may be filtered out.

[0124] In step S503 of some embodiments, the selected target phrases may be tested using the copywriting evaluation indicators to determine whether the target phrases meet the specifications and standards of the copywriting evaluation indicators. The detected abnormal target phrases are then summarized to form abnormal copywriting phrase information. For example, if the target copywriting data of the financial category contains the target phrase "risk-free investment", this target phrase will be considered an abnormal target phrase according to the copywriting evaluation indicators of the financial industry.

[0125] In steps S501 to S503 shown in this embodiment, the copywriting phrases in the target copywriting data are interpreted in detail to determine the phrase meanings of the copywriting phrases. Then, based on the phrase meanings, target phrases related to the target copywriting data evaluation indicators are screened out. Finally, the target phrases are subjected to a normative test to identify abnormal phrases that do not meet the financial industry standards or regulatory requirements, thereby obtaining copywriting abnormal phrase information, which can ensure the compliance and accuracy of the copywriting.

[0126] In step S405 of some embodiments, grammatical anomaly detection is to identify whether the grammar of the target copy data meets the copy evaluation index.

[0127] This application parses the copywriting phrases to determine the part of speech of each copywriting phrase, then uses the part of speech to perform grammatical analysis on the entire target copywriting data to construct the grammatical structure of the target copywriting data, and finally performs a normative test on the grammatical structure based on the copywriting evaluation indicators, thereby identifying abnormal grammatical structures that do not conform to language specifications or industry standards, and then generates copywriting abnormal grammatical information based on the abnormal grammatical structure.

[0128] For details, see Figure 6 In some embodiments, step S405 may include but is not limited to steps S601 to S603:

[0129] Step S601, parsing the copywriting phrase to obtain the phrase part of speech;

[0130] Step S602: parse the target text data based on the phrase part of speech to obtain the text grammar;

[0131] Step S603: Based on the copy evaluation index, the copy grammar is checked for grammatical standardization to obtain abnormal grammatical information of the copy.

[0132] In step S601 of some embodiments, the copy phrases can be parsed using natural language processing technology to determine the part-of-speech information of each copy phrase, i.e., the phrase part-of-speech. It should be noted that the phrase part-of-speech can be a noun, verb, adjective, or other part-of-speech.

[0133] In step S602 of some embodiments, the target text data may be parsed for grammatical structure based on the above-mentioned phrase parts of speech to generate a grammatical tree of the target text data, and then the grammatical tree may be subjected to grammatical recognition to obtain the text grammar of the target text data.

[0134] In step S603 of some embodiments, the copywriting evaluation index can be used to perform a grammatical structure normative check on the copywriting grammar of the target copywriting data, thereby identifying abnormal copywriting grammar in the target copywriting data. Furthermore, the detected abnormal copywriting grammar is summarized to obtain copywriting abnormal grammar information.

[0135] In steps S601 to S603 shown in this embodiment, the target copywriting data is parsed to determine the part of speech of each copywriting phrase, and then the whole target copywriting data is parsed using the part of speech to construct the grammatical structure of the target copywriting data. Finally, the copywriting grammar is checked for standardization based on the copywriting evaluation indicators, thereby identifying abnormal grammatical structures that do not meet the language specifications or industry standards in the copywriting evaluation indicators, obtaining abnormal grammatical information of the copywriting, and ensuring the grammatical accuracy and professionalism of the copywriting.

[0136] In step S406 of some embodiments, information merging is to integrate abnormal information of copy sentiment, copy phrases and copy grammar to form a comprehensive copy content abnormal information report.

[0137] This application can detect abnormal emotional information in the copywriting 、 Abnormal copywriting sentiment information and abnormal copywriting sentiment information are summarized to generate detailed abnormal copywriting content information. It should be noted that the abnormal copywriting content information can point out the areas that need to be modified in the target copywriting data and provide modification suggestions.

[0138] In steps S401 to S405 shown in this embodiment, sentiment analysis is used to identify the sentiment tendency in the target copy data, and then the appropriateness of the sentiment of the copy is evaluated based on the copy evaluation index. Then, the target copy data is segmented, and the compliance of the copy phrases is identified and tested. At the same time, grammar testing is performed to ensure the grammatical correctness of the target copy data. Finally, the abnormal information of sentiment, phrases and grammar is merged to form a comprehensive copy content abnormality information report, thereby improving the accuracy of abnormal copy rewriting, ensuring that the copy meets the requirements of the copy evaluation index in terms of emotional expression, word accuracy and grammatical standardization, and ensuring the quality and compliance of the copy.

[0139] In step S107 of some embodiments, the target copy data can be parsed using natural language processing technology to obtain copy context information of the target copy data, and then the abnormal part of the target copy data can be accurately located using the above-mentioned copy content abnormal information. Finally, the abnormal part can be modified in combination with the copy context information.

[0140] For details, see Figure 7 In some embodiments, step S107 may include but is not limited to steps S701 to S703:

[0141] Step S701: performing natural language processing on the target copy data to obtain copy context information;

[0142] Step S702: Based on the abnormal information of the copy content, the target copy data is abnormally located to obtain the abnormal information location;

[0143] Step S703: Based on the text context information, modify the text at the abnormal information location to obtain a standard text.

[0144] In step S701 of some embodiments, the copy review system can use mature natural language processing tools to perform entity recognition, relationship recognition and overall context recognition on the target copy data, and integrate and analyze the recognized copy to obtain copy context information, wherein the natural language processing tool can be a sentiment analyzer, a part-of-speech tagger and a syntactic analyzer.

[0145] In step S702 of some embodiments, based on the above-mentioned abnormal copy content information, the copy review system can determine the location where the copy abnormality occurs in the target copy data, that is, the abnormal information location, through information search.

[0146] In step S703 of some embodiments, after obtaining the location of the abnormal information, the copy review system can locate the abnormality in the target copy data, and then modify the abnormal part of the target copy data according to the copy context information and the copy evaluation indicators to obtain a modified copy. Furthermore, the copy review system can re-evaluate the modified copy until the modified copy meets the copy evaluation indicators to obtain a standard copy. For example, in financial target copy data, by modifying "guaranteed high returns" in the target copy data to "expected high returns, but please pay attention to investment risks", the copy modification of the target copy data can be achieved, thereby obtaining a standard copy.

[0147] In one application scenario, taking financial product promotional copy as an example, first, copy evaluation guide data for the financial industry is obtained. This copy evaluation guide data includes industry standards, regulatory requirements, and best practices. Then, a preset financial product promotional copy is classified, for example, as belonging to the "investment product" category. Next, based on the "investment product" category, copy review instructions are generated. These instructions may emphasize the need for special attention to risk disclosure and the accuracy of return forecasts. Based on the copy review instructions, specific copy evaluation indicators, such as "avoiding misleading statements" and "ensuring information transparency," are extracted from the copy evaluation guide data. Next, the copy evaluation indicators are used to detect anomalies in the target copy data, identifying anomalous content such as "guaranteed high returns" that may mislead consumers. Finally, based on the detected anomalies, the target copy data is rewritten, changing "guaranteed high returns" to "expected high returns, but please be aware of investment risks" to ensure the compliance and accuracy of the copy. This solution can improve the accuracy of product promotional copy reviews.

[0148] This application establishes the review standards for the copy by obtaining the copy evaluation guide data, then performs feature extraction on the preset target copy data to obtain the copy features, and classifies the target copy data according to the copy features to identify the field to which the target copy data belongs, then performs copy analysis on the copy features to obtain the basic copy information, and based on the copy category and the basic copy information, screens the preset review instruction template library to obtain the copy review instruction text, and provides targeted instructions for the copy review system, then extracts key copy evaluation indicators from the copy evaluation guide data, uses the copy evaluation indicators to perform anomaly detection on the target copy data, and identifies non-compliant or inaccurate content, and finally, performs necessary rewriting on the copy based on the detection results to ensure its accuracy, compliance, clarity and appropriateness, thereby reducing the subjectivity of manual review and improving the accuracy of abnormal copy rewriting.

[0149] See also Figure 8The embodiment of the present application further provides an abnormal text rewriting device, which can implement the above abnormal text rewriting method, and the device includes:

[0150] Evaluation guide acquisition module 801 is used to obtain copywriting evaluation guide data;

[0151] The document feature extraction module 802 is used to extract features from preset target document data to obtain document features;

[0152] The document category prediction module 803 is used to classify the target document data based on the document features to obtain the document category;

[0153] The instruction text generation module 804 is used to analyze the document features to obtain basic information of the document, and based on the document category and basic information, filter the preset review instruction template library to obtain the document review instruction text;

[0154] Evaluation index extraction module 805 is used to extract information from the copywriting evaluation guide data based on the copywriting review instruction text to obtain the copywriting evaluation index;

[0155] The copy anomaly detection module 806 is used to detect copy content anomalies on the target copy data based on the copy evaluation index to obtain copy content anomaly information;

[0156] The abnormal copy rewriting module 807 is used to rewrite the target copy data based on the abnormal copy content information.

[0157] The specific implementation of the abnormal text rewriting device is basically the same as the specific embodiment of the abnormal text rewriting method described above, and will not be repeated here.

[0158] The present application also provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-mentioned abnormal text rewriting method. The electronic device can be any smart terminal including a tablet computer, an in-vehicle computer, or the like.

[0159] See also Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is shown. The electronic device includes:

[0160] The processor 901 can be implemented as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0161] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 902 can store an operating system and other application programs. When the technical solutions provided by the embodiments of the present specification are implemented by software or firmware, the related program codes are stored in the memory 902 and are invoked and executed by the processor 901 to implement the abnormal case rewriting method of the embodiments of the present application.

[0162] The input / output interface 903 is configured to realize information input and output.

[0163] The communication interface 904 is configured to realize the communication interaction between the device and other devices. The communication can be realized by a wired manner (for example, a USB, a network cable, etc.) or a wireless manner (for example, a mobile network, WIFI, Bluetooth, etc.).

[0164] The bus 905 is configured to transmit information between various components (for example, the processor 901, the memory 902, the input / output interface 903, and the communication interface 904) of the device.

[0165] The processor 901, the memory 902, the input / output interface 903, and the communication interface 904 are connected to each other through the bus 905 to realize the communication connection between the device.

[0166] The embodiments of the present application also provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the above-mentioned abnormal case rewriting method.

[0167] The memory is a non-transitory computer readable storage medium, which can be used to store non-transitory software programs and non-transitory computer executable programs. In addition, the memory can include a high-speed random access memory and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged relative to the processor. These remote memories can be connected to the processor through a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0168] The abnormal document rewriting method, abnormal document rewriting device, electronic device, and storage medium provided in the embodiments of the present application are applied to a document review system. The method obtains document evaluation guide data, then performs feature extraction on preset target document data to obtain document features. The target document data is then classified based on the document features to obtain a document category. Furthermore, the document features are analyzed to obtain basic document information. Based on the document category and basic document information, a preset review instruction template library is screened to obtain the document review instruction text. Next, based on the document review instruction text, information is extracted from the document evaluation guide data to obtain document evaluation indicators. Based on the document evaluation indicators, the target document data is then tested for content anomalies to obtain content anomaly information. Finally, based on the content anomaly information, the target document data is rewritten.

[0169] The embodiments described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0170] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0171] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0172] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0173] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0174] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0175] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the above-mentioned units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. The mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0176] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0177] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0178] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes multiple instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: various media that can store programs, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0179] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A method for rewriting abnormal text, characterized in that: The method comprises: Obtain copywriting evaluation guide data; Perform feature extraction on the preset target copy data to obtain copy features; Classifying the target copy data based on the copy features to obtain a copy category; Performing a copy analysis on the copy features to obtain basic copy information, and screening a preset review instruction template library based on the copy category and the basic copy information to obtain a copy review instruction text; Based on the copy review instruction text, extract information from the copy evaluation guide data to obtain copy evaluation indicators; Performing sentiment analysis on the target copywriting data to obtain the copywriting sentiment; Based on the copy evaluation index, the copy emotion is evaluated to obtain abnormal copy emotion information; Performing word segmentation processing on the target copywriting data to obtain copywriting phrases; Performing phrase interpretation on the copywriting phrase to obtain the phrase meaning; Based on the meaning of the phrase, the copywriting phrase is screened to obtain a target phrase; Based on the copywriting evaluation index, the target phrase is tested for phrase normativity to obtain abnormal phrase information of the copywriting; Performing phrase analysis on the copywriting phrase to obtain the phrase part of speech; Based on the part of speech of the phrase, the target copywriting data is parsed to obtain the copywriting grammar; Based on the copy evaluation index, the copy grammar is tested for grammatical standardization to obtain abnormal grammatical information of the copy; Merging the abnormal sentiment information of the copy, the abnormal phrase information of the copy, and the abnormal grammatical information of the copy to obtain abnormal content information of the copy; Based on the abnormal information of the copy content, the target copy data is rewritten.

2. The method according to claim 1, characterized in that The target document data is classified based on the document features to obtain a document category, including: Performing feature mapping on the document feature to obtain a feature category and a category confidence, wherein the category confidence is used to indicate the degree of confidence that the document feature belongs to the feature category; Based on the category confidence copy feature, the feature category is selected to obtain the copy category.

3. The method according to any one of claims 1-2, characterized in that The step of extracting information from the copywriting evaluation guide data based on the copywriting review instruction text to obtain copywriting evaluation indicators includes: Extracting audit information from the document review instruction text to obtain key points for rewriting abnormal documents; Based on the abnormal copy rewriting key points, the copy evaluation guide data is screened to obtain the copy evaluation index.

4. The method according to any one of claims 1 to 2, characterized in that The rewriting of the target copy data based on the abnormal copy content information includes: Performing natural language processing on the target copy data to obtain copy context information; Based on the abnormal information of the copy content, the target copy data is abnormally located to obtain the abnormal information location; Based on the text context information, the text at the abnormal information location is modified to obtain a standard text.

5. An abnormal text rewriting device, characterized in that: The device comprises: Evaluation guide acquisition module, used to obtain copywriting evaluation guide data; The document feature extraction module is used to extract features from the preset target document data to obtain document features; A document category prediction module, configured to classify the target document data based on the document features to obtain a document category; An instruction text generation module is used to perform text analysis on the text features to obtain basic information of the text, and based on the text category and the basic information, screen a preset review instruction template library to obtain the text review instruction text; An evaluation index extraction module is used to extract information from the copywriting evaluation guide data based on the copywriting review instruction text to obtain the copywriting evaluation index; A copy anomaly detection module is configured to perform sentiment analysis on the target copy data to obtain the copy sentiment, evaluate the copy sentiment based on the copy evaluation index to obtain copy abnormal sentiment information, perform word segmentation processing on the target copy data to obtain copy phrases, perform phrase interpretation on the copy phrases to obtain phrase meanings, screen the copy phrases based on the phrase meanings to obtain target phrases, perform phrase standardization detection on the target phrases based on the copy evaluation index to obtain copy abnormal phrase information, perform phrase analysis on the copy phrases to obtain phrase parts of speech, perform grammatical analysis on the target copy data based on the phrase parts of speech to obtain copy grammar, perform grammatical standardization detection on the copy grammar based on the copy evaluation index to obtain copy abnormal grammatical information, and merge the copy abnormal sentiment information, the copy abnormal phrase information, and the copy abnormal grammatical information to obtain copy content abnormal information; The abnormal copy rewriting module is used to rewrite the target copy data based on the abnormal copy content information.

6. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the abnormal document rewriting method according to any one of claims 1 to 4 when executing the computer program.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the abnormal text rewriting method according to any one of claims 1 to 4 is implemented.

Images