Financial system comment processing method and device, electronic equipment and storage medium

Abstract

The invention provides a financial system comment processing method and device, electronic equipment and a storage medium, and belongs to the technical field of financial services. The method comprises the steps of obtaining business comment data from a target financial system; determining a target comment rule model matched with the target financial system in a plurality of preset candidate comment rule models according to the target financial system; performing multi-dimensional comment recognition processing on the business comment data based on the target comment rule model to obtain a target recognition result; analyzing the target identification result based on a preset monitoring response rule to obtain a service response strategy for the service comment data; and triggering a corresponding service processing flow according to the service response strategy. The comment processing efficiency of the financial system can be improved.

Description

Technical Field

[0001] The present invention relates to the field of financial business technology, and in particular to a financial system comment processing method and device, electronic equipment and storage medium. Background Art

[0002] In the financial business field, daily work needs to be carried out through various financial systems. In order to continuously optimize system functions and improve user experience, it is necessary to timely obtain and process system comment information from system users. Comment information includes feedback, function evaluation, complaints, etc. This information is of great guiding significance for system optimization and business development.

[0003] In related technologies, the processing of financial insurance system comments often relies on manual summary and reporting by local branches or automated tools based on simple keyword matching. The manual summary method has many levels of transmission, low efficiency, and is prone to information distortion, making it difficult to cope with massive feedback information. At the same time, the automated tools based on keyword matching have low recognition accuracy when processing insurance professional terms and complex business scenarios, making it difficult to meet the insurance business's requirements for precise and intelligent comment processing. Therefore, a more efficient and accurate method for processing financial system comments is urgently needed. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to propose a financial system comment processing method and device, electronic device and storage medium, aiming to improve the efficiency of financial system comment processing.

[0005] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application proposes a financial system comment processing method, the method comprising:

[0006] Obtain business review data from the target financial system;

[0007] Determining a target review rule model matching the target financial system from among a plurality of preset candidate review rule models of the target financial system;

[0008] Performing multi-dimensional comment recognition processing on the business comment data based on the target comment rule model to obtain a target recognition result;

[0009] Analyze the target identification result based on preset monitoring response rules to obtain a business response strategy for the business review data;

[0010] According to the business response strategy, the corresponding business processing flow is triggered.

[0011] In some embodiments, determining a target review rule model that matches the target financial system from a plurality of preset candidate review rule models of the target financial system includes:

[0012] Acquiring system characteristics of the target financial system, wherein the system characteristics include a business type identifier and a function module identifier;

[0013] Perform keyword processing on the business review data to extract key field features;

[0014] According to the system characteristics and the key field characteristics, a target review rule model matching the target financial system is determined from a plurality of preset candidate review rule models.

[0015] In some embodiments, determining a target review rule model matching the target financial system from a plurality of preset candidate review rule models according to the system characteristics and the key field characteristics includes:

[0016] Matching the system feature with the applicable scenario in the candidate comment rule model to obtain a first matching score;

[0017] Matching the key field feature with the rule item in the candidate comment rule model to obtain a second matching score;

[0018] Performing weighted calculation on the first matching score and the second matching score to obtain a target matching score;

[0019] The target matching scores of the candidate review rule models are sorted, and the candidate review rule model with the highest target matching score is determined as the target review rule model.

[0020] In some embodiments, the performing multi-dimensional comment recognition processing on the business comment data based on the target comment rule model to obtain a target recognition result includes:

[0021] Based on the sensitive word rules in the target review rule model, the business review data is scanned to identify financial sensitive information and obtain the review text to be processed;

[0022] Extract keywords and calculate feature vectors of the comment text to be processed to obtain comment text features;

[0023] Based on the target comment rule model, multi-dimensional comment recognition processing is performed on the comment text to be processed and the comment text features to obtain a target recognition result.

[0024] In some embodiments, the multi-dimensional comment recognition processing is performed on the comment text to be processed and the comment text features based on the target comment rule model to obtain a target recognition result, including:

[0025] Based on the comment text features, performing text clustering analysis on the comment text to be processed to obtain comment categories of the comment text to be processed;

[0026] Based on the semantic recognition rules in the target comment rule model, contextual semantic analysis is performed on the comment text to be processed to obtain the comment tendency of the comment text to be processed;

[0027] Calculating text similarities of different business review data based on the review text features, and marking the corresponding business review data as hot reviews when the text similarities are greater than a preset threshold;

[0028] The comment category, the comment tendency and the hot comments are integrated to generate the target recognition result.

[0029] In some embodiments, the integrating the comment category, the comment tendency and the hot comments to generate the target recognition result includes:

[0030] Generating a category label of the business review data according to the review category, wherein the category label is used to characterize the attribution type of the business review data;

[0031] According to the positive or negative tendency of the review and the preset sentiment grading standard, the business review data is graded to obtain the sentiment grade of the business review data;

[0032] Determining the attention level of the business review data based on the marking status of the hot review;

[0033] The target recognition result is obtained according to the attribution type, the emotion level and the attention level.

[0034] In some embodiments, the target identification result is analyzed based on a preset monitoring response rule to obtain a business response strategy for the business review data, including:

[0035] Determining the processing priority of the business review data according to the attribution type, sentiment level and attention level in the target recognition result;

[0036] Based on the processing priority, determining a target response rule from a plurality of preset response rules;

[0037] A business response strategy for the business review data is generated according to the target response rule, wherein the business response strategy includes a processing department, a processing time limit, and a processing method.

[0038] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides a financial system comment processing device, the device comprising:

[0039] An acquisition module, used to acquire business review data from a target financial system;

[0040] A matching module, configured to determine a target review rule model matching the target financial system from among a plurality of preset candidate review rule models of the target financial system;

[0041] An identification module, used to perform multi-dimensional comment identification processing on the business comment data based on the target comment rule model to obtain a target identification result;

[0042] A strategy module, used to analyze the target identification result based on a preset monitoring response rule to obtain a business response strategy for the business review data;

[0043] The response module is used to trigger the corresponding business processing flow according to the business response strategy.

[0044] To achieve the above objectives, a third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect is implemented.

[0045] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0046] The financial system comment processing method and device, electronic device and storage medium proposed in the present application obtain business comment data from the target financial system; determine a target comment rule model that matches the target financial system from multiple preset candidate comment rule models of the target financial system; perform multi-dimensional comment recognition processing on the business comment data based on the target comment rule model to obtain a target recognition result; analyze the target recognition result based on preset monitoring response rules to obtain a business response strategy for the business comment data; and trigger the corresponding business processing flow according to the business response strategy.

[0047] This application first obtains the business review data of the target financial system and introduces multiple preset candidate review rule models. It can select the most suitable review rule model according to the business characteristics of different financial systems to improve the targeted processing of reviews. The multi-dimensional recognition method based on the target review rule model avoids the lack of accuracy caused by single-dimensional recognition, and realizes systematic review response management through analysis through preset monitoring response rules to better meet the review processing needs of the financial system. Through the formulation of business response strategies, a rapid response mechanism between review recognition and business processing is established, and the corresponding business processing flow is triggered according to the business response strategy to improve the efficiency of financial system review processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is a flow chart of a method for processing financial system comments provided by an embodiment of the present application;

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

[0050] Figure 3 yes Figure 2 Flow chart of step S203 in FIG.

[0051] Figure 4 yes Figure 1 Flow chart of step S103 in FIG.

[0052] Figure 5 yes Figure 4 Flow chart of step S403 in FIG.

[0053] Figure 6 yes Figure 5 Flow chart of step S504 in FIG.

[0054] Figure 7 yes Figure 1 Flow chart of step S104 in FIG.

[0055] Figure 8 It is a schematic diagram of a financial system comment processing method provided in an embodiment of the present application;

[0056] Fig. 9 It is another schematic diagram of the financial system comment processing method provided by the embodiment of the present application;

[0057] Fig.10 It is a structural diagram of a financial system comment processing device provided in an embodiment of the present application;

[0058] Fig.11 It is a schematic diagram of the hardware structure of the electronic device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0060] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0062] First, some nouns involved in this application are analyzed:

[0063] Financial System: refers to the information system that provides business processing, transaction management, customer service and other functions for financial institutions. It includes but is not limited to banking systems, securities trading systems, insurance business systems, payment and clearing systems, etc. These systems carry the core business processes of financial institutions and have high requirements for system stability, security and user experience. For example, the securities trading system needs to support real-time transactions, market inquiries, funds transfers and other functions, and the banking system needs to support various financial services such as deposits and withdrawals, transfers, and wealth management.

[0064] Based on this, the embodiments of the present application provide a method and device for processing financial system comments, an electronic device and a storage medium, aiming to improve the efficiency of processing financial system comments.

[0065] The financial system comment processing method and device, electronic device and storage medium provided in the embodiments of the present application are specifically illustrated through the following embodiments. First, the financial system comment processing method in the embodiments of the present application is described.

[0066] The financial system comment processing method provided in the embodiment of the present application relates to the field of financial business technology. The financial system comment processing method provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server side, and can also be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the financial system comment processing method, etc., but is not limited to the above forms.

[0067] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0068] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0069] Figure 1 is an optional flowchart of the financial system comment processing method provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S105.

[0070] Step S101, obtaining business review data from a target financial system.

[0071] Step S102: determining a target review rule model matching the target financial system from among a plurality of preset candidate review rule models of the target financial system.

[0072] Step S103, performing multi-dimensional comment recognition processing on the business comment data based on the target comment rule model to obtain a target recognition result.

[0073] Step S104: Analyze the target recognition result based on the preset monitoring response rule to obtain a business response strategy for the business review data.

[0074] Step S105: triggering the corresponding business processing flow according to the business response strategy.

[0075] In steps S101 to S105 shown in the embodiment of the present application, business review data of the target financial system is first obtained, and multiple preset candidate review rule models are introduced. The most suitable review rule model can be selected according to the business characteristics of different financial systems, thereby improving the targeted nature of review processing; a multi-dimensional recognition method based on the target review rule model avoids the lack of accuracy caused by single-dimensional recognition, and realizes systematic review response management through analysis through preset monitoring response rules, thereby better meeting the review processing needs of the financial system; through the formulation of business response strategies, a rapid response mechanism between review recognition and business processing is established, and the corresponding business processing flow is triggered according to the business response strategies, thereby improving the efficiency of review processing in the financial system.

[0076] In step S101 of some embodiments, the business review data generated by the target financial system in various business scenarios by the user is obtained in real time through the API interface. The business review data includes but is not limited to various types of review information such as feedback, function evaluation, and complaints. Among them, feedback mainly involves users' suggestions for improvement on the functions and business processes of the target financial system; function evaluation includes users' ratings of various functional modules of the target financial system, description of usage experience, etc.; complaint information includes various problems and dissatisfaction encountered by users in the process of using the target financial system. The business review data is preliminarily processed, including removing duplicate data, supplementing metadata information such as comment time and source, and storing them in the corresponding database according to different comment types. At the same time, the context information of the comments is saved, such as user operation trajectory, system response time, etc., which is of great value for subsequent comment analysis and problem location.

[0077] Exemplarily, business review data in the target financial system is obtained through the database interface. For example, business review data of the feedback type include "the business operation process is complicated, it is recommended to simplify the approval process", "the mobile banking transfer limit setting is unreasonable, it is recommended to provide a more flexible limit adjustment method"; functional evaluation types include "the financial product purchase function is easy to operate and the interface is clear", "the online banking login verification method is cumbersome and needs to be improved"; complaint types include "the system responds slowly, affecting the timeliness of transactions", "the failure to update customer information leads to obstruction of business processing", etc.

[0078] In step S102 of some embodiments, multiple preset candidate comment rule models are evaluated and selected. The candidate comment rule models may include different rule models such as custom business feedback types, function evaluations, complaints, and feedback. Each candidate comment rule model contains a complete rule definition, covering rule configurations in multiple dimensions such as comment classification standards, keyword libraries, sensitive word libraries, and business terminology comparison tables. According to the specific characteristics of the target financial system, such as business type, user group characteristics, comment data characteristics, etc., a specific algorithm is used to select the most matching target comment rule model from multiple candidate comment rule models. In the model selection process, multiple factors such as the scope of application of the model, historical application effects, and rule coverage are comprehensively considered to ensure that the selected model can meet the comment processing needs of the target financial system to the greatest extent.

[0079] See also Figure 2 In some embodiments, step S102 may include but is not limited to steps S201 to S203:

[0080] Step S201, obtaining system characteristics of the target financial system.

[0081] Among them, the system characteristics include business type identification and function module identification.

[0082] Step S202: perform keyword processing on the business review data to extract key field features.

[0083] Step S203, determining a target review rule model that matches the target financial system from a plurality of preset candidate review rule models according to the system characteristics and key field characteristics.

[0084] In step S201 of some embodiments, the system characteristics of the target financial system are obtained by calling the interface or reading the configuration file, including basic parameters such as system type, deployment environment, and user scale. The business type identification is mainly extracted, such as business type information such as securities trading, insurance sales, and financial products, which directly reflects the business attributes and service areas of the target financial system. At the same time, the functional module identification is obtained, including specific functional module divisions, dependencies between modules, core functional characteristics, and other information. These system characteristic information is structured and stored for subsequent model matching analysis.

[0085] For example, taking the insurance sales system as an example, the business type identifications obtained include "personal insurance", "group insurance", "insurance claims", etc., and the functional module identifications include identification information of specific functional modules such as "product display", "online insurance", "policy management", and "claims service". For the securities trading system, the business type identifications may include "stock trading", "fund trading", "futures trading", etc., and the functional module identifications include "market inquiry", "trading order", "fund transfer", "account management", etc. These system features provide the basic basis for subsequent rule model matching.

[0086] In step S202 of some embodiments, a complete keyword processing mechanism is established to extract valuable key field features from business review data. First, the business review data is preprocessed, including basic operations such as text segmentation, stop word removal, and standardization. Then, the keywords and phrases in the business review data are identified through natural language processing technology. The keywords may involve specific business terms, functional descriptions, problem phenomena, etc. Then, the keywords are subjected to frequency statistics and importance analysis to screen out the most representative key fields. At the same time, the correlation between these key fields is analyzed to construct a key field network, so as to more comprehensively understand the characteristics of the content of the business review data.

[0087] For example, for business review data such as "slow system response, delayed transaction confirmation, and serious impact on transaction experience", key fields such as "slow response", "transaction confirmation", and "delay" are extracted, and their associations with system performance, transaction functions, etc. are marked. Based on historical data and expert experience, weights are assigned to different key fields, such as 0.8 for performance-related words, 0.6 for function-related words, and 0.4 for experience-related words, to highlight the impact of important features.

[0088] In step S203 of some embodiments, an intelligent model matching mechanism is constructed to select the most suitable target comment rule model from multiple preset candidate comment rule models. First, the candidate comment rule models are preliminarily screened according to system features such as business type identification and function module identification, and a set of models that may be applicable are filtered out. Then, the rule characteristics of these candidate models are analyzed, including dimensions such as rule coverage, applicable scenarios, and historical application effects. The extracted key field features are matched with the rule features of the candidate comment rule models. The matching degree of each candidate model with the current scenario is calculated to determine the target comment rule model that matches the target financial system.

[0089] For example, when the acquired system features indicate that this is a target financial system related to securities trading, and the key field features frequently contain terms related to trading, such as "stock trading", "order placement", and "market inquiry", the matching mechanism will give priority to the comment rule model optimized for securities trading scenarios. The candidate comment rule model with the highest score is selected as the target comment rule model by comprehensively scoring the business field matching, functional coverage, keyword matching, and historical effect scores.

[0090] Through the above steps S201 to S203, by obtaining the system characteristics of the target financial system and the key field characteristics of the business review data, it is possible to more accurately select a matching review rule model and improve the accuracy of model selection; by analyzing the system characteristics and key field characteristics, the errors that may be caused by relying solely on experience to select the model are avoided, and the intelligent model selection is achieved to better meet the personalized needs of different financial systems.

[0091] See also Figure 3 In some embodiments, step S203 may include but is not limited to steps S301 to S304:

[0092] Step S301, matching the system features with the applicable scenarios in the candidate comment rule model to obtain a first matching score.

[0093] Step S302, matching the key field features with the rule items in the candidate comment rule model to obtain a second matching score.

[0094] Step S303: Perform weighted calculation on the first matching score and the second matching score to obtain a target matching score.

[0095] Step S304, sorting the target matching scores of the candidate review rule models, and determining the candidate review rule model with the highest target matching score as the target review rule model.

[0096] In step S301 of some embodiments, an accurate matching evaluation is achieved by establishing a mapping relationship between system features and model applicable scenarios. First, the system features are standardized, including the normalized representation of business type identifiers and functional module identifiers, and then these standardized system features are matched with the applicable scenarios predefined in the candidate comment rule model, and the similarity between feature vectors is calculated to obtain a first matching score. During the matching process, algorithms such as cosine similarity can be used to calculate the distance between feature vectors, and the scoring range is a real number between 0 and 1.

[0097] For example, for a securities trading system, its system features include business type identifiers such as "stock trading" and "fund trading", as well as functional module identifiers such as "trading order" and "fund transfer". When matching with the applicable scenario of a candidate comment rule model, if the model is specifically designed for the securities trading field and covers the relevant functional modules, it may obtain a higher first matching score, such as 0.85; conversely, if the model is designed for other financial fields, it may obtain a lower score, such as 0.3. The matching score based on system features provides an important reference for the subsequent selection of the target comment rule model.

[0098] In step S302 of some embodiments, a refined scoring mechanism is established for the matching of key field features and rule items. The extracted key field features are vectorized to construct feature vectors, and the feature vectors are matched with the rule items in the candidate comment rule model to calculate the degree of match between them. The matching process not only considers the direct matching of keywords, but also considers the semantic relevance of synonyms and near-synonyms, and uses the word vector model to calculate the semantic similarity, and finally obtains the second matching score. For example, if the key field features contain performance-related words such as "transaction delay" and "slow system response", when matching with a candidate comment rule model that specializes in dealing with system performance issues, a higher second matching score, such as 0.9, may be obtained; and when matching with a candidate comment rule model that deals with other types of problems, a lower score, such as 0.4, may be obtained. This semantic-level matching score further improves the accuracy of model selection.

[0099] In step S303 of some embodiments, a scientific weighted calculation method is used to reasonably combine the first matching score and the second matching score. When setting the weight coefficient, it can be based on actual application experience, for example, the weight of the system feature matching is set to 0.6, and the weight of the key field feature matching is set to 0.4. The target matching score comprehensively reflects the applicability of the candidate review rule model in terms of the two dimensions of system features and key field features.

[0100] In step S304 of some embodiments, based on the calculated target matching score, a strict target review rule model selection mechanism is established. First, the target matching scores of all candidate review rule models are sorted in descending order to generate a sorted list, and then the candidate review rule model ranked first, that is, the candidate review rule model with the highest target matching score, is selected as the final target review rule model. If multiple candidate review rule models have the same target matching score, secondary factors such as the model's time efficiency and resource consumption are further considered for screening.

[0101] Through the above steps S301 to S304, by scoring the matching degree of system features and key field features respectively and combining them with weighted calculation, the applicability of candidate comment rule models can be evaluated more comprehensively and objectively; through the final sorting selection method, the selection of the optimal target comment rule model is ensured, thereby improving the accuracy and scientificity of model selection.

[0102] In step S103 of some embodiments, a complete multi-dimensional comment recognition and processing mechanism is established based on the selected target comment rule model. First, the business comment data is preprocessed, including basic processing such as text cleaning, word segmentation, and removal of stop words; then, the comment information is automatically classified through a deep learning algorithm, and the business comment data is accurately divided into corresponding business categories; then, the pre-trained language model is used to identify sensitive words, and content that may involve sensitive information or pose risks is discovered in a timely manner; at the same time, through semantic understanding and text clustering algorithms, hot issues and issues that users generally pay attention to are identified; finally, natural language processing technology is used to extract comment opinions, and key appeals, emotional tendencies, and specific suggestions in the business comment data are extracted.

[0103] For example, for a piece of business review data, "The mobile banking transfer function often fails, and transfers fail many times but funds have been deducted, which seriously affects the user experience. I hope it will be resolved as soon as possible", the following target recognition results can be obtained through multi-dimensional recognition processing: the comment type is a function complaint, involving mobile banking transfer business, the problem type is transaction abnormality, the emotional tendency is strong dissatisfaction, and the core demand is to repair the transfer function and ensure the safety of funds. These target recognition results provide a clear direction for subsequent processing.

[0104] See also Figure 4 In some embodiments, step S103 may include but is not limited to steps S401 to S403:

[0105] Step S401, based on the sensitive word rules in the target review rule model, the business review data is scanned to identify financial sensitive information and obtain the review text to be processed.

[0106] Step S402: extract keywords and calculate feature vectors for the comment text to be processed to obtain comment text features.

[0107] Step S403, performing multi-dimensional comment recognition processing on the comment text to be processed and the comment text features based on the target comment rule model to obtain a target recognition result.

[0108] In step S401 of some embodiments, a complete sensitive information identification mechanism is established in view of the particularity of the financial field. First, the sensitive word rule library predefined in the target comment rule model is loaded. The rule library contains sensitive terms of multiple categories such as account information, transaction data, personal privacy, etc. A multi-pattern matching algorithm is used to perform a full-text scan on the business review data to identify the sensitive information that may be contained therein, and the identified sensitive information is desensitized or marked to generate the comment text to be processed.

[0109] In step S402 of some embodiments, natural language processing technology is used to conduct in-depth analysis on the review text to be processed. First, the review text to be processed is segmented using a word segmentation algorithm to obtain basic morpheme units. Then, keywords in the review text to be processed are extracted based on the algorithm, and multiple feature dimensions such as word frequency, position, and semantics are considered at the same time. The extracted keywords are converted into feature vectors, and the review text information to be processed is mapped to a high-dimensional feature space using technologies such as word embedding to obtain quantifiable review text features.

[0110] For example, for the comment text to be processed "the system responds particularly slowly, often freezes, and affects transaction efficiency", morpheme units such as "system / response / particularly / slow / often / freezes / affect / transaction / efficiency" are obtained through word segmentation, and keywords such as "system response", "freezes", and "transaction efficiency" are extracted, and their feature weights are calculated to generate a comment text feature vector, which fully represents the core features of the comment text to be processed. These comment text features provide an important basis for subsequent multi-dimensional recognition.

[0111] In step S403 of some embodiments, based on the multi-dimensional recognition rules in the target comment rule model, the comment text to be processed is comprehensively analyzed. The comment type is identified according to the comment text features, and the comment text to be processed is divided into different categories such as function evaluation, complaint feedback, and suggestions. Then, sentiment analysis is performed to determine the positive, negative, or neutral features of the comment text to be processed. Then, through the topic cluster analysis, the specific business areas and problem types involved in the comment text to be processed are identified. Finally, combined with the context information, the specific demands and suggestions in the comment text to be processed are extracted. For example, for a certain comment text to be processed and its comment text features, the multi-dimensional comment recognition processing may obtain the following target recognition results: the comment type is "function complaint", the sentiment tendency is "strong dissatisfaction", the subject involved is "system performance problem", the specific demand is "improve system response speed", and the urgency is "high".

[0112] Through the above steps S401 to S403, sensitive information is first identified to ensure the compliance of business review data; then feature extraction and vectorization processing are performed to provide a basis for subsequent multi-dimensional identification; finally, comprehensive identification and analysis are performed based on the target review rule model to achieve accurate identification and classification of business review data.

[0113] See also Figure 5 In some embodiments, step S403 may include but is not limited to steps S501 to S504:

[0114] Step S501: Based on the comment text features, text clustering analysis is performed on the comment text to be processed to obtain the comment category of the comment text to be processed.

[0115] Step S502 , based on the semantic recognition rules in the target comment rule model, contextual semantic analysis is performed on the comment text to be processed to obtain the comment tendency of the comment text to be processed.

[0116] Step S503, calculating the text similarity of different business review data according to the review text features, and when the text similarity is greater than a preset threshold, marking the corresponding business review data as a hot review.

[0117] Step S504, integrating the comment categories, comment tendencies and hot comments to generate target recognition results.

[0118] In step S501 of some embodiments, text clustering technology is used to classify the comment text to be processed. A feature space is constructed based on the comment text features, and each comment text to be processed is represented as a vector in the feature space. A clustering algorithm such as K-means is used to automatically cluster the text vectors according to the distance relationship between them. The clustering results are semantically annotated in combination with business knowledge to obtain clear comment category divisions.

[0119] For example, for a batch of pending review texts of financial systems, cluster analysis may yield the following review categories: functional anomaly (such as "transaction failure", "system freeze"), business consultation (such as "product purchase process", "account opening requirements"), service experience (such as "complex operation", "unfriendly interface"), etc. Each review category has its own specific feature vector center, and the new pending review text can be determined by calculating the distance from these centers to determine the review category it belongs to.

[0120] In step S502 of some embodiments, a review tendency analysis is performed based on natural language processing technology based on deep learning. The semantic recognition rules in the target review rule model are loaded, including sentiment dictionary, semantic rule library, etc., and semantic dependency analysis is performed on the review text to be processed to identify the core viewpoints and sentiment table. Through context relationship analysis, the real review tendency of the review text to be processed is accurately understood to avoid judgment bias caused by relying solely on keywords.

[0121] For example, for the pending comment text "This function is well designed, but the response is too slow, which completely affects the user experience", contextual semantic analysis can identify that: although the function design is positively evaluated, the overall comment tendency is negative, and the main dissatisfaction lies in the system performance. Detailed comment tendency analysis results will be given, including positive factors, negative factors and their respective weights.

[0122] In step S503 of some embodiments, a text similarity calculation model is constructed to identify hot comments, and a word vector model is used to calculate the text similarity between different to-be-processed comment texts, and a cosine similarity algorithm is used for quantitative calculation. A preset threshold for text similarity is set, and when the text similarity between a group of to-be-processed comment texts exceeds the preset threshold, the group of to-be-processed comment texts is marked as a related comment cluster, and then based on factors such as the number of comments and time distribution, it is determined whether the comment cluster constitutes a hot comment.

[0123] For example, if multiple comment texts with similar content appear in a short period of time, such as "Unable to log into the system", "Login page cannot be opened", "Login verification failed", etc., and the text similarity between them exceeds the preset threshold, then this group of business review data will be marked as hot comments, indicating that there may be common problems with the login function and need to be processed as a priority.

[0124] In step S504 of some embodiments, a multi-dimensional comment result integration mechanism is adopted. First, a unified result representation model is established, and information such as comment categories, comment tendencies, and hot comment tags are organized in a predefined format. Weight coefficients are set for recognition results of different dimensions to reflect their importance in the overall evaluation of target recognition results, and a structured target recognition result is generated, which includes complete analysis information of the comment text to be processed.

[0125] For example, the target recognition result of a comment text to be processed may include: comment category = "abnormal function", comment tendency = "strong dissatisfaction", problem type = "system performance", hot comments = "yes", impact range = "widespread", urgency = "high" and other multi-dimensional information. This multi-dimensional structured target recognition result provides sufficient basis for subsequent accurate processing and analysis. The confidence of the target recognition result can also be calculated during the integration process to ensure the reliability of the result.

[0126] Through the above steps S501 to S504, accurate classification of comments is achieved through text clustering analysis, the tendency of comments is accurately grasped through semantic analysis, hot issues are identified through similarity calculation, and finally the multi-dimensional analysis results are integrated to form a comprehensive target recognition result, thereby improving the accuracy and efficiency of comment processing.

[0127] See also Figure 6 In some embodiments, step S504 may include but is not limited to steps S601 to S604:

[0128] Step S601, generating a category label for the business review data according to the review category.

[0129] Among them, the category label is used to characterize the attribution type of the business review data.

[0130] Step S602, classifying the business review data according to the positive or negative tendency of the review and a preset sentiment grading standard, and obtaining the sentiment grade of the business review data.

[0131] Step S603, determining the attention level of the business review data based on the marking status of the hot comment.

[0132] Step S604, obtaining a target recognition result according to the attribution type, the emotion level and the attention level.

[0133] In step S601 of some embodiments, a standardized category label system is established, and the review categories are mapped to a predefined label system based on the characteristics of the business field. The system includes two levels: main labels and sub-labels. The main labels represent the main business directions of the business review data, such as "functional category", "business category", "service category", etc.; the sub-labels further subdivide specific problem types, such as "system performance", "operation experience", "business consultation", etc. This hierarchical label system can accurately reflect the attribution type of the business review data.

[0134] In step S602 of some embodiments, a multi-level sentiment rating system is constructed, and a preliminary division is made according to the positivity and negativity of the review tendency, and then a refined classification is performed in combination with a preset sentiment grading standard. The sentiment grading standard can include 5 levels: extremely positive, positive, neutral, negative, and extremely negative. The final sentiment level is determined by evaluating factors such as the intensity of sentiment words and the degree of modifiers in the business review data.

[0135] For example, for the business review data "the system interface design is very bad, the operation is extremely inconvenient, and it seriously affects work efficiency", by analyzing the sentiment words "bad", "inconvenient" and the degree words "very", "extremely", and "serious", its sentiment level is determined to be extremely negative. At the same time, the detailed basis for the sentiment judgment is recorded, including keywords and their weights, to provide sufficient support for the determination of the sentiment level.

[0136] In step S603 of some embodiments, a scientific attention level classification method is adopted to establish an attention level evaluation model based on the marking status of hot comments, combined with the time density, user scope, and influence degree of business comment data, for example, the attention level is divided into four levels: urgent, high attention, ordinary attention, and general. For business comment data marked as hot comments, the specific attention level is determined according to its characteristic parameters.

[0137] For example, if a large number of similar contents appear in a certain type of business review data within a short period of time, and it involves core business functions, such as "the transfer function cannot be used", it may be classified as an emergency level of concern. The determination of the concern level will take into account multiple quantitative indicators, such as the number of similar comments exceeding the threshold, the number of affected users exceeding the benchmark value, etc., to ensure the scientific nature of the attention level classification.

[0138] In step S604 of some embodiments, a complete target recognition result generation mechanism is constructed, and a unified result representation format is established, which includes information on three core dimensions: attribution type, sentiment level, and attention level. Weight coefficients are set according to the importance of different dimensions, and a comprehensive evaluation score is generated to form a structured target recognition result, providing a detailed reference basis for subsequent processing and analysis.

[0139] Through the above steps S601 to S604, the refined classification and grading of business review data are achieved through standardized processing and hierarchical division of multi-dimensional features, which provides a clear priority basis for subsequent response processing and improves the pertinence and efficiency of review processing.

[0140] In step S104 of some embodiments, by constructing a monitoring response rule system, the target identification results are deeply analyzed and a business response strategy is formulated. The monitoring response rules include configurations in multiple dimensions, including response priority classification standards for different types of comments, response time requirements, processing flow definitions, response method selection standards, etc. According to the importance, urgency, scope of influence and other characteristics of the business review data, the most suitable business response strategy is automatically generated with reference to the monitoring response rules. In the process of strategy formulation, multiple factors such as resource conditions and processing capabilities are considered to ensure the executability of the business response strategy, and at the same time, a dynamic adjustment mechanism is established to optimize and adjust the business response strategy according to the actual implementation situation.

[0141] See also Figure 7 In some embodiments, step S104 may include but is not limited to steps S701 to S703:

[0142] Step S701, determining the processing priority of the business review data according to the attribution type, sentiment level and attention level in the target recognition result.

[0143] Step S702: determining a target response rule from a plurality of preset response rules based on the processing priority.

[0144] Step S703: Generate a business response strategy for the business review data according to the target response rule.

[0145] Among them, the business response strategy includes the processing department, processing time limit and processing method.

[0146] In step S701 of some embodiments, a basic weight is set for the attribution type of the business review data, such as a higher weight for functional issues than for general consulting issues, and then the weight is adjusted in combination with the emotional level. A negative emotional level will increase the processing priority. Finally, the impact of the attention level is considered, and a higher processing priority is given to hot issues or high-attention issues. The final processing priority is obtained through comprehensive calculation of these three dimensions. For example, for a piece of business review data, its attribution type is "Functional-Transaction Abnormality", the emotional level is "Extremely Negative", and the attention level is "Urgent". If the processing priority is divided into the first level, the second level, the third level, and the fourth level, the business review data can be classified as the highest level of the first level for priority processing.

[0147] In step S702 of some embodiments, standard processing procedures for different types of issues are defined in a preset response rule library, including processing steps, response time limits, escalation mechanisms, etc. According to the processing priority of the business review data, the target response rules of the corresponding level are matched. For issues with high processing priority, stricter target response rules will be matched, including shorter response time limits and higher-level processing personnel.

[0148] In step S703 of some embodiments, the processing department, such as the technical department, business department, customer service department, etc., is determined according to the target response rule, and a processing time limit is set, including multiple time nodes such as the first response time, the solution provision time, and the final processing completion time, and a specific processing method is determined, including specific measures such as problem repair, process optimization, and user communication.

[0149] Through the above steps S701 to S703, the processing priority is determined through a comprehensive evaluation of multi-dimensional features, and the corresponding target response rules are selected based on the priority, and finally a specific business response strategy is generated, which realizes the precision and standardization of comment processing and improves response efficiency.

[0150] In step S105 of some embodiments, based on the formulated business response strategy, a complete set of business processing flow triggering mechanism is established. First, according to the priority and urgency of the business response strategy, the startup order of the business processing flow is determined. For high-priority issues, such as serious system failures, important customer complaints, etc., the emergency processing process is immediately triggered, including creating an emergency work order, notifying relevant responsible persons, and starting emergency plans. For routine functional evaluations and suggestions, they are included in the product optimization library, regularly analyzed and evaluated, and included in the iteration plan. A full-process task tracking mechanism is established to ensure that each business review data can be processed in a timely and effective manner through functions such as task status updates, processing progress monitoring, and delay warnings. A variety of response notification channels are provided, including email notifications, SMS reminders, system messages, etc., and appropriate notification methods are selected according to the needs of different roles to ensure that information is transmitted in a timely and accurate manner during the processing process. After the processing is completed, a processing report is generated to record the processing process, solutions, and processing results, and relevant experience and best practices are deposited in the knowledge base to provide reference for the subsequent processing of similar problems. The entire business processing process is built on a high-availability architecture, and the stability and reliability of the business processing process are ensured through technical means such as caching mechanisms, task queues, and failure retries.

[0151] See also Figure 8 , Figure 8It is a schematic diagram of the financial system comment processing method provided by the embodiment of the present application, which is applied to the comment processing platform. First, the comment writing module obtains the business comment data from the target financial system, such as feedback, function evaluation, complaint information, etc., and collects various types of comment data generated by users in the process of using various business systems of the financial system in real time. The rule model module is responsible for managing and applying various comment rule models, including functions such as custom evaluation rules, sensitive word management, keyword management, rule matching analysis, etc., through which personalized rule configuration and model matching for different target financial systems are realized. The comment recognition module realizes multi-dimensional analysis and processing of business comment data, including automatic classification recognition, sensitive word recognition, hot issue recognition, context semantic recognition, text clustering, comment view extraction, etc., through which comprehensive analysis and understanding of comment data are realized. Then, the storage module can adopt a variety of storage solutions, including cache, database, big data, data stream, ES and other storage methods, to achieve efficient storage and rapid retrieval of different types of data. The combined use of these storage solutions ensures high performance and reliability. The list module can provide comment data display and management functions, including comment list, detail view, statistical analysis, metadata management and hot issue display, etc., so that managers can grasp the comment situation in real time and make analytical decisions. The monitoring response module is responsible for formulating and executing response strategies, including monitoring platform access, email monitoring, SMS monitoring, threshold alarms and manual replies, etc., through which timely response and processing of comment issues can be achieved. At the bottom layer of the entire financial system, a one-click access mechanism is used to achieve rapid access to different business systems. This mechanism supports multiple business systems to easily access the comment processing platform through access credentials, and realizes unified management and processing of comment functions.

[0152] See also Fig. 9 , Fig. 9It is a schematic diagram of the financial system comment processing method provided by the embodiment of the present application. The financial system comment processing method provided by the present embodiment adopts a high-availability architecture design. The entire architecture is divided into an API interface layer, a service layer, a processing layer and a storage layer from top to bottom, realizing the full process management from business comment data acquisition to target recognition result response. The API interface layer supports different target financial systems to quickly access the comment processing platform through an open and standardized API interface. The service layer adopts a distributed architecture and deploys multiple functional service nodes. Among them, the comment processing service node is responsible for the preprocessing and preliminary classification of business comment data; the rule matching service node performs the selection and application of the target comment rule model; and the response processing service node generates a business response strategy based on the target recognition result. Each service node supports horizontal expansion and ensures the high availability of the system through load balancing. The service nodes are decoupled through message queues to improve the fault tolerance of the system. The processing layer includes two core modules: hot spot processing and storage processing. The hot spot processing module is responsible for the deep processing of business comment data, including comment classification, sensitive word recognition, contextual semantic analysis, and comment viewpoint extraction. The storage layer can achieve efficient data management through mechanisms such as data sharding and multi-level caching, while providing functions such as data backup and recovery and archiving cleanup. The high availability of the entire architecture is guaranteed by a number of technical means. First, the service layer adopts multi-node deployment, and requests are distributed through a load balancer. The failure of a single node will not affect the overall service; second, the storage layer adopts master-slave replication and data sharding to improve data availability and access performance; finally, through mechanisms such as fuses and downgrades, the normal operation of core functions is guaranteed when the system pressure is too high. For example, when a service node is overloaded, the load balancer will automatically distribute new requests to other nodes; when a storage node fails, the system will automatically switch to the backup node to ensure service continuity. This multi-level fault-tolerant mechanism improves the stability and reliability of the system.

[0153] The financial system comment processing method and device, electronic device and storage medium proposed in the present application obtain business comment data from the target financial system; determine a target comment rule model that matches the target financial system from multiple preset candidate comment rule models of the target financial system; perform multi-dimensional comment recognition processing on the business comment data based on the target comment rule model to obtain a target recognition result; analyze the target recognition result based on preset monitoring response rules to obtain a business response strategy for the business comment data; and trigger the corresponding business processing flow according to the business response strategy.

[0154] This application first obtains the business review data of the target financial system and introduces multiple preset candidate review rule models. It can select the most suitable review rule model according to the business characteristics of different financial systems to improve the targeted processing of reviews. The multi-dimensional recognition method based on the target review rule model avoids the lack of accuracy caused by single-dimensional recognition, and realizes systematic review response management through analysis through preset monitoring response rules to better meet the review processing needs of the financial system. Through the formulation of business response strategies, a rapid response mechanism between review recognition and business processing is established, and the corresponding business processing flow is triggered according to the business response strategy to improve the efficiency of financial system review processing.

[0155] See also Fig.10 The embodiment of the present application also provides a financial system comment processing device, which can implement the above-mentioned financial system comment processing method, and the device includes:

[0156] An acquisition module, used to acquire business review data from a target financial system;

[0157] A matching module, used to determine a target review rule model that matches the target financial system from among a plurality of preset candidate review rule models of the target financial system;

[0158] The recognition module is used to perform multi-dimensional review recognition processing on the business review data based on the target review rule model to obtain the target recognition result;

[0159] A strategy module is used to analyze the target recognition results based on preset monitoring response rules to obtain a business response strategy for the business review data;

[0160] The response module is used to trigger the corresponding business processing flow according to the business response strategy.

[0161] The specific implementation of the financial system review processing device is basically the same as the specific implementation of the above-mentioned financial system review processing method, and will not be repeated here.

[0162] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above financial system comment processing method when executing the computer program. The electronic device can be any smart terminal including a tablet computer, a car computer, etc.

[0163] See also Fig.11 , Fig.11 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0164] The processor 1101 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0165] The memory 1102 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1102 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 1102, and the processor 1101 calls and executes the financial system comment processing method of the embodiment of this application;

[0166] Input / output interface 1103, used to implement information input and output;

[0167] The communication interface 1104 is used to realize the communication interaction between the device and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WI FI, Bluetooth, etc.);

[0168] A bus 1105 that transmits information between various components of the device (e.g., the processor 1101, the memory 1102, the input / output interface 1103, and the communication interface 1104);

[0169] The processor 1101 , the memory 1102 , the input / output interface 1103 and the communication interface 1104 are connected to each other in communication within the device via the bus 1105 .

[0170] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above-mentioned financial system comment processing method is implemented.

[0171] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0172] The financial system comment processing method and device, electronic device and storage medium proposed in the present application obtain business comment data from the target financial system; determine a target comment rule model that matches the target financial system from multiple preset candidate comment rule models of the target financial system; perform multi-dimensional comment recognition processing on the business comment data based on the target comment rule model to obtain a target recognition result; analyze the target recognition result based on preset monitoring response rules to obtain a business response strategy for the business comment data; and trigger the corresponding business processing flow according to the business response strategy.

[0173] This application first obtains the business review data of the target financial system and introduces multiple preset candidate review rule models. It can select the most suitable review rule model according to the business characteristics of different financial systems to improve the targeted processing of reviews. The multi-dimensional recognition method based on the target review rule model avoids the lack of accuracy caused by single-dimensional recognition, and realizes systematic review response management through analysis through preset monitoring response rules to better meet the review processing needs of the financial system. Through the formulation of business response strategies, a rapid response mechanism between review recognition and business processing is established, and the corresponding business processing flow is triggered according to the business response strategy to improve the efficiency of financial system review processing.

[0174] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0175] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0176] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0177] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0178] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0179] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one item (one) of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0180] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0181] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0182] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0183] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0184] The preferred embodiments of the present application are described above with reference to the accompanying drawings, but the scope of the rights of the present application is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present application should be within the scope of the rights of the present application.

Claims

1. A method for processing financial system comments, characterized in that: include: Obtain business review data from the target financial system; Determining a target review rule model matching the target financial system from among a plurality of preset candidate review rule models of the target financial system; Performing multi-dimensional comment recognition processing on the business comment data based on the target comment rule model to obtain a target recognition result; Analyze the target identification result based on preset monitoring response rules to obtain a business response strategy for the business review data; According to the business response strategy, the corresponding business processing flow is triggered.

2. The financial system review processing method according to claim 1, characterized in that: The step of determining a target review rule model matching the target financial system from among a plurality of preset candidate review rule models of the target financial system includes: Acquiring system characteristics of the target financial system, wherein the system characteristics include a business type identifier and a function module identifier; Perform keyword processing on the business review data to extract key field features; According to the system characteristics and the key field characteristics, a target review rule model matching the target financial system is determined from a plurality of preset candidate review rule models.

3. The financial system review processing method according to claim 2, characterized in that: The step of determining a target review rule model matching the target financial system from a plurality of preset candidate review rule models according to the system characteristics and the key field characteristics includes: Matching the system feature with the applicable scenario in the candidate comment rule model to obtain a first matching score; Matching the key field feature with the rule item in the candidate comment rule model to obtain a second matching score; Performing weighted calculation on the first matching score and the second matching score to obtain a target matching score; The target matching scores of the candidate review rule models are sorted, and the candidate review rule model with the highest target matching score is determined as the target review rule model.

4. The financial system review processing method according to claim 1, characterized in that: The performing multi-dimensional comment recognition processing on the business comment data based on the target comment rule model to obtain a target recognition result includes: Based on the sensitive word rules in the target review rule model, the business review data is scanned to identify financial sensitive information and obtain the review text to be processed; Extract keywords and calculate feature vectors of the comment text to be processed to obtain comment text features; Based on the target comment rule model, multi-dimensional comment recognition processing is performed on the comment text to be processed and the comment text features to obtain a target recognition result.

5. The financial system review processing method according to claim 4, characterized in that: The step of performing multi-dimensional comment recognition processing on the comment text to be processed and the comment text features based on the target comment rule model to obtain a target recognition result includes: Based on the comment text features, performing text clustering analysis on the comment text to be processed to obtain comment categories of the comment text to be processed; Based on the semantic recognition rules in the target comment rule model, contextual semantic analysis is performed on the comment text to be processed to obtain the comment tendency of the comment text to be processed; Calculating text similarities of different business review data based on the review text features, and marking the corresponding business review data as hot reviews when the text similarities are greater than a preset threshold; The comment category, the comment tendency and the hot comments are integrated to generate the target recognition result.

6. The financial system review processing method according to claim 5, characterized in that: The step of integrating the comment category, the comment tendency and the hot comments to generate the target recognition result includes: Generating a category label of the business review data according to the review category, wherein the category label is used to characterize the attribution type of the business review data; According to the positive or negative tendency of the review and the preset sentiment grading standard, the business review data is graded to obtain the sentiment grade of the business review data; Determining the attention level of the business review data based on the marking status of the hot review; The target recognition result is obtained according to the attribution type, the emotion level and the attention level.

7. The financial system review processing method according to claim 6, characterized in that: The target identification result is analyzed based on the preset monitoring response rule to obtain a business response strategy for the business review data, including: Determining the processing priority of the business review data according to the attribution type, sentiment level and attention level in the target recognition result; Based on the processing priority, determining a target response rule from a plurality of preset response rules; A business response strategy for the business review data is generated according to the target response rule, wherein the business response strategy includes a processing department, a processing time limit, and a processing method.

8. A financial system comment processing device, characterized in that: include: An acquisition module, used to acquire business review data from a target financial system; A matching module, configured to determine a target review rule model matching the target financial system from among a plurality of preset candidate review rule models of the target financial system; An identification module, used to perform multi-dimensional comment identification processing on the business comment data based on the target comment rule model to obtain a target identification result; A strategy module, used to analyze the target identification result based on a preset monitoring response rule to obtain a business response strategy for the business review data; The response module is used to trigger the corresponding business processing flow according to the business response strategy.

9. An electronic device, characterized in that: include: A memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method for processing financial system comments as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The storage medium stores a program, and the program is executed by a processor to implement the financial system review processing method as described in any one of claims 1 to 7.

Images