Automatic petition document circulation method and system and computer readable medium

Abstract

The invention provides a petition document automatic circulation method and system and a computer readable medium, and the method comprises the steps: receiving a petition document text from a user side, and extracting key features in the petition document text; classifying the petition text based on the key features and obtaining the category of the petition text; determining a processing department terminal corresponding to the petition text according to the category of the petition text; and receiving feedback information generated by the processing department side in response to the petition document text, and forwarding the feedback information to the user side. The influence of manual classification is reduced, and the processing efficiency and accuracy of the petition document text are improved. Through automatic data preprocessing, classification and dispatching, manual intervention is reduced, and the processing speed of the petition document is remarkably increased.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method and system for automatically circulating petitions, and a computer-readable medium. Background Art

[0002] Letters and visits refer to letters written by citizens, legal persons or other organizations to the people's governments at all levels and working departments of the people's governments at or above the county level in the form of letters, e-mails, faxes, telephone calls, visits, etc., to reflect situations, make suggestions, opinions or complaints, and are handled by the relevant administrative organs in accordance with the law.

[0003] In the existing related technologies, letters and visits are usually handled manually, including receiving, registering, classifying, distributing, processing and feedback. Most of these steps rely on manual operation, which is inefficient, easily affected by subjective factors, and prone to omissions and delays when handling a large number of letters and visits. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a method, system and computer-readable medium for the automatic circulation of letters and visits, which solves the problem in the prior art that letters and visits rely on manual operations, are inefficient and are easily affected by subjective factors.

[0005] In the first aspect, the present application provides a method for automatically circulating petitions, including:

[0006] Receiving a petition text from a user terminal and extracting key features from the petition text;

[0007] Classifying the petition text based on the key features and obtaining the category of the petition text;

[0008] Determine the processing department end corresponding to the petition text according to the category of the petition text;

[0009] Receive feedback information generated by the processing department in response to the petition text, and forward the feedback information to the user end.

[0010] In one embodiment, extracting key features from the petition text specifically includes:

[0011] Determine a vocabulary list containing all words in the petition text according to the petition text; the vocabulary list includes words and frequencies corresponding to the words;

[0012] Determining a target word in the vocabulary according to the frequency corresponding to the word;

[0013] Convert each of the target words into a target vector of fixed dimension through a preset word embedding model;

[0014] The target vector is aggregated according to a preset aggregation method and key features in the petition text are determined.

[0015] In one embodiment, before extracting the key features in the petition text, the method further includes:

[0016] Performing data cleaning on the petition text;

[0017] Determine the format data to be unified of the petition text after cleaning, and perform data unification processing on the format data to be unified;

[0018] Determine the category data to be converted of the cleaned petition text, and perform encoding conversion on the category data to be converted;

[0019] Determine the unstructured data of the petition text after cleaning, and convert the unstructured data into corresponding structured data.

[0020] In one embodiment, the classifying the petition text based on the key features and obtaining the category of the petition text specifically includes:

[0021] The key features are used as input, corresponding category labels are determined through a preset classification model, and the category of the petition text is determined based on the category labels.

[0022] In one embodiment, the classification model is trained by:

[0023] Obtaining the original text of the petition to be trained, and marking key information of the original text;

[0024] Divide the annotated original text into training set, validation set and test set;

[0025] The training set is trained by a deep learning model to obtain a primary classification model, the parameters of the primary classification model are adjusted by a validation set, and the performance of the classification model is tested by a test set.

[0026] In one embodiment, determining the processing department end corresponding to the petition text according to the category of the petition text specifically includes:

[0027] Based on the category of the petition text, a preset rule matching the petition text is determined, and according to the preset rule, a processing department end corresponding to the petition text is determined.

[0028] In one embodiment, after determining the processing department end corresponding to the petition text according to the category of the petition text, the method further includes:

[0029] Obtaining the processing progress of the petition text at the processing department end;

[0030] When the query interface receives a query task regarding the petition text, the processing progress is fed back to the query interface.

[0031] In one embodiment, after forwarding the feedback information to the user terminal, the method further includes:

[0032] Determine the target decision requirement according to the input decision information, and match the corresponding target decision model based on the target decision requirement;

[0033] Acquire internal attribute information and external attribute information about the petition text, and construct an attribute information data set according to the internal attribute information and the external attribute information;

[0034] The attribute information data set is used as input, and the decision analysis result of the petition text is determined through the target decision model.

[0035] In a second aspect, the present application provides a system for automatically circulating letters and visits, comprising a processor and a memory; wherein the memory stores a computer program, and the computer program is used to be loaded by the processor and execute the method for automatically circulating letters and visits as described in any one of the first aspects.

[0036] In a third aspect, the present application provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores instructions, and the instructions are used by a processor to load and execute the automatic circulation method of letters and visits as described in any one of the first aspects.

[0037] In a method, system and computer-readable medium for automatically circulating letters and visits of the present embodiment, the category of the letter and visit text is determined according to the key features of the letter and visit text, and the letter and visit text is sent to the corresponding processing department end, and the technical personnel of the processing department end processes the corresponding letter and visit text, and then the processed feedback information is sent to the user end that provides the letter and visit text, and the letter and visit text is automatically circulated to the corresponding processing department end after classification by the user end, and then the feedback information of the processing department end processing the letter and visit text is forwarded to the user end, which reduces the influence of manual labor on classification and improves the efficiency and accuracy of the processing of the letter and visit text. Through automated data preprocessing, classification and dispatching, manual intervention is reduced, and the processing speed of letters and visits is significantly accelerated. The automatic circulation method of letters and visits of the present embodiment can complete preliminary processing within a few minutes, ensuring timely response to the demands of the masses. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0039] Figure 1 This is a flow chart of the method for automatically circulating letters and visits in an embodiment of the present application. DETAILED DESCRIPTION

[0040] The specific embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the description of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] In the description of the present invention, unless otherwise clearly specified and limited, the terms "set", "install", "connection" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0042] The directions or positional relationships indicated by terms such as “upper”, “lower”, “left”, “right”, “front”, “back”, “top”, “bottom”, “inside” and “outside” are based on the directions or positional relationships shown in the accompanying drawings, or are the directions or positional relationships in which the inventive product is usually placed when used. They are only for the convenience of description and simplified description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as a limitation on the present invention.

[0043] The terms "first", "second", "third" and the like are merely used to distinguish elements of similar nature, and do not indicate or imply relative importance or a particular order.

[0044] The terms "comprises," "comprising," or any other variations thereof, are intended to cover a non-exclusive inclusion of the elements listed and may also include additional elements not expressly listed.

[0045] like Figure 1 As shown, this embodiment provides a method for automatically circulating petitions, including:

[0046] Step S100: receiving a petition text from a user terminal, and extracting key features from the petition text;

[0047] Step S200: classifying the petition text based on the key features and obtaining the category of the petition text;

[0048] Step S300: determining the processing department end corresponding to the petition text according to the category of the petition text;

[0049] Step S400: receiving feedback information generated by the processing department in response to the petition text, and forwarding the feedback information to the user end.

[0050] In the automatic circulation method of letters and visits of the present embodiment, the category of the letter and visit text is determined according to the key features of the letter and visit text, and the letter and visit text is sent to the corresponding processing department end, and the technical personnel of the processing department end processes the corresponding letter and visit text, and then the processed feedback information is sent to the user end that provides the letter and visit text, and the letter and visit text is automatically circulated to the corresponding processing department end after classification by the user end, and then the feedback information of the processing department end processing the letter and visit text is forwarded to the user end, which reduces the influence of manual labor on classification and improves the efficiency and accuracy of the processing of the letter and visit text. Through automated data preprocessing, classification and dispatching, manual intervention is reduced, and the processing speed of letters and visits is significantly accelerated. The automatic circulation method of letters and visits of the present embodiment can complete preliminary processing within a few minutes, ensuring timely response to the demands of the masses.

[0051] Step S100: receiving a petition text from a user terminal, and extracting key features from the petition text.

[0052] In step S100, the user can submit the petition text through the user interaction interface of the user terminal. The user terminal also records the user's name, contact information, occupation and other basic information, as well as the IP address of the user terminal, so as to facilitate the binding of the petition text with the corresponding user terminal. The key features in the petition text refer to information such as the user, the subject of the appeal, and the timestamp. After logging in to the user terminal, the user can provide the option of submitting the petition and querying the progress through the main interface. After clicking the option to submit the petition, the user fills in the detailed information of the petition in this form and submits it.

[0053] The extracting of key features from the petition text specifically includes:

[0054] Step S101: determining a vocabulary list including all words in the petition text according to the petition text; the vocabulary list includes words and frequencies corresponding to the words;

[0055] Step S102: Determine the target words in the vocabulary according to the frequencies corresponding to the words;

[0056] Step S103: Convert each of the target words into a target vector of a fixed dimension through a preset word embedding model;

[0057] Step S104: Aggregate the target vectors according to a preset aggregation method and determine the key features in the letter of complaint text.

[0058] In step S101, it is necessary to first clean and format the words in the letter of complaint text. First, tokenize the text to convert the continuous text string into a sequence of words or terms. Then, remove the stop words such as "de", "shi", "zai", etc. that do not carry important semantic information. Finally, organize the remaining words to obtain the vocabulary.

[0059] In step S102, methods such as the Bag of Words (BoW) model or TF-IDF can be used to convert the text into vector form. The Bag of Words model constructs a dictionary containing all the words and represents each document as a long vector, where each dimension corresponds to a word in the dictionary and the value represents the frequency of the word in the document. TF-IDF (Term Frequency-Inverse Document Frequency) takes into account the frequency of a word in a document (TF) and the inverse document frequency (IDF) in all documents to evaluate the importance of a word for a document in a document set or a corpus. When the frequency of a certain word in the vocabulary exceeds the set threshold frequency, then the word is the target word.

[0060] In addition, methods such as PCA (Principal Component Analysis) or LDA (Linear Discriminant Analysis) can be used for dimensionality reduction to reduce the complexity of the data and extract the target words.

[0061] In step S103, a pre-trained word embedding model, such as the Word2Vec model or the GloVe model, is used to convert each word into a target vector of a fixed dimension, and these target vectors can capture the semantic relationships between words.

[0062] In step S104, the extracted feature vectors are combined to form a document-level feature vector, which can be aggregated by taking the average of the word vectors or other aggregation methods to obtain the key features in the letter of complaint text.

[0063] Before extracting the key features in the letter of complaint text, it further includes:

[0064] Step S105: Clean the data of the letter of complaint text;

[0065] Step S106: determining the format data to be unified of the petition text after cleaning, and performing data unification processing on the format data to be unified;

[0066] Step S107: determining the category data to be converted of the cleaned petition text, and performing encoding conversion on the category data to be converted;

[0067] Step S108: determining the unstructured data of the petition text after cleaning, and converting the unstructured data into corresponding structured data.

[0068] In step S105, the data cleaning operation includes operations such as removing noise and filling missing values. Through data cleaning, functions such as removing irrelevant characters, correcting incorrect formats, and processing missing values ​​can be achieved. For example, "#1234!" can be cleaned to "1234", or "2024-01 / 22" can be corrected to the standard format "2024-01-22" or the standard format "2024 / 01 / 22", and if some fields such as user addresses are missing, default values ​​can be filled in according to the context, or statistical methods such as averages and medians can be used to fill in.

[0069] In step S106, when the petition text contains multiple different expression formats, such as "01 / 22 / 2024", "2024-01-22", and "Jan22,2024", the formats "01 / 22 / 2024" and "Jan22,2024" are unified into the standard format "2024 / 01 / 22". By unifying the standardized format, it is possible to ensure that the data remains consistent at different processing stages, reduce errors caused by inconsistent formats, improve processing efficiency, and facilitate analysis and comparison.

[0070] In step S107, since digital codes are easy to process, the text data in the petition text is converted into digital codes, which can improve the accuracy and processing efficiency of the petition text. For example, the categories of the petition may be in text form, such as "complaint", "suggestion", "help", and these text categories are converted into digital codes, such as "complaint" is converted into "1", "suggestion" is converted into "2", and "help" is converted into "3".

[0071] In step S108, by converting the unstructured data into corresponding structured data, it is possible to ensure that the data remains consistent at different processing stages, reduce errors caused by inconsistent formats, improve processing efficiency, and facilitate analysis and comparison. For example, by converting the unstructured data into a structured form, the text data of "petition ID", "submission date", "user", "category code", and "text content" are presented in a table.

[0072] Step S200: classifying the petition text based on the key features and obtaining the category of the petition text.

[0073] The step of classifying the petition text based on the key features and obtaining the category of the petition text specifically includes: taking the key features as input, determining the corresponding category label through a preset classification model, and determining the category of the petition text based on the category label.

[0074] The classification model can be obtained by training machine learning models such as support vector machine (SVM), random forest, neural network, and deep learning models such as recurrent neural network (RNN), long short-term memory network (LSTM), gated recurrent unit (GRU) or Transformer.

[0075] The classification model is trained by:

[0076] Step S201: obtaining the original text of the petition document to be trained, and marking the key information of the original text;

[0077] Step S202: dividing the annotated original text into a training set, a validation set and a test set;

[0078] Step S203: Train the training set through a deep learning model to obtain a primary classification model, adjust the parameters of the primary classification model through a validation set, and test the performance of the classification model through a test set.

[0079] In step S201, the petition texts to be trained include historical petition texts, and the historical petition texts are annotated in detail, such as the category of the petition (such as complaint, suggestion, request for help, etc.) and the urgency (such as high, medium, low), etc.

[0080] After labeling, the labeled historical petition texts need to be preprocessed, including noise removal, filling missing values, etc., and then natural language processing technology is used to extract key features, such as users, petition topics, timestamps and other information as key features.

[0081] In step S202 and step S203, the training set is used to train the model; the validation set is used to evaluate the performance of the model, such as accuracy, recall, and F1 score; the test set is used to evaluate the final performance of the model, and the model is adjusted and optimized according to the evaluation results to improve the accuracy and efficiency of classification. For example, 70% of the original text is divided into the training set, 15% of the original text is divided into the validation set, and 15% of the original text is divided into the test set. It can also be adjusted according to the amount of data and task requirements.

[0082] By training on a large amount of historical data, the system can accurately identify the categories of letters and visits, improve the accuracy of classification, and ensure that letters and visits can be handled correctly.

[0083] Step S300: Determine the processing department end corresponding to the petition text according to the category of the petition text.

[0084] Determining the processing department end corresponding to the petition text according to the category of the petition text specifically includes: determining a preset rule matching the petition text based on the category of the petition text, and determining the processing department end corresponding to the petition text according to the preset rule.

[0085] According to the preset rules of the "Rules for Handling Petitions and Appeals in Accordance with the Law" issued by the State Bureau of Petitions, petitions can be classified according to category and urgency. For example, petitions can be classified according to category into complaints (involving dissatisfaction with the official behavior of administrative agencies or their staff), suggestions (improvement suggestions for policies or services) and requests for help (requests for government help to solve personal difficulties or problems); according to urgency, they can be divided into high-level (urgent matters involving major social stability, public safety or personal life and property safety), medium-level (matters that need to be resolved within a certain period of time) and low-level (general consultations or suggestions, matters that can be handled step by step).

[0086] For another example, when dispatching based on case category: if the category of the petition is "complaint", it will be automatically dispatched to the complaint handling department. If the category of the petition is "suggestion", it will be dispatched to the policy research department. When dispatching based on urgency: if the petition is marked as "urgent", it will be dispatched to the emergency response team first. If it is marked as "non-urgent", it will be dispatched to the regular processing department. When dispatching based on geographic location: if the petition involves an event in a specific area, such as "Tianfu New District", it will be dispatched to the office in the corresponding area. When dispatching based on case complexity: if the content of the petition exceeds a certain length or contains multiple issues, it may require professional knowledge and be dispatched to the expert group. When dispatching based on the historical handling of the case: if the user has had a similar case handled by the same department before, in order to maintain continuity, it will be dispatched to the same department again. When dispatching based on resource availability: check the current workload of each department and dispatch the case to the department with a lighter workload to ensure processing efficiency. When dispatching based on handling expertise: If the letter involves specific expertise, such as legal, medical or engineering issues, dispatch it to a department with relevant professional background. When dispatching based on time sensitivity: If the letter involves a policy or activity that is about to expire, dispatch it to a department that can respond quickly. When dispatching based on the source of the case: If the letter comes from a specific channel, such as a government website, social media or written letter, dispatch it to the corresponding processing team based on the nature of the channel. When dispatching based on user information: If the user is elderly or disabled and may need special care, dispatch him to an experienced customer service representative.

[0087] By training the classification model, the category and urgency of the petition can be automatically determined based on the content of the petition, and it can be assigned to the corresponding processing department or individual, thereby improving the efficiency and quality of petition work. It can be integrated with the existing workflow system of government departments to ensure the smooth flow of petitions. Automatically assigning processing tasks based on the category and urgency of petitions and rationally allocating human resources not only reduces labor costs, but also enables staff to devote more time to complex and high-value tasks, improving overall work efficiency.

[0088] Step S400: receiving feedback information generated by the processing department in response to the petition text, and forwarding the feedback information to the user end.

[0089] After the handling department conducts in-depth analysis and processing of the petition, feedback information will be generated. This feedback information may include the results of the investigation of the petition problem, the measures taken, the solution, and the progress of the processing. Once the feedback information is received and sorted, the system will quickly forward it to the user. This can be achieved in a variety of ways, such as sending email notifications, pushing messages on the petition platform, or informing users in the form of text messages. After receiving the feedback information, the user can promptly understand the processing of the problem they reported, evaluate and feedback the processing results. If the feedback information is clear and the solution is feasible, the user may be satisfied with the processing result and end the petition process. However, if the user is not satisfied with the feedback information or believes that the problem has not been properly resolved, they can continue to raise further demands and opinions through the petition channel. At this time, the system needs to convey the user's feedback to the handling department again, prompting the handling department to re-examine the problem and further process it until the user is satisfied with the processing result.

[0090] After determining the processing department end corresponding to the petition text according to the category of the petition text, the method further includes:

[0091] Step S500: obtaining the processing progress of the petition text at the processing department end;

[0092] Step S600: When the query interface receives a query task regarding the petition text, the processing progress is fed back to the query interface.

[0093] In step S500 and step S600, the user can query the processing progress of the petition by entering the relevant identifier (such as the case number) on the main interface of the user end, and obtain the query progress after the query is completed. The processing progress can be divided into multiple stages, such as the acceptance stage, the investigation stage, the processing stage and the feedback stage. In the acceptance stage, the system can display that the petition has been received and is undergoing preliminary review; after entering the investigation stage, the user can understand that the processing department is conducting an in-depth investigation of the problem, which may include collecting evidence, questioning relevant personnel, etc.; the processing stage indicates that the processing department is formulating a solution or taking specific measures to solve the petition problem; finally, in the feedback stage, the user can know that the processing result has been generated and will be fed back to them. The user can quickly and quickly obtain the processing progress of the petition text through the query interface. In addition, the system can also set an automatic notification function to send a notification to the user in a timely manner when the processing progress of the petition changes. This can be achieved through email, SMS or message push on the petition platform. In this way, users can keep abreast of the processing dynamics of the petition without actively querying. The transparency of the processing progress enhances the public's trust in the government's work and improves satisfaction.

[0094] After forwarding the feedback information to the user terminal, the method further includes:

[0095] Step S700: determining a target decision requirement according to the input decision information, and matching a corresponding target decision model based on the target decision requirement;

[0096] Step S800: obtaining internal attribute information and external attribute information about the petition text, and constructing an attribute information data set according to the internal attribute information and the external attribute information;

[0097] Step S900: Using the attribute information data set as input, the decision analysis result of the petition text is determined through the target decision model.

[0098] In step S700, when the input decision information is received, it is necessary to conduct in-depth analysis and understanding of the information. The decision information may include various data, problem descriptions, constraints, goal settings, etc. For example, whether the target decision requirement is to improve the efficiency of petition processing, or to identify bottlenecks in the service, or whether the target decision requirement is to improve the efficiency and satisfaction of petition processing. Corresponding target decision models are set for different target decision requirements.

[0099] In steps S800 and S900, through in-depth mining and analysis of petition data, the decision support module can extract valuable information to help government departments identify service shortcomings and improvement directions. This data-driven decision-making method can improve the scientificity and effectiveness of policies.

[0100] For example, when the target decision requirement is to improve the efficiency of petition processing or to identify bottlenecks in the service, the internal attribute information includes the text content, processing time, processing results, etc. of the petition, and the external attribute information includes public sentiment on social media, relevant policy changes, etc. Then, the internal attribute information and external attribute information can be combined into a consistent attribute information data set through data cleaning (removing irrelevant characters, correcting incorrect formats, processing missing values, etc.), data conversion (converting unstructured data (such as text) into structured data (such as feature vectors)), and data integration.

[0101] When analyzing data, exploratory data analysis (EDA) can be used to perform preliminary analysis of the data through statistical charts and other methods to discover the basic characteristics and patterns of the data, and then select the most useful features for the analysis target through feature engineering, select appropriate algorithms according to the analysis target, such as classification, clustering, regression and other models for training, and use the validation set to evaluate the performance of the model. Then, the analysis results can be intuitively displayed through charts (such as bar charts, line charts, pie charts, heat maps, etc.), or combined with dashboards (such as Tableau or PowerBI) so that users can explore the data.

[0102] By analyzing petition data, the analysis results are converted into easy-to-understand language and strategic recommendations are given. Specific strategies and action plans are proposed based on the analysis results to provide data support for decision makers, helping them make more informed decisions. In addition, we work with relevant departments to promote strategy implementation and effect tracking.

[0103] For another example, when the target decision requirement is to improve the efficiency and satisfaction of petition handling, the internal attribute information includes the type of petition, processing time, processing results, user feedback, etc., and the external attribute information includes public sentiment analysis data on social media.

[0104] When analyzing the data, exploratory data analysis (EDA) can be used to find the types of petitions that take a long time to process, as well as common problems in user feedback. Then, feature engineering can be used to select features related to processing time and user satisfaction. Finally, a logistic regression model is used to predict the processing time of petitions. Then, a bar chart can be made to show the average processing time of petitions of different categories. Or a heat map can be made to show the frequency of keywords in user feedback.

[0105] By analyzing petition data, we can identify the petition categories that take a long time to process and analyze the possible reasons. We can also provide strategic recommendations: we recommend adding resources to handle these petition categories or improving processes. We can also provide decision-making support to help government departments optimize policies and services.

[0106] In summary, in the automatic circulation method of letters and visits of this embodiment, automatic data preprocessing: format standardization and key information extraction of letters and visits are performed through natural language processing technology to improve the efficiency and accuracy of data processing. Machine learning and deep learning algorithms are used to realize automatic classification of letters and visits, reduce human errors, and improve classification accuracy. According to the category and urgency of the letters and visits, the letters and visits are automatically assigned to the corresponding processing departments or individuals to optimize resource allocation. Real-time monitoring of the processing status of letters and visits is realized, progress transparency is enhanced, and public satisfaction is improved. Through in-depth analysis of letters and visits data, scientific decision-making support is provided to help government departments optimize services and policies.

[0107] Based on the same inventive concept as the above-mentioned embodiment, this embodiment provides a method and device for automatically circulating petitions, including:

[0108] A feature extraction module is used to receive the petition text from the user end and extract key features from the petition text;

[0109] A classification module, used to classify the petition text based on the key features and obtain the category of the petition text;

[0110] A department determination module, used to determine the processing department end corresponding to the petition text according to the category of the petition text;

[0111] A feedback module is used to receive feedback information generated by the processing department in response to the petition text, and forward the feedback information to the user end.

[0112] Based on the same inventive concept as the above embodiment, this embodiment also provides a system for automatic circulation of letters and visits, including: a processor and a memory; wherein the memory stores a computer program, and the computer program is used to be loaded by the processor and execute the automatic circulation method of letters and visits as described above.

[0113] Based on the same inventive concept as the above embodiment, this embodiment further provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, and the instructions are used by a processor to load and execute the above-mentioned method for automatically circulating letters and visits.

[0114] In the embodiments of the mobile terminal and computer-readable storage medium provided in the present application, all technical features of the above-mentioned control method embodiments are included, and the expansion and explanation content of the specification are basically the same as those of the above-mentioned method embodiments, which will not be repeated here.

[0115] The embodiment of the present application further provides a computer program product, which includes a computer program code. When the computer program code runs on a computer, the computer executes the methods in the above various possible implementation modes.

[0116] An embodiment of the present application also provides a chip, including a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that a device equipped with the chip executes the methods in various possible implementation modes as described above.

[0117] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0118] In the present application, the same or similar terminology concepts, technical solutions and / or application scenario descriptions are generally described in detail only the first time they appear. When they appear again later, they are generally not repeated for the sake of brevity. When understanding the technical solutions and other contents of the present application, for the same or similar terminology concepts, technical solutions and / or application scenario descriptions that are not described in detail later, reference can be made to the previous related detailed descriptions.

[0119] In the present application, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0120] The various technical features of the technical solution of the present application can be arbitrarily combined. In order to make the description concise, not all possible combinations of the various technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present application.

[0121] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product is stored in a storage medium as above, including a number of instructions for a terminal device to execute the method of each embodiment of the present application. The above is only a preferred embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of this application, or directly or indirectly used in other related technical fields, is similarly included in the patent protection scope of the present application.

[0122] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0123] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the attached claims.

Claims

1. A method for automatically circulating petitions, characterized in that: include: Receiving a petition text from a user terminal and extracting key features from the petition text; Classifying the petition text based on the key features and obtaining the category of the petition text; Determine the processing department end corresponding to the petition text according to the category of the petition text; Receive feedback information generated by the processing department in response to the petition text, and forward the feedback information to the user end.

2. The automatic circulation method of petitions according to claim 1 is characterized in that: The extracting of key features from the petition text specifically includes: Determine a vocabulary list containing all words in the petition text according to the petition text; the vocabulary list includes words and frequencies corresponding to the words; Determining a target word in the vocabulary according to the frequency corresponding to the word; Convert each of the target words into a target vector of fixed dimension through a preset word embedding model; The target vector is aggregated according to a preset aggregation method and key features in the petition text are determined.

3. The automatic circulation method of petitions according to claim 1 is characterized in that: Before extracting the key features in the petition text, it also includes: Performing data cleaning on the petition text; Determine the format data to be unified of the petition text after cleaning, and perform data unification processing on the format data to be unified; Determine the category data to be converted of the cleaned petition text, and perform encoding conversion on the category data to be converted; Determine the unstructured data of the petition text after cleaning, and convert the unstructured data into corresponding structured data.

4. The automatic circulation method of petitions according to claim 1 is characterized in that: The step of classifying the petition text based on the key features and obtaining the category of the petition text specifically includes: The key features are used as input, corresponding category labels are determined through a preset classification model, and the category of the petition text is determined based on the category labels.

5. The automatic circulation method of petitions according to claim 4 is characterized in that: The classification model is trained by: Obtaining the original text of the petition to be trained, and marking key information of the original text; Divide the annotated original text into training set, validation set and test set; The training set is trained by a deep learning model to obtain a primary classification model, the parameters of the primary classification model are adjusted by a validation set, and the performance of the classification model is tested by a test set.

6. The automatic circulation method of petitions according to claim 1 is characterized in that: Determining the processing department end corresponding to the petition text according to the category of the petition text specifically includes: Based on the category of the petition text, a preset rule matching the petition text is determined, and according to the preset rule, a processing department end corresponding to the petition text is determined.

7. The automatic circulation method of petitions according to claim 1 is characterized in that: After determining the processing department end corresponding to the petition text according to the category of the petition text, the method further includes: Obtaining the processing progress of the petition text at the processing department end; When the query interface receives a query task regarding the petition text, the processing progress is fed back to the query interface.

8. The method for automatically circulating petitions according to any one of claims 1 to 7, characterized in that: After forwarding the feedback information to the user terminal, the method further includes: Determine the target decision requirement according to the input decision information, and match the corresponding target decision model based on the target decision requirement; Acquire internal attribute information and external attribute information about the petition text, and construct an attribute information data set according to the internal attribute information and the external attribute information; The attribute information data set is used as input, and the decision analysis result of the petition text is determined through the target decision model.

9. A system for automatically circulating petitions, characterized in that: It comprises a processor and a memory; wherein the memory stores a computer program, and the computer program is used for the processor to load and execute the automatic circulation method of letters and visits as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and the instructions are used by the processor to load and execute the automatic circulation method of letters and visits as described in any one of claims 1-8.

Images