Administrative penalty case-involved amount extraction method and system
Through the combination of rule algorithms and deep learning models, administrative penalties are handled, and the problem of high misjudgment rate of withdrawal of administrative penalties in the existing technology is solved, and more efficient and accurate withdrawal of the amount involved is achieved.
Patent Information
- Application Number
- CN202410810950.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-21
- Publication Date
- 2025-01-10
AI Technical Summary
The existing risk monitoring auxiliary system has a high misjudgment rate when withdrawing administrative penalty amounts, especially when the penalty content is complex or the amount is repeatedly mentioned, resulting in the problem of extraction errors and missing data extraction.
The classification model based on rules algorithm is used to obtain the feature label of administrative penalty content, combine regular expressions and deep learning models, and extract sentences and amounts of paragraphs, and handle administrative penalties without amount, single amount and multiple amounts of feature, and obtain them through training models. The amount of fine and the amount of illegal income confiscated, and the sum will be added.
The efficiency and accuracy of withdrawing amounts involved in administrative penalties have been improved, and the effectiveness and operation efficiency of different content processing plans have been ensured.
Smart Images
Figure CN119293541A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of natural language processing, and in particular to a method and system for extracting the amount involved in an administrative penalty case. Background Art
[0002] Administrative punishment refers to the specific administrative behavior of the administrative subject to impose administrative sanctions on the counterparty who violates administrative law norms but has not yet committed a crime in accordance with the statutory powers and procedures, and property fines are the most widely used form of punishment in administrative punishment, which mainly includes fines and confiscation of property. As an important dimension for judging the operating risks of market entities, with the continuous growth of administrative punishment data, more and more artificial intelligence and data mining technologies are also being applied to administrative punishment data practice, and the extraction of the amount involved is the most important and challenging content.
[0003] In the existing risk monitoring auxiliary system, a simple regular expression matching method is often used to extract the amount of administrative penalties. Although the above method can handle simple situations, when the administrative penalty content is more complicated or the amount is mentioned repeatedly, the misjudgment rate is too high, resulting in extraction errors, missed data and other problems. Summary of the invention
[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method and system for extracting the amount involved in an administrative penalty.
[0005] The purpose of the present invention is achieved through the following technical solution: A method for extracting the amount involved in an administrative penalty case, comprising the following steps:
[0006] S1: Input the administrative penalty content into the classification model based on the rule algorithm to obtain the corresponding administrative penalty content feature label;
[0007] S2: For administrative penalties without amount features, the extracted amount is output as "0" and the operation ends; for administrative penalties with a single amount feature, the amount in the paragraph is extracted using a regular expression; for administrative penalties with multiple amount features, the next step is entered;
[0008] S3: Use the deep learning model to divide the paragraph into sentences, extract the amount in each sentence according to the rules, and then obtain the fine amount and the amount of confiscated illegal income through the training model;
[0009] S4: Add up the amounts extracted from each clause of the fine amount and the amount of confiscated property amount to obtain the amount of administrative penalty involved.
[0010] Preferably, in step S1, the keywords and the amount values before the keywords in the administrative penalty content are matched by rules, and the number of occurrences is counted to obtain the corresponding administrative penalty content feature labels.
[0011] Preferably, the keywords include yuan and ten thousand yuan.
[0012] Preferably, for the single amount characteristic administrative penalty, the following steps are also included:
[0013] A1: Extract the sentences containing the amount. The sentence format is feature word + number + unit.
[0014] A2: The regular expression uses enumeration to extract the amount in the clause and converts it into Arabic numerals.
[0015] A3: Determine the category of the amount based on the feature words. If it is determined to be a fine amount, determine the validity of the amount by extracting keywords. If it is determined to be an invalid amount, the penalty amount is assigned a value of "0". If it is determined to be a valid amount, output the penalty amount by unit.
[0016] A4: When it is determined to be a confiscated property amount, the validity of the amount is determined by extracting keywords. When it is determined to be an invalid amount, the penalty amount is assigned a value of "0". When it is determined to be a valid amount, the confiscated property amount is output.
[0017] Preferably, in step A3 and step A4, the keywords include should and if.
[0018] Preferably, for administrative penalties with multiple amounts of money, the following steps are also included:
[0019] B1: Extract sentences containing multiple amounts. The sentence format is feature word + number + unit.
[0020] B2: Determine the validity of the amount of each extracted clause. If the amount is determined to be invalid, remove the clause.
[0021] B3: Split the effectiveness clauses into feature words + numbers + amounts + description words through the training model;
[0022] B4: Repeat steps A2 to A4 to extract the penalty amount and confiscated property amount of each clause, and add them up to obtain the fine amount and confiscated property amount.
[0023] A system for extracting the amount of money involved in administrative penalties includes the above-mentioned method for extracting the amount of money involved in administrative penalties, and also includes a text classification model module, a sentence module, a valid amount extraction module and a summation module.
[0024] The text classification model module is used to assign different administrative penalty content feature labels such as no amount, single amount or multiple amounts to administrative penalty content;
[0025] The sentence segmentation module is used to extract sentences involving amounts;
[0026] The effective amount extraction module is used to input the sentence into the amount extraction model, and determine the effective amount and invalid amount corresponding to the fine and confiscation in the amount involved in the case through the model;
[0027] The summation module is used to add up the extracted valid amounts to obtain the fine amount and the confiscated property amount respectively.
[0028] The present invention has the following advantages: the present invention obtains the corresponding administrative penalty content feature labels through rules and NLP algorithm models. When it is a single amount, it is processed through a regular expression, and when it is multiple amounts, it is processed through an algorithm model, thereby providing corresponding solutions for different contents, which not only ensures the effectiveness of regular processing data, but also improves the operating efficiency of model processing data, thereby improving the overall efficiency and accuracy of extracting the amount involved in administrative penalties. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 This is a structural diagram of the process of extracting the amount involved in administrative penalties;
[0030] Figure 2 This is a structural diagram of the system for extracting the amount involved in administrative penalties. DETAILED DESCRIPTION
[0031] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions 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 part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.
[0032] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments 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.
[0033] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.
[0034] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.
[0035] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the product of the invention is usually placed when in use, or the positions or positional relationships commonly understood by those skilled in the art, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.
[0036] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" 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 it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0037] In this embodiment, if Figure 1 As shown, a method for extracting the amount involved in an administrative penalty case includes the following steps:
[0038] S1: Input the administrative penalty content into the classification model based on the rule algorithm to obtain the corresponding administrative penalty content feature label;
[0039] S2: For administrative penalties without amount features, the extracted amount is output as "0" and the operation ends; for administrative penalties with a single amount feature, the amount in the paragraph is extracted using a regular expression; for administrative penalties with multiple amount features, the next step is entered;
[0040] S3: Use the deep learning model to divide the paragraph into sentences, extract the amount in each sentence according to the rules, and then obtain the fine amount and the amount of confiscated illegal income through the training model;
[0041] S4: Add up the amounts extracted from each sentence of the fine amount and the amount of confiscated property to obtain the amount of administrative penalty involved. Obtain the corresponding administrative penalty content feature label through rules and NLP algorithm models. When it is a single amount, it is processed through regular expressions, and when it is multiple amounts, it is processed through algorithm models, so as to provide corresponding solutions for different contents, which not only ensures the effectiveness of regular processing data, but also improves the operating efficiency of model processing data, thereby improving the efficiency and accuracy of extracting the amount of administrative penalty involved.
[0042] Further, in step S1, keywords and the amount values before the keywords in the administrative penalty content are matched by rules, and the number of occurrences is counted to obtain the corresponding administrative penalty content feature tags. Preferably, the keywords include "yuan" and "ten thousand yuan".
[0043] In this embodiment, for the administrative penalty with a single amount feature, the following steps are further included:
[0044] A1: Extract the clauses containing amounts. The format of the clause is feature word + number + unit. For example: Confiscate 20,000 yuan of illegal income.
[0045] A2: Use regular expressions to extract the amounts in the clauses by enumeration method and process them into Arabic numerals. Specifically, for example, process the Chinese character "two hundred and thirty yuan" into "230 yuan"; or process "at 50% of the maximum fine of 50,000 yuan" into "50,000 * 50% = 25,000 yuan".
[0046] A3: Determine the category to which the amount belongs according to the feature word. When it is determined to be a fine amount, determine the validity of the amount by extracting keywords. When it is determined to be an invalid amount, the penalty amount is assigned "0". When it is determined to be a valid amount, the penalty amount is output according to the unit. Specifically, for example: [Confiscate] 234.4 yuan [of illegal income], because it hits the feature word of the amount of illegal property, it is processed as confiscating the amount of illegal income.
[0047] A4: When it is determined to be the amount of confiscated property, determine the validity of the amount by extracting keywords. When it is determined to be an invalid amount, the penalty amount is assigned "0". When it is determined to be a valid amount, the amount of confiscated property is output. Preferably, in steps A3 and A4, the keywords include "should" and "if". Specifically, the feature word before the invalid amount contains non-deterministic words. Here, "should" and "if" are just examples, not all non-deterministic words.
[0048] In this embodiment, for the administrative penalty with multiple amount features, the following steps are further included:
[0049] B1: Extract the clauses containing multiple amounts. The format of the clause is feature word + number + unit;
[0050] B2: Judge the validity of the amounts in each of the extracted clauses. When it is determined to be an invalid amount, the clause where it is located is excluded;
[0051] B3: Split the valid clauses through a trained model. The split format is feature word + number + amount + description word. Specifically, for example: Impose fines of XXXX yuan and YYYY yuan on the XX illegal act and YY illegal act of XXXX company respectively, and split it into "Impose a fine of XXXX yuan on the XX illegal act of XXXX company", "Impose a fine of YYYY yuan on the YY illegal act of XXXX company".
[0052] B4: Repeat steps A2 to A4 to extract the penalty amount and confiscated property amount of each clause, and add them up to obtain the fine amount and confiscated property amount. Specifically, invalid amounts generally appear in data ranges. If the clauses are split first, it is easy to cause the invalidity of the amount validity judgment. Therefore, the validity judgment is performed first, and then the clauses are split.
[0053] In this embodiment, if Figure 2 As shown, a system for extracting the amount involved in an administrative penalty case includes the above-mentioned method for extracting the amount involved in an administrative penalty case, and also includes a text classification model module, a sentence module, a valid amount extraction module and a summation module.
[0054] The text classification model module is used to assign different administrative penalty content feature labels such as no amount, single amount or multiple amounts to administrative penalty content;
[0055] The sentence segmentation module is used to extract sentences involving amounts;
[0056] The effective amount extraction module is used to input the sentence into the amount extraction model, and determine the effective amount and invalid amount corresponding to the fine and confiscation in the amount involved in the case through the model;
[0057] The summation module is used to add up the extracted valid amounts to obtain the fine amount and the confiscated property amount respectively. Specifically, if the "no amount" label text is directly assigned an output result, the penalty amount is assigned to "0", and the confiscated property amount is assigned to "0"; if the "single amount" or "multiple amounts" feature label text flows to the sentence module, the sentence module extracts sentences involving the amount based on the deep learning algorithm model, and puts forward relevant sentences involving the amount in the text. If it is a "single amount" label text, it flows directly to the effective amount extraction module. If it is a "multiple amounts" label text, the sentence is processed into a standard sentence of the minimum order, and then flows to the effective amount extraction module. The effective amount extraction module inputs the sentence-related amount text into the amount extraction model through the amount extraction unit, the amount classification unit, and the effective and invalid amount judgment unit, and judges the effective amount and invalid amount corresponding to the fine and confiscation in the amount involved through the model. Among them, the "single amount" label text directly outputs the corresponding result after passing through the effective amount extraction module, and the "multiple amounts" label text flows to the summation module. The summation module superimposes and sums the extracted effective amounts according to the amount classification through the amount addition unit to obtain the fine amount and the confiscated property amount respectively.
[0058] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for extracting the amount involved in an administrative penalty, characterized by: The following steps are involved: S1: Input the administrative penalty content into the classification model based on the rule algorithm to obtain the corresponding administrative penalty content feature label; S2: For administrative penalties without amount features, the extracted amount is output as "0" and the operation ends; for administrative penalties with a single amount feature, the amount in the paragraph is extracted using a regular expression; for administrative penalties with multiple amount features, the next step is entered; S3: Use the deep learning model to divide the paragraph into sentences, extract the amount in each sentence according to the rules, and then obtain the fine amount and the amount of confiscated illegal income through the training model; S4: Add up the amounts extracted from each clause of the fine amount and the amount of confiscated property amount to obtain the amount of administrative penalty involved.
2. The method for extracting the amount of money involved in administrative penalties according to claim 1 is characterized by: In step S1, the keywords and the amount before the keywords in the administrative penalty content are matched by rules, and the number of occurrences is counted to obtain the corresponding administrative penalty content feature label.
3. The method for extracting the amount involved in administrative penalties according to claim 2 is characterized by: Keywords include yuan and ten thousand yuan.
4. The method for extracting the amount involved in administrative penalties according to claim 3 is characterized by: For administrative penalties with single amount characteristics, the following steps are also included: A1: Extract the sentences containing the amount. The sentence format is feature word + number + unit. A2: The regular expression uses enumeration to extract the amount in the clause and converts it into Arabic numerals. A3: Determine the category of the amount based on the feature words. If it is determined to be a fine amount, determine the validity of the amount by extracting keywords. If it is determined to be an invalid amount, the penalty amount is assigned a value of "0". If it is determined to be a valid amount, output the penalty amount by unit. A4: When the amount is determined to be a confiscated amount, the validity of the amount is determined by extracting keywords. When the amount is determined to be invalid, the penalty amount is assigned a value of "0". When the amount is determined to be valid, the confiscated amount is output.
5. The method for extracting the amount involved in administrative penalties according to claim 4 is characterized by: In step A3 and step A4, the keywords include should and if.
6. The method for extracting the amount of money involved in administrative penalties according to claim 5 is characterized by: For administrative penalties with multiple amounts, the following steps are also included: B1: Extract sentences containing multiple amounts. The sentence format is feature word + number + unit. B2: Determine the validity of the amount of each extracted clause. If the amount is determined to be invalid, remove the clause. B3: Split the effectiveness clauses into feature words + numbers + amounts + description words through the training model; B4: Repeat steps A2 to A4 to extract the penalty amount and confiscated property amount of each clause, and add them up to obtain the fine amount and confiscated property amount.
7. A system for extracting the amount of money involved in an administrative penalty case, comprising the method for extracting the amount of money involved in an administrative penalty case according to any one of claims 1 to 6, characterized in that: It also includes a text classification model module, a sentence segmentation module, a valid amount extraction module, and a summation module. The text classification model module is used to assign different administrative penalty content feature labels such as no amount, single amount or multiple amounts to the administrative penalty content; The sentence segmentation module is used to extract sentences involving amounts; The effective amount extraction module is used to input the sentence into the amount extraction model, and determine the effective amount and invalid amount corresponding to the fine and confiscation in the amount involved in the case through the model; The summing module is used to add up and sum the extracted valid amounts to obtain the fine amount and the confiscated property amount respectively.