Financial dispute mediation information processing method, electronic equipment and storage medium
By converting the voice information in the dispute mediation process into text and matching keywords, the parties' willingness to repay is determined, which solves the problem that the mediator cannot know the parties' willingness to repay, and improves the success rate and efficiency of mediation.
Patent Information
- Application Number
- CN202510153229.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-12
- Publication Date
- 2025-05-30
AI Technical Summary
During the mediation of financial disputes, the mediator was unable to effectively know the parties' willingness to repay, resulting in poor mediation results.
By obtaining voice information during the dispute mediation process, converting it into text information, and matching keywords, the parties' willingness to repay are determined.
This method can automatically analyze the parties' repayment intentions, provide a reliable basis for the mediator, and improve the mediation success rate and efficiency.
Smart Images
Figure CN120067310A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of dispute mediation, and more particularly to a method for processing financial dispute mediation information, an electronic device, and a storage medium. Background Art
[0002] With the continuous implementation by the central government of the transformation of the contradiction resolution mode to the form of "centering on mediation and using litigation as the ultimate guarantee" to improve the legalization level of the prevention and resolution of contradictions and disputes, the number of court mediation cases has also increased sharply. During the mediation process, mediators need to communicate with a large number of parties multiple times to negotiate mediation plans. Since the repayment capabilities and repayment intentions of the parties are different, the mediation success rates will also vary greatly.
[0003] However, in the current mediation scenario, mediators can only obtain some original materials of the parties and cannot well understand the repayment intentions of the parties, which results in mediators not being able to fully utilize their advantages to improve the final effect of mediation.
[0004] In view of this, the present invention is specifically proposed. Summary of the Invention
[0005] In consideration of the above problems, the present invention is proposed. According to one aspect of the present invention, there is provided a method for processing financial dispute mediation information, including: Obtaining voice information during the dispute mediation process; Converting the voice information into corresponding text information, where the text information includes party text information corresponding to the parties; Performing keyword matching on the party text information in the text information to determine the repayment intention keywords included in the party text information; Determining the repayment intention of the party based at least on the repayment intention keywords; Generating a processing result, where the processing result includes the repayment intention of the party.
[0006] Exemplarily, the determining the repayment intention of the party based at least on the repayment intention keywords includes: Searching in a user intention keyword library for a target keyword set including the repayment intention keywords; wherein, the user intention keyword library includes multiple keyword sets, the multiple keyword sets correspond to multiple repayment intentions one by one, and the multiple keyword sets include the target keyword set; Determining the repayment intention of the party as the repayment intention corresponding to the target keyword set.
[0007] Exemplarily, the determining the repayment intention of the party based at least on the repayment intention keywords includes: When the number of the repayment intention keywords is 0, input the party text information into a preset prompt template to obtain a target prompt; use a large language model to process the target prompt to determine the repayment willingness of the party.
[0008] Exemplarily, there are multiple parties; the method further includes: for each of the multiple parties, obtain the party information of this party; determine the repayment ability of this party according to the party information; rank the multiple parties based on the repayment ability of each of the multiple parties to obtain a repayment ability ranking result; wherein, the processing result includes the repayment willingness of each of the multiple parties and the repayment ability ranking result.
[0009] Exemplarily, the party information includes age, gender, income, the number of co-debt cases, and the average overdue days; the determining the repayment ability of this party according to the party information includes: input the age, gender, income, the number of co-debt cases, and the average overdue days of this party into a pre-trained repayment ability prediction model to determine the repayment ability of this party.
[0010] Exemplarily, the pre-trained repayment ability prediction model is trained in the following manner: input sample information into the repayment ability prediction model to obtain a predicted repayment ability, where the sample information includes the age, gender, income, the number of co-debt cases, and the average overdue days of the sample party; optimize the repayment ability prediction model based on the difference between the predicted repayment ability and the actual repayment ability to obtain the pre-trained repayment ability model; wherein, the actual repayment ability is the total repayment amount ratio of the sample party.
[0011] Exemplarily, before performing keyword matching on the party text information in the text information, the method further includes: preprocess the text information to obtain preprocessed text information; wherein, the performing keyword matching on the party text information in the text information includes: perform keyword matching on the party text information in the preprocessed text information.
[0012] Exemplarily, the text information includes a plurality of text segments and role tags corresponding to the plurality of text segments one by one; the preprocessing of the text information includes: Based on a preset correspondence, the role tags corresponding to the plurality of text segments one by one are unified into mediators and parties; For any two adjacent text segments among the plurality of text segments, when the role tags corresponding to the two text segments are the same, the two text segments are merged into the same text segment; Wherein, the party text information includes each text segment with a role tag of a party; And / or, The preprocessing of the text information includes: Performing text cleaning on the text information to remove redundant information in the text information, where the redundant information includes any one or several of special symbols, modal particles, and repeated words.
[0013] According to another aspect of the present invention, there is provided an electronic device, including a processor and a memory, where a computer program is stored in the memory, and the processor is configured to execute the computer program to implement the method as described above.
[0014] According to still another aspect of the present invention, there is provided a computer-readable storage medium storing a computer program / instructions, and when the computer program / instructions are executed by a processor, the method as described above is implemented.
[0015] The above technical solution can automatically analyze the repayment willingness of the parties based on the voice information in the dispute mediation process. Thus, it can provide a relatively reliable basis for the subsequent mediation process of the mediator, which helps to promote the efficient progress of the mediation work and improve the final effect of the mediation. In short, this method can assist the mediator's mediation work, promote the mediator to handle financial disputes in a more scientific and systematic manner, thereby improving the mediation success rate and optimizing the allocation of mediation resources.
[0016] The above description is only an overview of the technical solution of the present invention. In order to be able to understand the technical means of the present invention more clearly, it can be implemented according to the content of the specification. And in order to make the above and other objects, features, and advantages of the present invention more obvious and understandable, the following specifically illustrates the specific embodiments of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] By describing the embodiments of the present invention in more detail in conjunction with the drawings, the above and other objects, features, and advantages of the present invention will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings, the same reference numerals generally represent the same components or steps.
[0018] Figure 1 Schematic flowchart showing a method for processing financial dispute mediation information according to an embodiment of the present invention; Figure 2 Schematic block diagram showing an electronic device according to an embodiment of the present invention. Detailed implementation manners
[0019] In order to make the objectives, technical solutions and advantages of the present invention more apparent, exemplary embodiments according to the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited by the exemplary embodiments described herein. Based on the embodiments of the present invention described herein, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present invention.
[0020] As described above, in the current mediation scenario, mediators can only obtain some original materials of the parties and cannot well understand the repayment willingness of the parties, which results in mediators being unable to give full play to their advantages to improve the final effect of mediation. In view of this, the present invention provides a method for processing financial dispute mediation information, an electronic device and a storage medium. This method can process financial dispute mediation information to obtain the repayment willingness of the parties, thereby providing a basis for the mediation work of mediators, which helps to improve the final effect of mediation.
[0021] According to one aspect of the embodiments of the present invention, a method for processing financial dispute mediation information is provided. Figure 1 Schematic flowchart showing a method for processing financial dispute mediation information according to an embodiment of the present invention. As Figure 1 shown, the method may include the following steps S110, step S120, step S130, step S140 and step S150.
[0022] In step S110, voice information during the dispute mediation process is obtained.
[0023] It can be understood that the dispute mediation process refers to the telephone communication between the mediator and the parties. The voice information during the dispute mediation process is the audio data during the above telephone communication. This audio data can be recorded through the built-in recording software of the telephone or other recording devices during the telephone communication. In step S110, the audio data during the above telephone communication can be obtained in real time.
[0024] It can be understood that during the financial dispute mediation process, it may be necessary to communicate with the parties multiple times. The voice information in this example can be the voice information corresponding to the first phone call during the financial dispute mediation process. By analyzing this voice information, the repayment intention of the parties can be initially understood, thus providing a reliable basis for the subsequent mediation work.
[0025] In step S120, the voice information is converted into corresponding text information, and the text information includes the party text information corresponding to the parties.
[0026] After obtaining the voice information, the voice information can be converted into text information. Optionally, the voice information can be converted into corresponding text information through any existing or future-developed speech-to-text technology. For example, the voice information can be converted into text information through an Automatic Speech Recognition (ASR) algorithm. Of course, the text information can also be obtained through the built-in speech-to-text function in the existing mediation platform system.
[0027] As described above, the dispute mediation process refers to the phone communication between the mediator and the parties. When the voice information is converted in real time, the identity of the speaker can be distinguished. That is to say, in the text information, the text of the mediator and the text of the parties can be distinguished and marked respectively to determine the party text information.
[0028] In step S130, keyword matching is performed on the party text information in the text information to determine the repayment intention keywords included in the party text information.
[0029] In this example, keyword matching can be performed on the party text information in the text information, that is, querying in the party text information to determine whether the party text information contains repayment intention keywords. It can be understood that the repayment intention keywords can be used to represent the repayment intention of the parties. In some embodiments, multiple keywords for representing the repayment intention can be preset (the multiple keywords can include each keyword in the user intention keyword library below). In step S130, each keyword can be retrieved in the party text information to determine the repayment intention keywords included in the party text information. Of course, the repayment intention keywords included in the party text information can also be determined by means of regular expressions, etc., which will not be elaborated.
[0030] In step S140, the repayment intention of the parties is determined based at least on the repayment intention keywords.
[0031] According to the above description, the repayment intention keywords can be used to represent the repayment willingness of the parties. After determining the repayment intention keywords included in the text information of the parties, the repayment willingness of the parties can be determined based on the repayment intention keywords. The repayment willingness of the parties can be represented by the repayment willingness level. For example, low repayment willingness, relatively low repayment willingness, relatively high repayment willingness, and high repayment willingness can be represented by 1, 2, 3, and 4 respectively. Of course, the repayment willingness can also be represented in other ways, which will not be elaborated.
[0032] In step S150, a processing result is generated, and the processing result includes the repayment willingness of the parties.
[0033] In the solution of this example, the processing result can be generated at least based on the repayment willingness of the parties. This processing result can provide a basis for the mediation work of the mediator. Of course, the processing result can also include other information. For example, it can also include the basic information of the parties, which will not be elaborated.
[0034] The above technical solution can automatically analyze the repayment willingness of the parties based on the voice information in the dispute mediation process. Thus, it can provide a relatively reliable basis for the subsequent mediation process of the mediator, which helps to promote the efficient progress of the mediation work and improve the final effect of the mediation. In short, this method can assist the mediator's mediation work, promote the mediator to handle financial disputes in a more scientific and systematic way, thereby improving the mediation success rate and optimizing the allocation of mediation resources.
[0035] Exemplarily, in step S140, to determine the repayment willingness of the parties based at least on the repayment intention keywords, it includes: searching for a target keyword set including the repayment intention keywords in the user intention keyword library; wherein, the user intention keyword library includes multiple keyword sets, the multiple keyword sets correspond to multiple repayment willingness levels one by one, and the multiple keyword sets include the target keyword set; determining that the repayment willingness of the parties is the repayment willingness corresponding to the target keyword set.
[0036] In this example, the user intention keyword library is a pre-established library including multiple keyword sets. In some embodiments, multiple keywords that can reflect the repayment willingness of the parties can be determined in advance according to historical financial dispute mediation data, and the user intention keyword library can be constituted based on these keywords. Of course, a large language model can also be used to expand the keywords representing different repayment willingness levels to obtain more keywords. In this way, it helps to improve the accuracy of judging the repayment willingness of the parties.
[0037] Optionally, the initial keywords can be expanded in the following way: input the initial keywords into a keyword template; use a large language model to process the keyword template to obtain multiple keywords with the same meaning as the initial keywords.
[0038] In a specific embodiment, the keyword template can be: ``` The following word is a keyword in the call record between the mediator and the parties, indicating a high willingness to repay. Please give me 3 words with the same meaning, in a more colloquial form, separated by 1., 2., 3.: {Initial keyword} ``` When the initial keyword is "willing", it can be input into the above keyword template. In this embodiment, the output results are: 1. willing; 2. willing to repay; 3. no problem.
[0039] As described above, multiple keyword sets correspond one-to-one with multiple repayment intentions. That is, multiple keyword sets respectively represent different repayment intentions. Each keyword set can include multiple different keywords, and multiple different keywords in the same keyword set all correspond to the same repayment intention. In some embodiments, the multiple keyword sets included in the user intention keyword library can be: Low - keywords = ["get out", "your mother", "his mother", "get lost", "not accept", "not willing"]; Lower - keywords = ["poor", "have no money", "can't", "have no way"]; High - keywords = ["willing", "accept"].
[0040] The above technical solution can accurately determine the repayment intention of the parties by performing keyword matching on the repayment intention keywords in the parties' text in the user keyword library, thereby providing a more accurate basis for the mediation work of the mediator.
[0041] Exemplarily, determining the repayment intention of the parties based at least on the repayment intention keywords includes: when the number of repayment intention keywords is 0, inputting the parties' text information into a preset prompt word template to obtain a target prompt word; using a large language model to process the target prompt word to determine the repayment intention of the parties.
[0042] Optionally, the large language model can adopt any existing or future - developed open - source or closed - source large language model. The large language model can include but is not limited to GPT - 3.5, Tongyi Qianwen, Wenxin Yiyan, LLAMA, ChatGLM, etc.
[0043] The preset prompt word template can be set as needed. For example, the preset prompt word template can be: ``` You are a mediator handling debt disputes. Now there is a call record between a robot mediator and a party, and you need to judge the party's willingness to repay. The judgment rules are as follows: 1. There are four levels of repayment willingness, namely [low, relatively low, relatively high, high], corresponding to the numbers [1, 2, 3, 4] respectively; 2. If the party doesn't say anything, or says abusive words, or clearly states that they are not willing to accept mediation, then the level of the party's repayment willingness is low; 3. If the party reveals information such as having no money, then the level of repayment willingness is relatively low; 4. If the party doesn't clearly indicate whether they are willing to accept mediation, then the level of the party's repayment willingness is relatively high; 5. If the party clearly indicates acceptance of mediation, then the level of repayment willingness is high. The following is the call record between the robot mediator and the party: {Dialogue record} Please output the corresponding repayment level of this party, just output the number corresponding to one level, and do not output other content ``` In the solution of this example, the repayment intention keywords in the party's text information can be determined first. If there are repayment intention keywords in the party's text information, the party's repayment willingness can be directly determined based on these repayment intention keywords. If there are no repayment intention keywords in the party's text information, the large language model can be used to analyze the party's text information to determine the party's repayment willingness.
[0044] The above solution can directly use the large language model to determine the party's repayment willingness when there are no repayment intention keywords in the party's text information, which can ensure that the party's repayment willingness can be determined in different situations (regardless of whether there are repayment intention keywords), thus better assisting the mediator in the mediation work.
[0045] Exemplarily, there are multiple parties; the method further includes: for each party among the multiple parties, obtaining the party's information; according to the party's information, determining the party's repayment ability; based on the repayment ability of each party among the multiple parties, sorting the multiple parties to obtain a repayment ability sorting result; where the processing result includes the repayment willingness of each of the multiple parties and the repayment ability sorting result.
[0046] The party's information may include the party's income. By analyzing the party's information, a more accurate portrait of the party can be drawn, so that the financial situation, credit risk, etc. of each party can be more accurately understood, and thus the party's repayment ability can be more accurately determined.
[0047] In the solution of this example, after obtaining the repayment capabilities of multiple parties respectively, the parties can be sorted according to their respective repayment capabilities. For example, the multiple parties can be sorted from high to low according to their repayment capabilities. This sorting result can provide a relatively accurate basis for the mediator, and the mediator can allocate the mediation order according to this sorting result. For example, the parties with higher repayment capabilities can be given priority for mediation, which helps to improve the efficiency of the mediation work.
[0048] In some embodiments, the determined repayment capability of the party can be the ratio of the total recovered amount of the party predicted based on the party information. Of course, it can also be the repayment capability score predicted based on the party information, which will not be elaborated.
[0049] The above technical solution can accurately evaluate the repayment capability and repayment willingness of the parties, thereby providing strong decision-making support and intuitive decision-making reference for the mediator, and further effectively improving the efficiency and success rate of the mediation work.
[0050] Exemplarily, the party information includes age, gender, income, the number of co-debt cases, and the average overdue days; determining the repayment capability of the party according to the party information includes: inputting the age, gender, income, the number of co-debt cases, and the average overdue days of the party into a pre-trained repayment capability prediction model to determine the repayment capability of the party.
[0051] The repayment capability prediction model can be any existing or future-developed neural network model. For example, it can be a convolutional neural network model, a fully connected neural network model, etc., which will not be elaborated.
[0052] In the solution of this example, by combining the party portrait and the repayment capability prediction model, the repayment capability of the party can be predicted more accurately, thereby providing an intuitive decision-making reference for the mediator and improving the scientific nature of the mediation work.
[0053] Exemplarily, the pre-trained repayment capability prediction model is trained in the following manner: inputting the sample information into the repayment capability prediction model to obtain the predicted repayment capability, where the sample information includes the age, gender, income, the number of co-debt cases, and the average overdue days of the sample party; optimizing the repayment capability prediction model based on the difference between the predicted repayment capability and the actual repayment capability to obtain the pre-trained repayment capability model; where the actual repayment capability is the ratio of the total recovered amount of the sample party.
[0054] The sample information can be the data of the parties in historical financial disputes. The actual repayment ability can be the total repayment amount ratio of the parties in historical financial disputes, that is, the ratio of the total amount of the repaid amount to the total amount of the amount to be repaid by the parties. In a specific embodiment, training sets can be constructed using data such as banks, WeChat Pay Loans, and Huabei / Jiebei data related to Ant Financial. The training sets include information of multiple sample parties. In the solution of this example, the age, gender, income, number of co-debt cases, and average overdue days of each sample party can be used as the input items of the repayment ability prediction model, and the repayment ability prediction model can be optimized through the difference between the output result of the model and the actual repayment ability. Thereby, it helps to improve the prediction accuracy of the model.
[0055] Exemplarily, before performing keyword matching on the party text information in the text information, the method further includes: preprocessing the text information to obtain the preprocessed text information; wherein, performing keyword matching on the party text information in the text information includes: performing keyword matching on the party text information in the preprocessed text information.
[0056] In the solution of this example, after converting the voice information into the corresponding text information, the text information can be preprocessed in advance. The preprocessing methods include but are not limited to unifying words with the same meaning in different expressions, unifying the labels of each text segment, merging text segments with the same role label, and removing special characters, modal particles, repeated words, etc. in the text information. The specific preprocessing method can be selected according to actual needs and will not be elaborated. Through preprocessing the text information, the solution of this example can remove the noise in the text information, thereby avoiding the noise in the text information from affecting the accuracy of keyword matching in subsequent steps.
[0057] Exemplarily, the text information includes multiple text segments and role labels corresponding to the multiple text segments one by one; preprocessing the text information includes: based on a preset correspondence, unifying the role labels corresponding to the multiple text segments one by one into mediators and parties; for any two adjacent text segments among the multiple text segments, when the role labels corresponding to the two text segments are the same, merging the two text segments into the same text segment; wherein, the party text information includes each text segment with the role label of party. and / or, Preprocessing the text information includes: cleaning the text information to remove redundant information in the text information, and the redundant information includes any one or more of special symbols, modal particles, and repeated words.
[0058] Optionally, the text information includes multiple text segments and role tags corresponding one-to-one to the multiple text segments; preprocessing the text information includes: based on a preset correspondence, uniformly converting the role tags corresponding one-to-one to the multiple text segments into mediators and parties; for any two adjacent text segments among the multiple text segments, when the role tags corresponding to the two text segments are the same, merging the two text segments into the same text segment; wherein, the party text information includes each text segment with the role tag of party.
[0059] In this article, when converting to text information using a mediation platform system or the like, the mediation platform system will add prefixes (i.e., role tags) to both parties of the call. Thus, the obtained text information includes multiple text segments and role tags corresponding one-to-one to the multiple text segments. For example, one text segment and the corresponding role tag can be: "Collector: May I ask when you will repay the money on your side." However, in the related art, although some mediation platform systems or speech-to-text algorithms will add role tags to both parties of the call when converting speech to text, the role tags added by different systems or algorithms are not unified. In the solution of this example, all role tags can be uniformly converted into mediators and parties based on a preset correspondence, thereby facilitating the determination of the party text information in the subsequent steps.
[0060] The preset correspondence can be set as needed. For example, the preset correspondence can be set according to the selected mediation platform system or algorithm for speech-to-text conversion to uniformly map the role tags automatically generated by the above mediation platform system or algorithm into mediators and parties. In one embodiment, the preset correspondence can be: Agent, mediator, collector correspond to mediator, Customer, party, debtor correspond to party. Of course, the above preset correspondence is only an example, and the user can adjust the preset correspondence according to needs, which will not be elaborated.
[0061] After unifying the role tags, each text segment corresponds to a role tag of either party or mediator. In this case, the role tags of each text segment can be traversed in sequence. If the role tags corresponding to two adjacent text segments are the same, the two text segments can be merged into the same text segment. At this time, the text segments corresponding to the mediator and the party in the text information appear alternately. In this case, the latest text segment among the text segments with the role tag of party is the latest question of the party, that is, the party text information.
[0062] This optional solution can improve the quality of the text information by unifying the role tags of the text segments and merging adjacent text segments with the same role tag, thereby helping to more accurately determine the party text information and accurately determine the party's repayment intention keywords in the party text information.
[0063] Optionally, preprocess the text information, including: cleaning the text information to remove redundant information in the text information, where the redundant information includes any one or several of special symbols, modal particles, and repeated words.
[0064] In some embodiments, the redundant information may include special symbols. Cleaning the text information to remove the redundant information in the text information may include: using regular expressions to remove special symbols in the text information. The regular expression may be: "content = re.sub(r"[\s+! / _,$%^*(+\''\):+——()?【】“”!,。?、~@#￥%……&*()]+", '', content)".
[0065] In some embodiments, the redundant information may include modal particles. In some embodiments, modal particles (such as "um", "ah", "uh", etc.) may be pre-stored in a modal particle stop word list. Then, jieba segmentation can be used to determine whether each token after segmentation exists in the modal particle stop word list. If it exists, the token is removed. The above method of removing modal particles is only an example, and those skilled in the art can also use other methods to remove modal particles. For example, modal particles in the text information can be removed through regular expressions.
[0066] In some embodiments, the redundant information may include repeated words. Repeated words may be, for example, meaningless repeated words such as "wait for you, wait for you" and "let me think, let me think". In some embodiments, jieba segmentation can be used to loop through each token for judgment. If the token is the same as the previous token, the token is removed. In other embodiments, the tokens can be pre-screened to prevent key information such as dates and amounts from being removed. In a specific embodiment, three functions can be preset: a Chinese date judgment function, an English date judgment function, and a number judgment function. After using jieba segmentation to process the text information, each token after segmentation can be looped for determination. If the token does not belong to a Chinese date, an English date, or a number, then continue to determine whether it is equal to the previous token. If the token is the same as the previous token, the token is removed. The Chinese date judgment function, the English date judgment function, and the number judgment function can be selected as needed. For example, regular expressions can be used to construct the Chinese date judgment function, the English date judgment function, and the number judgment function. Taking the Chinese date judgment as an example, regular expressions can be used to determine whether there is "number + year", "number + month", or "number + day". Those skilled in the art can understand the specific construction method of regular expressions and will not be elaborated.
[0067] In a preferred embodiment, the redundant information includes special symbols, modal particles, and repeated words. The solution of this embodiment can remove special symbols, modal particles, and repeated words, thereby improving the quality of the obtained text information.
[0068] The above optional solution can improve the quality of the finally obtained text information by removing redundant information in the text information by means of text cleaning, which is beneficial to improving the accuracy of the finally obtained target response words and / or the finally obtained guiding words.
[0069] Optionally, the preprocessing of the text information may include the above-mentioned preprocessing based on role tags, or may include the above-mentioned preprocessing based on text cleaning. When the preprocessing of the text information may include the above-mentioned preprocessing based on role tags and the above-mentioned preprocessing based on text cleaning, the two preprocessing methods may be executed in sequence. For example, the preprocessing based on role tags may be executed first, and then the preprocessing based on text cleaning may be executed. Another example is that the preprocessing based on text cleaning may be executed first, and then the preprocessing based on role tags may be executed. Thus, it helps to further improve the quality of the obtained text information.
[0070] According to another aspect of the embodiments of the present invention, an electronic device is further provided. Figure 2 A schematic block diagram of an electronic device according to an embodiment of the present invention is shown. As Figure 2 shown, the electronic device 200 includes: a processor 210 and a memory 220. A computer program is stored in the memory 220, and the processor 210 is configured to execute the computer program to implement the above method.
[0071] According to still another aspect of the embodiments of the present invention, a computer-readable storage medium is further provided. A computer program / instruction is stored in the storage medium, and when the computer program / instruction is executed by a processor, the above method is implemented. The storage medium may include, for example, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a portable compact disc read-only memory (CD-ROM), a USB memory, or any combination of the above storage media. The computer-readable storage medium may be any combination of one or more computer-readable storage media.
[0072] Those of ordinary skill in the art can easily understand the implementation structure, working principle, and beneficial effects of the electronic device and the computer-readable storage medium by reading the above method. For the sake of brevity, they will not be elaborated here.
[0073] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely exemplary and are not intended to limit the scope of the present invention thereto. Those of ordinary skill in the art can make various changes and modifications therein without departing from the scope and spirit of the present invention. All such changes and modifications are intended to be included within the scope of the present invention as claimed in the appended claims.
[0074] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.
[0075] In several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed.
[0076] In the specification provided herein, a large number of specific details are set forth. However, it can be understood that the embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures, and technologies have not been shown in detail so as not to obscure the understanding of this specification.
[0077] Similarly, it should be understood that, in order to streamline the present invention and assist in understanding one or more of the various inventive aspects, in the description of the exemplary embodiments of the present invention, the various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, the method of the present invention should not be construed as reflecting the intention that the claimed present invention requires more features than are expressly recited in each claim. Rather, as reflected in the corresponding claims, the inventive point lies in that the technical problem can be solved with fewer features than all the features of a single disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into the detailed description, where each claim itself serves as a separate embodiment of the present invention.
[0078] Those skilled in the art will appreciate that, except where features are mutually exclusive, any combination can be employed of all the features disclosed in this specification (including the accompanying claims, abstract, and drawings), as well as all the processes or units of any method or device so disclosed. Each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by alternative features serving the same, equivalent, or similar purpose, unless expressly stated otherwise.
[0079] In addition, those skilled in the art will understand that, although some embodiments described herein include certain features included in other embodiments but not others, combinations of features of different embodiments are meant to be within the scope of the present invention and form different embodiments. For example, in the claims, any one of the claimed embodiments can be used in any combination.
[0080] Each component embodiment of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor (DSP) can be used in practice to implement some or all of the functions of some of the modules in the electronic device according to the embodiments of the present invention. The present invention can also be implemented as a device program (such as a computer program and a computer program product) for performing part or all of the methods described herein. Such a program implementing the present invention can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, or provided on a carrier signal, or in any other form.
[0081] It should be noted that the above embodiments illustrate rather than limit the present invention, and those skilled in the art can design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in the claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In a unit claim listing several devices, several of these devices may be embodied by the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words can be interpreted as names.
[0082] As described above, it is only the specific implementation manner of the present invention or the description of the specific implementation manner. The protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should be covered within the protection scope of the present invention. The protection scope of the present invention shall be subject to the protection scope of the claims.
Claims
1. A method for processing financial dispute mediation information, characterized in that: include: Obtaining voice information during dispute mediation; Converting the voice information into corresponding text information, wherein the text information includes party text information corresponding to the party concerned; Performing keyword matching on the text information of the parties in the text information to determine the repayment intention keywords contained in the text information of the parties; Determining the repayment intention of the party at least based on the repayment intention keyword; A processing result is generated, wherein the processing result includes the repayment willingness of the party concerned.
2. The method according to claim 1, characterized in that The step of determining the repayment intention of the party at least based on the repayment intention keyword includes: Searching for a target keyword set including the repayment intention keyword in a user intention keyword library; wherein the user intention keyword library includes a plurality of keyword sets, the plurality of keyword sets correspond one to one with a plurality of repayment intentions, and the plurality of keyword sets include the target keyword set; The repayment willingness of the party concerned is determined to be the repayment willingness corresponding to the target keyword set.
3. The method according to claim 1, characterized in that The step of determining the repayment intention of the party at least based on the repayment intention keyword includes: When the number of the repayment intention keywords is 0, Inputting the text information of the party into a preset prompt word template to obtain a target prompt word; The target prompt word is processed using a large language model to determine the repayment willingness of the party concerned.
4. The method according to claim 1, characterized in that: There are multiple parties involved; the method further includes: For each of the plurality of parties, Obtain the party information of the party concerned; Determine the repayment ability of the party concerned based on the party concerned's information; Based on the repayment ability of each of the multiple parties, the multiple parties are ranked to obtain a repayment ability ranking result; The processing result includes the repayment willingness of each of the multiple parties and the repayment ability ranking result.
5. The method according to any one of claim 4, characterized in that: The party information includes age, gender, income, number of joint debt cases, and average overdue days; the party's repayment ability is determined based on the party information, including: The client's age, gender, income, number of co-debt cases and average number of overdue days are input into the pre-trained repayment ability prediction model to determine the client's repayment ability.
6. The method according to claim 5, characterized in that The pre-trained repayment ability prediction model is trained in the following way: Inputting sample information into a repayment capacity prediction model to obtain predicted repayment capacity, the sample information including the age, gender, income, number of co-debt cases and average number of overdue days of the sample party; Based on the difference between the predicted repayment ability and the actual repayment ability, optimizing the repayment ability prediction model to obtain the pre-trained repayment ability model; Among them, the actual repayment ability is the ratio of the total repayment amount of the sample parties.
7. The method according to any one of claims 1 to 6, characterized in that: Before performing keyword matching on the text information of the parties in the text information, the method further includes: Preprocessing the text information to obtain preprocessed text information; The step of performing keyword matching on the text information of the parties in the text information includes: Keyword matching is performed on the text information of the parties in the preprocessed text information.
8. The method according to claim 7, characterized in that The text information includes a plurality of text segments and role tags corresponding to the plurality of text segments one by one; the preprocessing of the text information includes: Based on the preset corresponding relationship, the role labels corresponding to the multiple text segments are unified into mediator and party; For any two adjacent text segments among the plurality of text segments, when the role labels corresponding to the two text segments are the same, the two text segments are merged into the same text segment; The text information of the party includes each text segment with the role label of the party; and / or, The preprocessing of the text information includes: The text information is cleaned to remove redundant information in the text information, wherein the redundant information includes any one or more of special symbols, modal particles and repeated words.
9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program, and the processor is used to execute the computer program to implement the method according to any one of claims 1 to 8.
10. A computer-readable storage medium, characterized in that: A computer program / instruction is stored, and when the computer program / instruction is executed by a processor, the method according to any one of claims 1 to 8 is implemented.