Auxiliary compiling method and device of scheme

Abstract

The invention discloses an auxiliary compiling method and device for a scheme, and relates to the field of artificial intelligence or other related technical fields, and the method comprises the steps: receiving a scheme creation request of a target user in an interaction system, obtaining user information of a target user, scheme information of a target scheme created by the target user and a historical scheme writing record of the target user based on the scheme creation request; performing preprocessing and vector coding processing on the user information, the scheme information and the historical scheme writing record to obtain a coding vector; inputting the coding vector into an auxiliary writing model, and outputting prompt information of each writing process of the target scheme; and pushing prompt information of the target scheme in each writing process to a function module corresponding to each writing process. The technical problem that in the prior art, the error rate is high when a user writes a scheme on various operation platforms is solved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence or other related technical fields, and in particular to a method and device for assisting in writing a solution. Background Art

[0002] Under the current background of informatization and digital transformation, many enterprises or institutions handle related businesses or complete related work through online operation platforms / systems. Enterprise platform systems play an increasingly important role in business process automation, data management, decision support, etc. However, with the continuous enrichment of system functions and the increasing complexity of the operation interface, users are facing more and more challenges when using these systems, especially when filling out various applications, reports, statements and other plans. Filling out plans not only requires users to have certain business knowledge, but also to follow strict format requirements and logical rules. Any negligence in details may lead to the failure of the plan to be reviewed, which requires repeated revisions, seriously affecting work efficiency and user experience.

[0003] In the related art, user guidance of various operating systems mainly relies on static help documents, operation manuals or the user's own work experience. These methods often seem to be inadequate when facing a large number of users and complex business scenarios. On the one hand, static documents and operation manuals are not updated in a timely manner, which makes it difficult to meet the user's needs for real-time and personalized guidance; on the other hand, users need to have rich work experience to ensure that they make as few mistakes as possible when writing various solutions, which places high demands on users. Based on this, most users have a high error rate when writing solutions on various operating platforms, resulting in low user writing efficiency.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiment of the present invention provides a method and device for assisting in writing a solution, so as to at least solve the technical problem in the related art that the error rate is high when users write solutions on various operating platforms.

[0006] According to one aspect of an embodiment of the present invention, a method for assisting the writing of a solution is provided, comprising: receiving a solution creation request from a target user in an interactive system, and obtaining user information of the target user, solution information of a target solution created by the target user, and historical solution writing records of the target user based on the solution creation request; preprocessing and vector encoding the user information, the solution information, and the historical solution writing records to obtain a coding vector; inputting the coding vector into an assisting writing model, and outputting prompt information of each writing process of the target solution, wherein the assisting writing model is a pre-built deep learning model for generating prompt information for user solution writing; and pushing the prompt information of the target solution in each of the writing processes to the functional modules corresponding to each of the writing processes.

[0007] Optionally, after pushing the prompt information of the target solution in each of the writing processes to the functional modules corresponding to each of the writing processes, it also includes: receiving a solution submission request from the target user, and obtaining the target solution submitted by the target user based on the solution submission request; calling the historical version of the target solution, calculating the similarity between the target solution and the historical version, and obtaining the solution pass rate of the target solution.

[0008] Optionally, after obtaining the solution pass rate of the target solution, it also includes: when the solution pass rate of the target solution is less than a preset first pass rate threshold, obtaining error information of the target solution, and returning the target solution and the error information to the terminal where the target user is located; when the solution pass rate of the target solution is greater than or equal to the first pass rate threshold and less than a preset second pass rate threshold, sending the target solution to the terminal used by the reviewer; when the solution pass rate of the target solution is greater than or equal to the second pass rate threshold, determining that the target solution has passed the review.

[0009] Optionally, after pushing the prompt information of the target scheme in each of the writing processes to the functional modules corresponding to each of the writing processes, it also includes: in the process of the target user writing the target scheme, monitoring the duration of the target user's stay in each writing operation; when the duration of the target user's stay in the target operation is greater than a preset duration threshold, determining that the target user has a writing abnormality, calling the historical scheme corresponding to the target scheme, and extracting the historical scheme content of the target operation from the historical scheme; generating prompt information of the target operation based on the historical scheme content of the target operation, and pushing the prompt information of the target operation to the functional module corresponding to the target operation.

[0010] Optionally, the auxiliary writing model is pre-constructed, and the steps of constructing the auxiliary writing model include: collecting historical plan data, historical review records and historical modification records corresponding to historical plans within a historical time period and historical user data for writing the historical plans to obtain historical data; performing feature extraction on the historical data to obtain user behavior features and plan content features; constructing sample data based on the historical data, the user behavior feature data and the plan content features, and dividing the sample data to obtain a training set and a test set; constructing a deep learning model based on a convolutional neural network, and training the deep learning model based on the training set to obtain an initial auxiliary writing model; testing the initial auxiliary writing model based on the test set to obtain a test result, and obtaining the auxiliary writing model when the test result indicates that the initial auxiliary writing model passes the test.

[0011] Optionally, the step of extracting features from the historical data includes: extracting the operation sequence and operation path of the historical user in the process of writing the historical plan based on the historical plan data and the historical user data; obtaining the operation duration of each writing process and the operation time interval between each writing process in the process of writing the historical plan based on the historical plan data and the historical user data; extracting the user behavior features based on the operation sequence, the operation path, the operation duration and the operation time interval.

[0012] Optionally, the step of extracting features from the historical data also includes: counting the frequency of use of each writing function by historical users when writing the historical plans based on the historical plan data and the historical user data; obtaining the latest timestamp of the historical users using each writing function; and extracting the user behavior features based on the frequency of use of each writing function by historical users and the latest timestamp of the use of each writing function.

[0013] Optionally, the step of extracting features from the historical data also includes: extracting text data of the historical scheme based on the historical scheme data, wherein the text data includes at least: keywords in the historical scheme, phrases in the historical scheme and the sentence structure of the historical scheme; extracting numerical data of the historical scheme based on the historical scheme data; extracting option data of the historical scheme based on the historical scheme data; extracting scheme content features of the historical scheme based on the text data of the historical scheme, the numerical data and the option data.

[0014] Optionally, the step of extracting features from the historical data also includes: counting the error causes, error types and error frequencies of the historical plans written by users based on the historical audit records; counting the historical plans written by users based on the historical modification records, and obtaining the modification methods and modification effects of the historical plans that failed the audit; and extracting the plan content features of the historical plans based on the error causes, error types and error frequencies of the historical plans as well as the modification methods and modification effects.

[0015] According to another aspect of an embodiment of the present invention, an auxiliary writing device for a scheme is also provided, comprising: a receiving unit, for receiving a scheme creation request from a target user in an interactive system, and obtaining user information of the target user, scheme information of a target scheme created by the target user, and historical scheme writing records of the target user based on the scheme creation request; a processing unit, for preprocessing and vector encoding the user information, the scheme information, and the historical scheme writing records to obtain a coding vector; an output unit, for inputting the coding vector into an auxiliary writing model, and outputting prompt information of each writing process of the target scheme, wherein the auxiliary writing model is a pre-built deep learning model for generating prompt information for user scheme writing; a pushing unit, for pushing the prompt information of the target scheme in each of the writing processes to the functional modules corresponding to each of the writing processes.

[0016] Optionally, the auxiliary writing device of the solution also includes: a first receiving module, used to receive a solution submission request from a target user, and obtain a target solution submitted by the target user based on the solution submission request; a first calculation module, used to call a historical version of the target solution, calculate the similarity between the target solution and the historical version, and obtain the solution pass rate of the target solution.

[0017] Optionally, the auxiliary writing device of the scheme also includes: a first acquisition module, used to obtain error information of the target scheme when the scheme pass rate of the target scheme is less than a preset first pass rate threshold, and return the target scheme and the error information to the terminal where the target user is located; a first sending module, used to send the target scheme to the terminal used by the reviewer when the scheme pass rate of the target scheme is greater than or equal to the first pass rate threshold and less than a preset second pass rate threshold; a first determination module, used to determine that the target scheme has passed the review when the scheme pass rate of the target scheme is greater than or equal to the second pass rate threshold.

[0018] Optionally, the auxiliary writing device of the scheme also includes: a first monitoring module, used to monitor the target user's stay time in each writing operation during the process of the target user writing the target scheme; a first judgment module, used to determine that the target user has a writing abnormality when the target user's stay time in the target operation is greater than a preset time threshold, call the historical scheme corresponding to the target scheme, and extract the historical scheme content of the target operation from the historical scheme; a first generation module, used to generate prompt information of the target operation based on the historical scheme content of the target operation, and push the prompt information of the target operation to the functional module corresponding to the target operation.

[0019] Optionally, the auxiliary writing device of the scheme also includes: a first acquisition module, which is used to collect historical scheme data, historical review records and historical modification records corresponding to historical schemes within a historical time period, and historical user data for writing the historical schemes to obtain historical data; a first extraction module, which is used to extract features from the historical data to obtain user behavior features and scheme content features; a first division module, which is used to construct sample data based on the historical data, the user behavior feature data and the scheme content features, and divide the sample data to obtain a training set and a test set; a first construction module, which is used to construct a deep learning model based on a convolutional neural network, and train the deep learning model based on the training set to obtain an initial auxiliary writing model; a first testing module, which is used to test the initial auxiliary writing model based on the test set to obtain a test result, and when the test result indicates that the initial auxiliary writing model passes the test, the auxiliary writing model is obtained.

[0020] Optionally, the first extraction module includes: a first extraction sub-module, used to extract the operation sequence and operation path of the historical user in the process of writing the historical plan based on the historical plan data and the historical user data; a first acquisition sub-module, used to acquire the operation duration of each writing process and the operation time interval between each writing process in the process of writing the historical plan by the historical user based on the historical plan data and the historical user data; a second extraction sub-module, used to extract the user behavior characteristics based on the operation sequence, the operation path, the operation duration and the operation time interval.

[0021] Optionally, the first extraction module also includes: a first statistical sub-module, used to count the frequency of use of each writing function by historical users when writing the historical plans based on the historical plan data and the historical user data; a second acquisition sub-module, used to obtain the latest timestamp of the historical user's use of each writing function; and a second extraction sub-module, used to extract the user behavior characteristics based on the frequency of use of each writing function by historical users and the latest timestamp of the use of each writing function.

[0022] Optionally, the first extraction module also includes: a third extraction sub-module, used to extract text data of the historical scheme based on the historical scheme data, wherein the text data at least includes: keywords in the historical scheme, phrases in the historical scheme and the sentence structure of the historical scheme; a fourth extraction sub-module, used to extract numerical data of the historical scheme based on the historical scheme data; a fifth extraction sub-module, used to extract option data of the historical scheme based on the historical scheme data; and a sixth extraction sub-module, used to extract scheme content features of the historical scheme based on the text data of the historical scheme, the numerical data and the option data.

[0023] Optionally, the first extraction module also includes: a second statistical sub-module, which is used to count the error causes, error types and error frequencies of the historical plans written by users based on the historical audit records; a third statistical sub-module, which is used to count the historical plans written by users based on the historical modification records, and obtain the modification methods and modification effects of the historical plans that fail the audit; and a seventh extraction sub-module, which is used to extract the plan content features of the historical plans based on the error causes, error types and error frequencies of the historical plans as well as the modification methods and the modification effects.

[0024] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute an auxiliary writing method of any of the above-mentioned schemes.

[0025] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the auxiliary writing method of any one of the above-mentioned schemes.

[0026] According to another aspect of an embodiment of the present invention, a computer program product is further provided, the computer program product comprising a computer program, wherein the computer program implements the auxiliary writing method of any one of the above solutions when executed by a processor.

[0027] In the present application, the following steps are performed: first, a solution creation request from a target user in an interactive system is received; based on the solution creation request, user information of the target user, solution information of the target solution created by the target user, and historical solution writing records of the target user are obtained; the user information, solution information, and historical solution writing records are pre-processed and vector encoded to obtain a coding vector; the coding vector is then input into an auxiliary writing model; prompt information of each writing process of the target solution is output; the auxiliary writing model is a pre-built deep learning model for generating prompt information for user solution writing; and finally, the prompt information of the target solution in each writing process is pushed to the functional modules corresponding to each writing process.

[0028] In the present application, an auxiliary writing model is used to extract and analyze features of the target user's real-time user information, the plan information of the current writing plan, and the target user's historical plan writing records, identify the target user's plan writing habits, and automatically output real-time and personalized prompt information through the auxiliary writing model. The prompt information is pushed to the corresponding functional modules of each writing process of the target plan on the operating platform, and the user is assisted in writing the plan in real time, thereby achieving the purpose of real-time generation of prompt information, and obtaining the technical effect of improving user writing efficiency and accuracy, thereby solving the technical problem in the related technology that the error rate of users when writing plans on various operating platforms is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0030] Figure 1 A hardware structure block diagram of a computer terminal for implementing the auxiliary writing method of the scheme is shown;

[0031] Figure 2 is a flow chart of an auxiliary writing method according to an optional solution of an embodiment of the present invention;

[0032] Figure 3 is a schematic diagram of an optional auxiliary writing model construction process according to an embodiment of the present invention;

[0033] Figure 4 is a schematic diagram of an auxiliary writing device according to an optional solution of an embodiment of the present invention;

[0034] Figure 5 It is a hardware structure block diagram of an electronic device (or mobile device) for assisting writing method according to an execution scheme of an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0036] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention 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 that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0037] It should be noted that the auxiliary writing method and device of the scheme in the present application can be used in the field of artificial intelligence when prompts and automatic review of scheme writing are provided based on artificial intelligence, and can also be used in any field other than the field of artificial intelligence when prompts and automatic review of scheme writing are provided based on artificial intelligence. The application field of the auxiliary writing method and device of the scheme in the present application is not limited.

[0038] It should be noted that the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with the relevant laws, regulations and standards of the relevant regions, take necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or organizations. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.

[0039] It should be noted that in this application, when collecting and analyzing customer information, corresponding operation entrances are provided for users to choose to agree or reject the automated decision-making results; if the user chooses to reject, the expert decision-making process will be entered.

[0040] The following embodiments of the present invention can be applied to various auxiliary writing systems / applications / devices for plans. The present invention provides real-time prompt information for plan writing to help enterprise employees or customers correctly fill in / write plans when using relevant operating platforms / systems, improve work efficiency and plan quality, and automatically review plans, reduce the workload of reviewers, and improve the accuracy of plan writing and the efficiency of plan writing and operation.

[0041] The present invention is described in detail below in conjunction with various embodiments.

[0042] Embodiment 1

[0043] According to an embodiment of the present invention, an embodiment of an auxiliary writing method of a scheme is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0044] The method embodiment provided in the first embodiment of the present invention may be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal (or mobile device) for implementing the auxiliary writing method of the solution is shown. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more (102a, 102b, ..., 102n are used to illustrate) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0045] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0046] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the auxiliary writing method of the scheme in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the auxiliary writing method of the above-mentioned scheme is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via 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 combinations thereof.

[0047] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0048] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0049] Under the above operating environment, the present invention provides Figure 2 The auxiliary writing method of the scheme shown is implemented by the auxiliary writing system of the scheme.

[0050] The embodiment of the present invention is described in detail below in conjunction with the following steps.

[0051] Figure 2is a flowchart of an auxiliary writing method according to an optional solution of an embodiment of the present invention, such as Figure 2 As shown, the method comprises the following steps:

[0052] Step S201 : receiving a solution creation request from a target user in an interactive system, and acquiring user information of the target user, solution information of a target solution created by the target user, and historical solution writing records of the target user based on the solution creation request.

[0053] It should be noted that the embodiments of the present invention can be applied to an operation platform or system for assisting the writing of a plan, and is used to solve the problems encountered by users when filling out or writing plans in complex operation platforms / systems, as well as the efficiency and accuracy problems in the subsequent review process. The application scenarios are wide, for example: the business operation platform, loan application system, insurance claim platform, etc. in the financial industry can use this intelligent prompt to help business personnel or users quickly fill out plans, and provide an automatic review mechanism to improve the accuracy of business applications and the efficiency of review. Another example is the healthcare industry, electronic medical record systems, medical insurance reimbursement application platforms, etc. These systems often require doctors or patients to fill out a large amount of information, and the accuracy of the information is crucial to the subsequent diagnosis and treatment or reimbursement process. The technical solution of the embodiment of the present invention can provide real-time prompt information for doctors to fill out medical record plans or reimbursement forms. For example, in scenarios such as tax declaration, public project application, license processing, etc., users on the e-government platform can assist users in writing declaration plans and automatically review them, thereby improving the accuracy of declarations.

[0054] It should be noted that the embodiment of the present invention does not limit the type and content of the plan. The target plan written by the user can be a loan application, project report, financial statement, project application plan, etc., which involves multiple scenarios and multiple fields. Documents or images. When the target user logs in to the interactive system (i.e., the operating system) and initiates a plan creation request, the system first identifies and records this operation. Plan creation is often the starting point of the business process. In order to respond to this request, the system needs to have functions such as user identity authentication, authority management, and operation log recording to ensure the legality and security of the operation.

[0055] After receiving the target user's request to create a plan, the plan-assisted writing system extracts information related to the target user, including but not limited to the user's basic attributes (such as work number, name, department, position), obtains user information, and obtains the target user's historical plan writing records, including historical plans created in the interactive system, and obtains the target user's activity in the interactive system, the number of plans created before, and the review pass rate of historical plans, etc. It can also obtain the target user's personalized settings for the operating system (such as language preference, interface layout preference). The acquisition of this information helps the system evaluate the user's proficiency and possible needs, and provides a basis for subsequent intelligent prompts and operating instructions.

[0056] When a user creates a new target plan, he or she can determine the plan information of the target plan based on the name, identification information, basic description information, etc. of the target plan, including but not limited to the plan type (such as financial report, loan application, project proposal), the content of each field filled in, the selected options, the uploaded attachments, etc.

[0057] Step S202, pre-processing and vector encoding processing are performed on the user information, solution information and historical solution writing records to obtain a coding vector.

[0058] It should be noted that after obtaining the user information of the target user, the solution information of the target solution and the historical solution writing records, preprocessing is performed to obtain the preprocessed data, and then vector encoding is performed on the processed data to obtain a data format that can be recognized by the model. Specifically, user information preprocessing may include: removing redundant, repeated, irrelevant or erroneous data from user information, for example, checking whether fields such as employee number and name have format errors or spelling errors; adopting reasonable strategies to fill in missing user attributes, such as years of service and department, such as using averages, medians or predicted values ​​based on other attributes; unifying data formats, such as converting all dates to a unified format, converting text fields to lowercase or removing special characters to ensure data consistency and comparability; when processing user information, sensitive information can also be desensitized, such as using hashing or random encoding instead of real names and employee numbers; when encoding user information, attributes that can be directly quantified, such as age and years of service, are directly converted into their numerical forms; for categorical attributes, such as gender, department, and position, one-hot encoding is used to convert them into binary vectors, where each dimension represents a category, 1 indicates that the category appears, and 0 indicates that it does not appear; for attributes with rich semantics, such as user descriptions, pre-trained word embedding models are used to convert them into fixed-length vectors to capture their intrinsic meanings.

[0059] When preprocessing the plan information, it can include: removing noise in the plan text, such as extra spaces, punctuation marks and numbers, to ensure the consistency of the text; extracting structured data from the plan, such as amounts, dates, and answers to multiple-choice questions, for specific numerical analysis; using natural language processing technology to process the unstructured text in the plan, extracting keywords, phrases and topics, and preparing for the encoding of text features, and then using different encoding methods for different preprocessed data, such as converting structured numerical data into numerical vectors, or converting text data into feature vectors, etc.

[0060] It should be noted that when preprocessing the historical plan writing records, it includes analyzing and integrating the historical plan writing records, data cleaning, missing value processing, feature extraction, etc., and then the preprocessed historical plan writing records are vector encoded to obtain a data structure that can be recognized by the model.

[0061] Step S203: input the coding vector into the auxiliary writing model, and output prompt information of each writing process of the target solution.

[0062] It should be noted that the embodiment of the present invention is based on a pre-built auxiliary writing model to extract and analyze the coding vector features to identify sequence features and capture long-term dependencies. After receiving these coding vectors, the auxiliary writing model will perform complex nonlinear transformations and feature combinations through a multi-layer neural network to predict the problems that may occur when the target user writes the target solution, and output prompt information based on the possible problems, thereby obtaining prompt information for each writing process of the target solution. Specifically, the auxiliary writing model extracts multiple vector features and fuses different types of vector features together to form a comprehensive user and solution status representation, and based on the current user and solution status, predicts possible errors or possible missing information of the user, and identifies the processes and fields that need prompts; then, the pre-trained text generation module is used to generate specific prompt information according to the predicted error types and user needs, such as filling instructions, sample data, common error warnings, etc.

[0063] Figure 3 is a schematic diagram of an optional auxiliary writing model construction process according to an embodiment of the present invention, such as Figure 3 As shown, the steps to build an assisted writing model include:

[0064] Step S301, collecting historical scheme data, historical review records, historical modification records and historical user data for compiling historical schemes corresponding to historical schemes within a historical time period, to obtain historical data;

[0065] It should be noted that the construction of the model cannot be separated from the support of big data. The historical plan data, historical review records and historical modification records corresponding to the historical plans, as well as the historical user data for writing historical plans are the data basis of the auxiliary writing model in the embodiment of the present invention. Therefore, when constructing the auxiliary writing model, it is first necessary to collect the historical plan data corresponding to the historical plans within the historical time period. Based on the historical plan data, the specific content of various types of plans can be obtained, and the review records of the historical plans can be obtained at the same time to obtain the historical review records. The historical review records can be used to learn about the errors that users are prone to when writing historical plans and the historical plans with low review rates, etc. The historical modification records of the historical plans can also be obtained to obtain the modification methods of the erroneous plans. Finally, it is necessary to obtain the historical user data for writing historical plans, match the historical plans with specific users, and learn about the writing habits of each user when writing plans, so as to analyze the content that users are prone to make mistakes in for individual users.

[0066] Step S302, extracting features from historical data to obtain user behavior features and solution content features;

[0067] It should be noted that feature extraction aims to mine features from users' historical behaviors and plan contents that are helpful in predicting users' filling behaviors and plan quality, thereby providing users with more personalized guidance and assistance. Specific feature extraction includes user behavior feature extraction and plan content feature extraction. User behavior features can reflect users' operating habits and proficiency on the platform. Plan content features involve the specific details and structure of the plan, and can provide information on plan quality, content completeness and compliance.

[0068] Optionally, the step of extracting features from historical data includes: extracting the operation sequence and operation path of historical users in the process of writing historical plans based on historical plan data and historical user data; obtaining the operation duration of each writing process and the operation time interval between each writing process in the process of writing historical plans based on historical plan data and historical user data; extracting user behavior features based on the operation sequence, operation path, operation duration and operation time interval.

[0069] Specifically, when extracting user behavior features, the operation sequence of the user when filling out the plan, that is, the operation order and operation path, including clicks, inputs, selections and other actions and their order, is recorded and analyzed. Through sequence analysis, the model can learn the fluency and habits of user operations and identify which links or operations the user may encounter difficulties in; record the time the user spends on each operation or writing step, as well as the target user's operation time interval between each writing process, which helps to judge the user's understanding or proficiency of a specific part. Steps that stay for a long time or are frequently operated may mean that the user is unfamiliar or prone to making mistakes. The model can provide additional guidance on this basis, thereby extracting user behavior features based on the user's operation sequence, operation path, operation duration and operation time interval.

[0070] Optionally, the step of extracting features from historical data also includes: counting the frequency of use of each writing function by historical users when writing historical plans based on historical plan data and historical user data; obtaining the latest timestamp of the historical users using each writing function; and extracting user behavior features based on the frequency of use of each writing function by historical users and the latest timestamp of the use of each writing function.

[0071] It should be noted that the frequency of use of each writing function by historical users when writing historical plans is counted based on historical plan data and historical user data, and the writing functions that users may be unfamiliar with or often make mistakes are identified. The latest timestamp of each writing function used by historical users is obtained, such as the time when the user last used the function. For example, if the user has used the drawing function many times, but the use records are all from 1 year ago and have not been used recently, it is judged that the user's proficiency has decreased. Then, based on the frequency of use of each writing function by historical users and the latest timestamp of using each writing function, the proficiency characteristics of the user in writing plans are obtained and used as one of the user behavior characteristics.

[0072] Optionally, the step of extracting features from historical data also includes: extracting text data of historical plans based on historical plan data, wherein the text data includes at least: keywords in the historical plans, phrases in the historical plans and sentence structures of the historical plans; extracting numerical data of the historical plans based on the historical plan data; extracting option data of the historical plans based on the historical plan data; extracting plan content features of the historical plans based on the text data, numerical data and option data of the historical plans.

[0073] It should be noted that the extraction of solution content features specifically includes: text content analysis, which can use natural language processing technology to analyze the text content in the solution and extract features such as keywords, phrases, and sentence structures; digital features, extracting numerical data in the solution, such as amount, quantity, proportion, etc., and performing statistical analysis; multiple-choice analysis, if there are multiple-choice questions in the solution, analyzing the user's selection patterns and common errors; logical consistency, checking whether the various parts of the solution are logically consistent, such as whether the previous and subsequent data are contradictory, etc., so as to obtain the solution content features of the historical solution.

[0074] Optionally, the step of extracting features from historical data also includes: counting the error causes, error types, and error frequencies of historical plans written by users based on historical audit records; counting the historical plans written by users based on historical modification records, and obtaining the modification methods and modification effects of historical plans that failed the audit; and extracting plan content features of historical plans based on the error causes, error types, and error frequencies as well as the modification methods and modification effects of historical plans.

[0075] It should be noted that the extraction of solution content features also includes: calculating the error causes, error types and error frequencies of historical solutions written by users based on historical records, and obtaining the modification methods and modification effects of historical solutions. According to the error causes, modification methods and modification effects, the user can be provided with writing prompts from a positive perspective, and high-frequency errors can be reminded to prevent users from making mistakes when writing solutions in real time. Finally, the solution content features of the historical solution are obtained based on the error causes, error types and error frequencies of historical solutions as well as the modification methods and modification effects.

[0076] Step S303, constructing sample data based on historical data, user behavior feature data and solution content features, and dividing the sample data to obtain a training set and a test set;

[0077] Step S304, building a deep learning model based on a convolutional neural network, and training the deep learning model based on a training set to obtain an initial auxiliary writing model;

[0078] Step S305, testing the initial auxiliary writing model based on the test set to obtain a test result, and obtaining an auxiliary writing model when the test result indicates that the initial auxiliary writing model passes the test.

[0079] It should be noted that sample data is constructed based on historical data, user behavior feature data, and solution content features, and used as sample data for model training and testing. A deep learning model is constructed based on a convolutional neural network, and the deep learning model is iteratively trained and tested based on the sample data to obtain the final assisted writing model. In a convolutional neural network, the same convolution kernel is applied to different parts of the input data, which means that the network can learn only one set of parameters and then reuse them on the entire data. This parameter sharing greatly reduces the number of parameters in the model and reduces the risk of overfitting. The trained assisted writing model will predict possible errors and areas where users need help based on user behavior data and real-time solution content, and give corresponding prompts and operating instructions to achieve the purpose of real-time and personalized prompts.

[0080] The deep learning model can be represented as:

[0081] Y=f(W,X),

[0082] Among them, X represents the input data, including user behavior characteristics, solution content characteristics, audit record characteristics, and quality control group inspection problem summary characteristics; W represents the weight parameters of the model; Y represents the output results, namely intelligent prompts and operation instructions; f represents the forward propagation process of the deep learning model.

[0083] Step S204: Push prompt information of the target solution in each writing process to the functional modules corresponding to each writing process.

[0084] It should be noted that after outputting the prompt information of the target solution in each writing process, the prompt information needs to be displayed on the interface of the user terminal and pushed to the corresponding work modules of each writing process in the interactive system. For example, the attachment upload requirement is pushed to the attachment upload module to prompt the user that attachments need to be uploaded here and the conditions that need to be met for uploading attachments.

[0085] In addition, after users write a plan, auditors are required to review the written plan before they can perform subsequent related operations. Auditors also face huge work pressure and efficiency bottlenecks when handling a large number of plan review tasks. Audit work usually requires a high degree of professional knowledge and meticulous attention. Manual review is not only time-consuming, but also prone to fatigue, resulting in fluctuations in audit quality and efficiency. Especially when encountering complex business logic or situations that require cross-departmental coordination, the limitations of manual review are more obvious.

[0086] Optionally, after pushing the prompt information of the target solution in each writing process to the functional module corresponding to each writing process, it also includes: receiving the solution submission request of the target user, and obtaining the target solution submitted by the target user based on the solution submission request; calling the historical version of the target solution, calculating the similarity between the target solution and the historical version, and obtaining the solution pass rate of the target solution.

[0087] It should be noted that the embodiment of the present invention also configures an automatic review mechanism for the solutions written by users. The automatic review mechanism can identify errors in the target solution based on the historical version or standard version of the target solution and calculate the solution pass rate of the target solution.

[0088] Specifically, after the user submits the target plan, the system receives the plan submission request from the target user, automatically enters the plan review process, calls the historical version of the target plan according to the plan information of the target plan, calculates the similarity value between the currently written target plan and the historical version, and obtains the plan pass rate of the target plan through the similarity value.

[0089] Optionally, after obtaining the solution pass rate of the target solution, it also includes: when the solution pass rate of the target solution is less than a preset first pass rate threshold, obtaining error information of the target solution, and returning the target solution and the error information to the terminal where the target user is located; when the solution pass rate of the target solution is greater than or equal to the first pass rate threshold and less than a preset second pass rate threshold, sending the target solution to the terminal used by the reviewer; when the solution pass rate of the target solution is greater than or equal to the second pass rate threshold, determining that the target solution has passed the review.

[0090] It should be noted that the calculation results of the solution pass rate will be used in the subsequent decision-making process. When the pass rate is high, the system can automatically release it to reduce the burden of manual review. When the pass rate is low, the system will mark the solution as high risk and require the user to provide additional explanations or modifications, or it may directly enter the manual review stage. When the pass rate is in the middle range, the system may prompt the user to check certain parts and may combine manual review for comprehensive evaluation.

[0091] Optionally, after pushing the prompt information of the target solution in each writing process to the functional modules corresponding to each writing process, it also includes: in the process of the target user writing the target solution, monitoring the target user's stay time in each writing operation; when the target user's stay time in the target operation is greater than a preset time threshold, determining that the target user has a writing abnormality, calling the historical solution corresponding to the target solution, and extracting the historical solution content of the target operation from the historical solution; generating prompt information of the target operation based on the historical solution content of the target operation, and pushing the prompt information of the target operation to the functional module corresponding to the target operation.

[0092] It should be noted that the auxiliary writing system of the embodiment of the present invention can also monitor the real-time interactive behavior of the target user when writing the target solution in real time. In the process of the user writing the target solution, the user's stay time on each writing operation is monitored in real time. When the user's stay time on a certain operation (called the target operation) exceeds the preset time threshold, the system will automatically determine that there is a writing abnormality. The preset time threshold is obtained based on the statistical analysis of historical data. It takes into account factors such as the average user's stay time, operation complexity, and user proficiency, so that possible writing problems or errors can be effectively identified. Once it is determined that the target user has a writing abnormality, the system will call all historical solution data related to the target solution type, and pay special attention to the part related to the target operation. From these historical solutions, the historical solution content of the target operation is extracted, including the text filled in by the user, the selected options, the uploaded attachments, and the modification records and review feedback related to the operation. Personalized prompt information is generated using a deep learning model. The model can understand the semantics and context of the historical solution data, thereby generating prompts that are relevant to the current operation, specific and easy to understand. The prompt information may include filling examples, common error warnings, operating steps instructions, and related rules and guidance. The generated prompt information will be pushed to the functional module corresponding to the target operation in real time and displayed directly on the user interface, solving the problems encountered by users in the solution writing process in real time and improving user writing efficiency.

[0093] Through the above steps, the target user's solution creation request in the interactive system is first received, and based on the solution creation request, the target user's user information, the solution information of the target solution created by the target user, and the target user's historical solution writing record are obtained, and the user information, solution information and historical solution writing record are preprocessed and vector encoded to obtain the encoding vector, and then the encoding vector is input into the auxiliary writing model to output the prompt information of each writing process of the target solution, wherein the auxiliary writing model is a pre-built deep learning model for generating prompt information for the user's solution writing, and finally the prompt information of the target solution in each writing process is pushed to the functional module corresponding to each writing process.

[0094] In this embodiment, the auxiliary writing model is used to extract and analyze the real-time user information of the target user, the plan information of the current writing plan, and the historical plan writing records of the target user, identify the plan writing habits of the target user, and automatically output real-time and personalized prompt information through the auxiliary writing model. The prompt information is pushed to the corresponding functional modules of each writing process of the target plan on the operating platform, and the user is assisted in writing the plan in real time, thereby achieving the purpose of real-time generation of prompt information, and obtaining the technical effect of improving the user's writing efficiency and accuracy, thereby solving the technical problem in the related technology that the error rate of users when writing plans on various operating platforms is high.

[0095] The following is a detailed description in conjunction with another embodiment.

[0096] Embodiment 2

[0097] The auxiliary writing device of a solution provided in this embodiment includes multiple implementation units, each implementation unit corresponds to each implementation step in the above-mentioned embodiment 1. Its specific implementation method and beneficial effects can refer to the above-mentioned method embodiment, which will not be repeated here.

[0098] Figure 4 is a schematic diagram of an auxiliary writing device according to an optional solution of an embodiment of the present invention, such as Figure 4 As shown, the auxiliary writing device of the scheme includes: a receiving unit 41, a processing unit 42, an output unit 43, and a push unit 44, wherein:

[0099] The receiving unit 41 is used to receive a solution creation request from a target user in the interactive system, and obtain user information of the target user, solution information of a target solution created by the target user, and a historical solution writing record of the target user based on the solution creation request;

[0100] The processing unit 42 is used to perform preprocessing and vector encoding processing on the user information, the scheme information and the historical scheme writing record to obtain a coding vector;

[0101] The output unit 43 is used to input the encoding vector into the auxiliary writing model, and output the prompt information of each writing process of the target solution, wherein the auxiliary writing model is a pre-built deep learning model for generating prompt information for the user's solution writing;

[0102] The push unit 44 is used to push the prompt information of the target solution in each writing process to the functional modules corresponding to each writing process.

[0103] The auxiliary writing device of the above-mentioned scheme receives a scheme creation request from a target user in an interactive system through a receiving unit 41, and obtains user information of the target user, scheme information of a target scheme created by the target user, and historical scheme writing records of the target user based on the scheme creation request; pre-processes and vector encodes the user information, scheme information, and historical scheme writing records through a processing unit 42 to obtain an encoding vector; inputs the encoding vector into an auxiliary writing model through an output unit 43, and outputs prompt information of each writing process of the target scheme, wherein the auxiliary writing model is a pre-built deep learning model for generating prompt information for user scheme writing; and pushes the prompt information of the target scheme in each writing process to the functional modules corresponding to each writing process through a pushing unit 44.

[0104] In this embodiment, the auxiliary writing model is used to extract and analyze the real-time user information of the target user, the plan information of the current writing plan, and the historical plan writing records of the target user, identify the plan writing habits of the target user, and automatically output real-time and personalized prompt information through the auxiliary writing model. The prompt information is pushed to the corresponding functional modules of each writing process of the target plan on the operating platform, and the user is assisted in writing the plan in real time, thereby achieving the purpose of real-time generation of prompt information, and obtaining the technical effect of improving the user's writing efficiency and accuracy, thereby solving the technical problem in the related technology that the error rate of users when writing plans on various operating platforms is high.

[0105] Optionally, the auxiliary writing device of the plan also includes: a first receiving module, used to receive a plan submission request from a target user, and obtain a target plan submitted by the target user based on the plan submission request; a first calculation module, used to call a historical version of the target plan, calculate the similarity between the target plan and the historical version, and obtain the plan pass rate of the target plan.

[0106] Optionally, the auxiliary writing device of the plan also includes: a first acquisition module, which is used to obtain error information of the target plan when the plan pass rate of the target plan is less than a preset first pass rate threshold, and return the target plan and the error information to the terminal where the target user is located; a first sending module, which is used to send the target plan to the terminal used by the reviewer when the plan pass rate of the target plan is greater than or equal to the first pass rate threshold and less than a preset second pass rate threshold; and a first determination module, which is used to determine that the target plan has passed the review when the plan pass rate of the target plan is greater than or equal to the second pass rate threshold.

[0107] Optionally, the auxiliary writing device of the plan also includes: a first monitoring module, used to monitor the target user's stay time in each writing operation during the process of the target user writing the target plan; a first judgment module, used to determine that the target user has a writing abnormality when the target user's stay time in the target operation is greater than a preset time threshold, call the historical plan corresponding to the target plan, and extract the historical plan content of the target operation from the historical plan; a first generation module, used to generate prompt information of the target operation based on the historical plan content of the target operation, and push the prompt information of the target operation to the functional module corresponding to the target operation.

[0108] Optionally, the auxiliary writing device of the plan also includes: a first acquisition module, which is used to collect historical plan data, historical review records and historical modification records corresponding to historical plans within a historical time period, and historical user data for writing historical plans to obtain historical data; a first extraction module, which is used to extract features from historical data to obtain user behavior features and plan content features; a first division module, which is used to construct sample data based on historical data, user behavior feature data and plan content features, and divide the sample data to obtain a training set and a test set; a first construction module, which is used to construct a deep learning model based on a convolutional neural network, and train the deep learning model based on the training set to obtain an initial auxiliary writing model; a first testing module, which is used to test the initial auxiliary writing model based on the test set to obtain a test result, and obtain an auxiliary writing model when the test result indicates that the initial auxiliary writing model passes the test.

[0109] Optionally, the first extraction module includes: a first extraction sub-module, used to extract the operation sequence and operation path of historical users in the process of writing historical plans based on historical plan data and historical user data; a first acquisition sub-module, used to acquire the operation duration of each writing process and the operation time interval between each writing process in the process of writing historical plans by historical users based on historical plan data and historical user data; a second extraction sub-module, used to extract user behavior characteristics based on the operation sequence, operation path, operation duration and operation time interval.

[0110] Optionally, the first extraction module also includes: a first statistical sub-module, which is used to count the frequency of use of each writing function by historical users when writing historical plans based on historical plan data and historical user data; a second acquisition sub-module, which is used to obtain the latest timestamp of the historical user's use of each writing function; and a second extraction sub-module, which is used to extract user behavior characteristics based on the frequency of use of each writing function by historical users and the latest timestamp of the use of each writing function.

[0111] Optionally, the first extraction module also includes: a third extraction sub-module, used to extract text data of historical schemes based on historical scheme data, wherein the text data includes at least: keywords in the historical schemes, phrases in the historical schemes and sentence structures of the historical schemes; a fourth extraction sub-module, used to extract numerical data of historical schemes based on historical scheme data; a fifth extraction sub-module, used to extract option data of historical schemes based on historical scheme data; and a sixth extraction sub-module, used to extract scheme content features of historical schemes based on the text data, numerical data and option data of historical schemes.

[0112] Optionally, the first extraction module also includes: a second statistical sub-module, which is used to count the error causes, error types and error frequencies of historical plans written by users based on historical audit records; a third statistical sub-module, which is used to count the historical plans written by users based on historical modification records, and obtain the modification method and modification effect of the historical plans that failed the audit; and a seventh extraction sub-module, which is used to extract the plan content characteristics of the historical plans based on the error causes, error types and error frequencies of the historical plans as well as the modification method and modification effect.

[0113] It should be noted that the receiving unit 41, the processing unit 42, the output unit 43, and the push unit 44 correspond to steps S201 to S204 in the first embodiment, and the examples and application scenarios implemented by the above units and the corresponding steps are the same, but are not limited to the contents disclosed in the first embodiment. It should be noted that the above modules or units can be hardware components or software components stored in a memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n), and the above modules or units can also be part of the device and can be run in the computer terminal 10 provided in the first embodiment.

[0114] The present invention is described below in conjunction with another optional embodiment.

[0115] Embodiment 3

[0116] An embodiment of the present invention may also provide an electronic device, Figure 5 is a structural block diagram of an electronic device for a method for assisting writing a solution according to an embodiment of the present invention, such as Figure 5 As shown, the electronic device may include: one or more ( Figure 5 (only one is shown) processor 502, memory 504, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0117] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via 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 combinations thereof.

[0118] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: receive a solution creation request from the target user in the interactive system, and obtain the user information of the target user, the solution information of the target solution created by the target user, and the historical solution writing record of the target user based on the solution creation request; pre-process and vector encode the user information, solution information and historical solution writing record to obtain a coding vector; input the coding vector to the auxiliary writing model, and output prompt information for each writing process of the target solution, wherein the auxiliary writing model is a pre-built deep learning model for generating prompt information for the user's solution writing; push the prompt information of the target solution in each writing process to the functional module corresponding to each writing process.

[0119] The processor can also call the information and application stored in the memory through the transmission device to perform the following steps: after pushing the prompt information of the target solution in each writing process to the functional module corresponding to each writing process, it also includes: receiving the solution submission request of the target user, and obtaining the target solution submitted by the target user based on the solution submission request; calling the historical version of the target solution, calculating the similarity between the target solution and the historical version, and obtaining the solution pass rate of the target solution.

[0120] The processor can also call the information and application stored in the memory through the transmission device to perform the following steps: after obtaining the solution pass rate of the target solution, it also includes: when the solution pass rate of the target solution is less than a preset first pass rate threshold, obtaining the error information of the target solution, and returning the target solution and the error information to the terminal where the target user is located; when the solution pass rate of the target solution is greater than or equal to the first pass rate threshold and less than a preset second pass rate threshold, sending the target solution to the terminal used by the reviewer; when the solution pass rate of the target solution is greater than or equal to the second pass rate threshold, determining that the target solution has passed the review.

[0121] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: after pushing the prompt information of the target solution in each writing process to the functional modules corresponding to each writing process, it also includes: in the process of the target user writing the target solution, monitoring the target user's stay time in each writing operation; when the target user's stay time in the target operation is greater than a preset time threshold, determining that the target user has a writing abnormality, calling the historical solution corresponding to the target solution, and extracting the historical solution content of the target operation from the historical solution; generating prompt information of the target operation based on the historical solution content of the target operation, and pushing the prompt information of the target operation to the functional module corresponding to the target operation.

[0122] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: the auxiliary writing model is pre-built, and the steps of building the auxiliary writing model include: collecting historical plan data, historical review records and historical modification records corresponding to historical plans within the historical time period, and historical user data for writing historical plans to obtain historical data; extracting features from the historical data to obtain user behavior features and plan content features; constructing sample data based on historical data, user behavior feature data and plan content features, and dividing the sample data to obtain training sets and test sets; constructing a deep learning model based on a convolutional neural network, and training the deep learning model based on the training set to obtain an initial auxiliary writing model; testing the initial auxiliary writing model based on the test set to obtain test results, and obtaining an auxiliary writing model when the test results indicate that the initial auxiliary writing model passes the test.

[0123] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: The step of extracting features from historical data includes: extracting the operation sequence and operation path of historical users in the process of writing historical plans based on historical plan data and historical user data; obtaining the operation duration of each writing process and the operation time interval between each writing process in the process of writing historical plans by historical users based on historical plan data and historical user data; extracting user behavior features based on the operation sequence, operation path, operation duration and operation time interval.

[0124] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: The step of extracting features from historical data also includes: based on historical plan data and historical user data, counting the frequency of use of each writing function by historical users when writing historical plans; obtaining the latest timestamp of the historical user's use of each writing function; extracting user behavior features based on the frequency of use of each writing function by historical users and the latest timestamp of the use of each writing function.

[0125] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: the step of extracting features from historical data also includes: extracting text data of historical schemes based on historical scheme data, wherein the text data at least includes: keywords in the historical schemes, phrases in the historical schemes and sentence structures of the historical schemes; extracting numerical data of historical schemes based on historical scheme data; extracting option data of historical schemes based on historical scheme data; extracting scheme content features of historical schemes based on the text data, numerical data and option data of historical schemes.

[0126] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: The step of extracting features from historical data also includes: based on historical audit records, counting the error causes, error types and error frequencies of historical plans written by users; based on historical modification records, counting the historical plans written by users, and obtaining the modification methods and modification effects of historical plans that failed the review; based on the error causes, error types and error frequencies of historical plans as well as the modification methods and modification effects, extracting the plan content features of historical plans.

[0127] Through the auxiliary writing model, the real-time user information of the target user, the plan information of the current writing plan, and the historical plan writing records of the target user are feature extracted and analyzed, the plan writing habits of the target user are identified, and the auxiliary writing model automatically outputs real-time and personalized prompt information. The prompt information is pushed to the corresponding functional modules of each writing process of the target plan on the operating platform, and the user is assisted in writing the plan in real time, achieving the purpose of real-time generation of prompt information, and obtaining the technical effect of improving user writing efficiency and accuracy, thereby solving the technical problem in related technologies that the error rate of users is high when writing plans on various operating platforms.

[0128] By adopting the embodiment of the present invention, a technical solution for scheme writing based on personalized prompt information is provided. The auxiliary writing model is used to extract and analyze the real-time user information of the target user, the scheme information of the current scheme being written, and the historical scheme writing records of the target user, so as to identify the scheme writing habits of the target user. The auxiliary writing model is used to automatically output real-time and personalized prompt information, and the prompt information is pushed to the functional modules corresponding to each writing process of the target scheme on the operating platform, so as to assist the user in writing the scheme in real time, achieve the purpose of real-time generation of prompt information, and obtain the technical effect of improving the user's writing efficiency and accuracy, thereby solving the technical problem in the related technology that the error rate of users when writing schemes on various operating platforms is high.

[0129] It can be understood by those skilled in the art that Figure 5The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Figure 5 The structure of the electronic device is not limited. Figure 5 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 5 Different configurations are shown.

[0130] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0131] The present invention is described below in conjunction with another optional embodiment.

[0132] Embodiment 4

[0133] The embodiment of the present invention further provides a computer-readable storage medium. Optionally, in the embodiment of the present invention, the computer-readable storage medium can be used to store the program code executed by the auxiliary writing method of the solution provided in the first embodiment.

[0134] Optionally, in an embodiment of the present invention, the above-mentioned storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0135] An embodiment of the present invention provides a computer program product, which, when executed on a data processing device, is suitable for executing the steps of an auxiliary writing method for a solution: receiving a solution creation request from a target user in an interactive system, obtaining user information of the target user, solution information of a target solution created by the target user, and historical solution writing records of the target user based on the solution creation request; preprocessing and vector encoding the user information, solution information, and historical solution writing records to obtain a coding vector; inputting the coding vector into an auxiliary writing model, and outputting prompt information for each writing process of the target solution, wherein the auxiliary writing model is a pre-built deep learning model for generating prompt information for the user's solution writing; and pushing the prompt information of the target solution in each writing process to the functional modules corresponding to each writing process.

[0136] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0137] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0138] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0139] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0140] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0141] 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 invention, in essence, 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, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0142] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for assisting writing a solution, characterized in that: include: Receiving a solution creation request from a target user in an interactive system, and acquiring user information of the target user, solution information of a target solution created by the target user, and a historical solution writing record of the target user based on the solution creation request; Preprocessing and vector encoding the user information, the scheme information and the historical scheme writing record to obtain a coding vector; Input the encoding vector into an auxiliary writing model, and output prompt information of each writing process of the target solution, wherein the auxiliary writing model is a pre-built deep learning model for generating prompt information for user solution writing; The prompt information of the target solution in each of the writing processes is pushed to the functional modules corresponding to each of the writing processes.

2. The writing method according to claim 1, characterized in that: After pushing the prompt information of the target solution in each of the writing processes to the functional modules corresponding to each of the writing processes, the method further includes: Receiving a solution submission request from a target user, and obtaining a target solution submitted by the target user based on the solution submission request; The historical version of the target solution is called, the similarity between the target solution and the historical version is calculated, and the solution pass rate of the target solution is obtained.

3. The writing method according to claim 2, characterized in that: After obtaining the pass rate of the target solution, the method further includes: When the solution pass rate of the target solution is less than a preset first pass rate threshold, obtaining error information of the target solution, and returning the target solution and the error information to the terminal where the target user is located; When the pass rate of the target solution is greater than or equal to the first pass rate threshold and less than a preset second pass rate threshold, sending the target solution to a terminal used by an auditor; When the solution pass rate of the target solution is greater than or equal to the second pass rate threshold, it is determined that the target solution has passed the review.

4. The writing method according to claim 1, characterized in that: After pushing the prompt information of the target solution in each of the writing processes to the functional modules corresponding to each of the writing processes, the method further includes: During the process of the target user writing the target solution, monitoring the duration of the target user's stay in each writing operation; When the target user stays for a target operation for a time period longer than a preset time threshold, it is determined that the target user has a writing abnormality, the historical scheme corresponding to the target scheme is called, and the historical scheme content of the target operation is extracted from the historical scheme; Prompt information of the target operation is generated based on the historical solution content of the target operation, and the prompt information of the target operation is pushed to a functional module corresponding to the target operation.

5. The writing method according to claim 1, characterized in that: The auxiliary writing model is pre-constructed, and the steps of constructing the auxiliary writing model include: Collect historical scheme data, historical review records, and historical modification records corresponding to historical schemes within a historical time period, as well as historical user data for compiling the historical schemes, to obtain historical data; Extracting features from the historical data to obtain user behavior features and solution content features; Constructing sample data based on the historical data, the user behavior feature data and the solution content feature, and dividing the sample data to obtain a training set and a test set; Building a deep learning model based on a convolutional neural network, and training the deep learning model based on the training set to obtain an initial auxiliary writing model; The initial auxiliary writing model is tested based on the test set to obtain a test result, and when the test result indicates that the initial auxiliary writing model passes the test, the auxiliary writing model is obtained.

6. The writing method according to claim 5, characterized in that: The step of extracting features from the historical data includes: Extracting the operation sequence and operation path of the historical user in the process of writing the historical plan based on the historical plan data and the historical user data; Based on the historical scheme data and the historical user data, obtaining the operation duration of each writing process and the operation time interval between each writing process in the process of the historical user writing the historical scheme; The user behavior feature is extracted based on the operation sequence, the operation path, the operation duration and the operation time interval.

7. The writing method according to claim 5, characterized in that: The step of extracting features from the historical data also includes: Based on the historical scheme data and the historical user data, statistics are collected on the frequency of use of various writing functions by historical users when writing the historical schemes; Get the latest timestamp of each writing function used by historical users; The user behavior feature is extracted based on the historical frequency of use of each writing function by the user and the latest timestamp of use of each writing function.

8. The writing method according to claim 5, characterized in that: The step of extracting features from the historical data also includes: Extracting text data of the historical scheme based on the historical scheme data, wherein the text data at least includes: keywords in the historical scheme, phrases in the historical scheme, and sentence structures of the historical scheme; Extracting numerical data of the historical scheme based on the historical scheme data; Extracting option data of the historical solution based on the historical solution data; The solution content features of the historical solution are extracted based on the text data of the historical solution, the numerical data and the option data.

9. The writing method according to claim 5, characterized in that: The step of extracting features from the historical data also includes: Counting the error causes, error types and error frequencies of the historical schemes written by users based on the historical audit records; Based on the historical modification records, the historical plans written by users are counted, and the modification methods and modification effects of the historical plans that have not passed the review are obtained; Based on the error causes, error types and error frequencies of the historical solutions as well as the modification methods and modification effects, solution content features of the historical solutions are extracted.

10. An auxiliary writing device of a solution, characterized in that: include: A receiving unit, configured to receive a solution creation request from a target user in an interactive system, and acquire user information of the target user, solution information of a target solution created by the target user, and a historical solution writing record of the target user based on the solution creation request; A processing unit, configured to perform preprocessing and vector encoding processing on the user information, the scheme information and the historical scheme writing record to obtain a coding vector; An output unit, used to input the encoding vector into an auxiliary writing model, and output prompt information of each writing process of the target solution, wherein the auxiliary writing model is a pre-built deep learning model used to generate prompt information for the user's solution writing; The pushing unit is used to push the prompt information of the target solution in each of the writing processes to the functional modules corresponding to each of the writing processes.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the auxiliary writing method of the solution described in any one of claims 1 to 9.

12. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the auxiliary writing method of the scheme described in any one of claims 1 to 9.

13. A computer program product, characterized in that The computer program product comprises a computer program, wherein the computer program, when executed by a processor, implements the auxiliary writing method of any one of claims 1 to 9.

Images