Call recording quality inspection method and system based on logic expression

Abstract

The invention discloses a call recording quality inspection method and system based on a logic expression, and relates to the field of call quality inspection, and the method comprises the steps: receiving a logic expression and a call recording text inputted by a user; logic operators in the logic expression are extracted, and operation logic of the logic operators is determined; when it is detected that the customized operator exists in the logic expression, obtaining a detection logic setting of the customized operator in a customized logic database, and determining a customized logic of the customized operator based on the detection logic setting; determining a detection logic tree corresponding to the call recording text; traversing all leaf nodes of the detection logic tree, and determining a node calculation result of the call recording text at the leaf nodes; determining an output result of the detection logic tree at the root node; and according to the logic expression, the call recording text and the output result, generating a recording quality inspection report. By implementing the method, a complex language mode can be identified and processed more accurately, and errors in call recording quality inspection can be avoided to a greater extent.

Description

Technical Field

[0001] The present application relates to the field of call quality inspection, and in particular to a call recording quality inspection method and system based on logical expressions. Background Art

[0002] With the development of the express delivery industry, the number of franchise express delivery outlets has continued to increase. How to quantify the service quality of these outlets has become an important issue. Telephone communication is one of the main channels for customer service. It is particularly important to strictly control the quality of customer service answering calls. In order to ensure the high quality of customer service, companies will conduct quality inspections on customer service call recordings to evaluate and improve the service level of customer service personnel.

[0003] Among the related technologies, the technical means for recording quality inspection mainly rely on speech recognition and natural language processing technologies. These technologies convert call recordings into text, then parse the text to detect whether there are specific keywords or key sentences in the call, and then classify and evaluate them based on predefined rules.

[0004] However, the same keywords have different meanings in different contexts. For example, "you go and give feedback" and "help you give feedback" both contain "you go and give feedback", but may reflect two completely different processing attitudes; the relevant recording quality inspection technology is difficult to accurately identify and process complex language patterns, which may lead to errors in call recording quality inspection. Summary of the invention

[0005] The present application provides a call recording quality inspection method and system based on logical expressions, which are used to more accurately identify and process complex language patterns and avoid errors in call recording quality inspection to a greater extent.

[0006] In a first aspect, the present application provides a call recording quality inspection method based on logical expressions, which is applied to a quality inspection system, and the method includes: receiving a logical expression input by a user and a call recording text to be inspected; extracting the logical operator in the logical expression, and determining the operation logic of the logical operator; when a custom operator is detected in the logical expression, obtaining the detection logic setting of the custom operator in the custom logic database, and determining the custom logic of the custom operator based on the detection logic setting; the custom operator is an operator different from the logical operator; according to the operation logic, custom logic and lexical units of the logical expression, determining the detection logic tree corresponding to the call recording text; traversing all leaf nodes of the detection logic tree, and determining the node calculation results of the call recording text at the leaf nodes; according to the node calculation results of all leaf nodes, determining the output result of the detection logic tree at the root node; generating a recording quality inspection report according to the logical expression, the call recording text and the output result.

[0007] In the above embodiment, the quality inspection system introduces customized operators and expands the expressive power of logical expressions. These customized operators can capture more complex language patterns and timing relationships, so that the quality inspection system can identify more subtle contextual differences; for example, through the before and after operators, the system can accurately locate the sequence of keywords, thereby more accurately judging the context and quality of the conversation; the accuracy and flexibility of call recording quality inspection are improved, complex language patterns can be more accurately identified and processed, and errors in call recording quality inspection can be avoided to a greater extent.

[0008] In combination with some embodiments of the first aspect, in some embodiments, a detection logic tree corresponding to the call recording text is determined based on the operation logic, customized logic and lexical units of the logical expression, specifically including: lexical segmentation of the logical expression to obtain multiple lexical units; determining the grammatical position of the lexical unit based on the operation logic or customized logic of the lexical unit; and generating a detection logic tree corresponding to the call recording text based on the lexical unit and the grammatical position.

[0009] In the above embodiment, the quality inspection system converts complex logical expressions containing customized operators into structured detection logic trees through lexical analysis and syntactic analysis. It can not only handle traditional and, or logic, but also handle more complex customized semantic relationships such as before and after. It can better understand and process complex quality inspection rules, thereby improving the accuracy and efficiency of quality inspection.

[0010] In combination with some embodiments of the first aspect, in some embodiments, traversing all leaf nodes of the detection logic tree to determine the node calculation results of the call recording text at the leaf nodes specifically includes: traversing all leaf nodes of the detection logic tree to determine the node units and node logic of the leaf nodes; generating node calculation tasks corresponding to the leaf nodes based on the node units and node logic; and determining the node calculation results of the leaf nodes based on the call recording text and the node calculation tasks.

[0011] In the above embodiment, the quality inspection system realizes accurate analysis of the call recording text by traversing the leaf nodes of the detection logic tree and generating corresponding computing tasks for each node; in particular, the use of customized operators such as before, after, etc. can accurately detect the temporal relationship and context information in the text, thereby improving the accuracy and reliability of the quality inspection results.

[0012] In combination with some embodiments of the first aspect, in some embodiments, after the step of generating node computing tasks corresponding to leaf nodes according to node units and node logic, the method also includes: determining the computing priority of the node computing tasks corresponding to the leaf nodes according to the node position of the leaf nodes in the detection logic tree; collecting all node computing tasks, and generating a computing task queue according to the computing priority.

[0013] In the above embodiment, the quality inspection system can improve the processing efficiency of recording quality inspection by prioritizing node computing tasks and generating a more efficient computing task queue.

[0014] In combination with some embodiments of the first aspect, in some embodiments, before the step of collecting all node computing tasks and generating a computing task queue according to computing priority, the method also includes: determining the node unit and node logic of each leaf node; when the node units and node logic of two leaf nodes meet preset merging conditions, merging the node computing tasks of the two leaf nodes.

[0015] In the above embodiment, the quality inspection system optimizes the calculation process by merging leaf node calculation tasks that meet preset conditions. Especially for customized operators, the system can intelligently merge related calculation tasks, reduce repeated calculations, and improve the efficiency of quality inspection.

[0016] In combination with some embodiments of the first aspect, in some embodiments, before the step of receiving the logical expression input by the user and the call recording text to be detected, the method also includes: displaying multiple preset lexical units and grammatical logic annotations corresponding to the preset lexical units according to the detection scenario option selected by the user; generating a logical preset formula with the preset lexical unit selected by the user as the target unit; and generating a logical expression according to the user's parameter setting operation on the logical preset formula.

[0017] In the above embodiment, the quality inspection system introduces the element of user interaction, allowing the user to select preset lexical units and customized operators according to specific detection scenarios, and set corresponding parameters, thereby improving the flexibility and customizability of recording quality inspection, which can better adapt to different business needs and complex language environments, simplifies user operations, and improves the ease of use of the system.

[0018] In combination with some embodiments of the first aspect, in some embodiments, after the step of generating a recording quality inspection report based on the logical expression, the call recording text and the output results, the method also includes: determining the customer service terminal corresponding to the call recording text; sending the recording quality inspection report to the customer service terminal in response to the user's quality inspection distribution instruction; in response to the appeal request returned by the customer service terminal, using the call recording text as the review text, determining the review voice file corresponding to the review text; collecting the review feedback entered by the user after confirming the appeal request and the review voice file, and sending the review feedback to the customer service terminal.

[0019] In the above embodiment, the quality inspection system introduces a feedback mechanism for quality inspection results, allowing customer service terminals to make appeals and conduct reviews, ensuring the fairness and accuracy of the quality inspection results, providing data feedback for continuous improvement of the quality inspection system, and improving the accuracy of the system and overall service quality.

[0020] In a second aspect, an embodiment of the present application provides a quality inspection system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the quality inspection system to perform the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer program product comprising instructions, which, when executed on a quality inspection system, enables the quality inspection system to execute the method described in the first aspect and any possible implementation of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions. When the instructions are executed on a quality inspection system, the quality inspection system executes the method described in the first aspect and any possible implementation of the first aspect.

[0023] It is understandable that the quality inspection system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the method provided in the embodiment of the present application. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method, which will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. Due to the adoption of a call recording quality inspection method based on customized operators, a more flexible and accurate detection logic tree can be constructed, and comprehensive analysis can be performed by traversing leaf nodes, which effectively solves the problem of the difficulty in accurately identifying complex language patterns and timing relationships in the existing technology, thereby achieving higher accuracy and more comprehensive call recording quality inspection, and avoiding errors in call recording quality inspection to a greater extent.

[0025] 2. Due to the use of lexical segmentation and grammatical analysis methods for logical expressions, the system can more accurately construct the detection logic tree and accurately process complex expressions containing customized operators, effectively solving the problem of insufficient expression capabilities of complex language patterns and specific business scenarios in existing technologies, and thus achieving more flexible and powerful quality inspection rule definition capabilities.

[0026] 3. Since the system adopts a method of generating logical expressions based on user-selected detection scenarios and preset lexical units, it can provide a more friendly and flexible user interface, while supporting the application of complex customized operators, effectively solving the problem of complex and unintuitive quality inspection rules in the existing technology, and difficulty in adapting to specific business needs, thereby realizing a more customizable and adaptable quality inspection system. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a call recording quality inspection method based on logical expressions in an embodiment of the present application; Figure 2 is another flowchart of the call recording quality inspection method based on logical expressions in an embodiment of the present application; Figure 3 It is a schematic diagram of the structure of a physical device of the quality inspection system in the embodiment of the present application. DETAILED DESCRIPTION

[0028] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification of the present application, the singular expressions "one", "a kind of", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more of the listed items.

[0029] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0030] For ease of understanding, the application scenarios of the embodiments of the present application are introduced below.

[0031] A large express delivery company, STQ Express, was facing a rising rate of customer complaints. An investigation found that many complaints were caused by the inappropriate words and deeds of customer service staff when handling customer inquiries and problems. The company needed an effective quality inspection system to monitor and improve the quality of customer service calls, but was faced with a huge amount of data from hundreds of thousands of calls every day. The traditional manual spot check method was inefficient and prone to deviations, and could not detect and correct problems in a timely manner.

[0032] In the related art, a simple automatic detection system based on keyword matching can be used to achieve basic monitoring of customer service call quality. This method can detect preset specific words or phrases to identify calls that may have problems. The following describes a scenario using the call recording quality inspection method based on logical expressions in the related art.

[0033] STQ Express tried to use a simple automatic quality inspection system based on keyword matching. The system is able to detect preset negative words such as "impossible" and "no". However, this approach has limitations. For example, when a customer service representative said, "It is impossible for this express to be delivered in more than three days," the system mistakenly marked it as a negative response. On the other hand, some euphemistic but actually impolite expressions, such as "I will be angry if you say that again," could not be recognized by the system. In addition, the system could not understand the context, resulting in a large number of misjudgments and missed judgments, and the reliability of the quality inspection results was low.

[0034] The call recording quality inspection method based on logical expressions in the embodiment of the present application realizes in-depth semantic analysis of the call content by introducing complex logical operators and customized operators, which can not only handle simple keyword matching, but also understand the logical relationship and context between words. The following introduces the scenario of using the call recording quality inspection method based on logical expressions in the present application.

[0035] First, we explain several types of custom operators. It should be noted that the operation logic of custom operators is not limited, that is, their corresponding naming and usage should not be limited to the following examples. Custom operators can be: Before means "before", for example, "{(A: Hello, STQ) before (A: How can I help you?)}", if "Hello, STQ" exists within 20 characters (the value can be set as needed) before "How can I help you?" in the text to be inspected, it is a hit; a hit means a successful match, and the corresponding output value can be set according to the forward or reverse match, such as a Boolean value of true or false, to implement the basic matching logic of recording quality inspection.

[0036] After means "afterwards", for example, "{(A: you want my work number) after (A: it's useless)}", if "it's useless" exists within 20 characters (the value can be set as needed) after "you want my work number" in the text to be checked, it is a hit; after can be used to connect multiple keywords, for example, "{(A: hello) after (A: STQ) after (A: at your service)}", if "hello", "STQ", and "at your service" appear consecutively in the text to be checked, it is a hit; Near means "nearby", for example, "{(A: your phone number) near (A: can't get through)" is a hit if "can't get through" exists within 20 characters before and after "your phone number" in the text to be checked; Except means "filter", for example, "{(A: I go) except (A: I go and take a look)}", if "I go" exists in the text to be checked but it is in the case of "I go and take a look", it will not be a hit, but it will be a hit in other cases.

[0037] The following is a scenario: After STQ Express Company adopted the call recording quality inspection method based on logical expressions in this application, the quality inspection effect was improved, and the new system was able to handle complex language patterns and contexts. For example, by setting the logical expression "{(A: you go to give feedback) except (A: help you to give feedback)}", the system can correctly distinguish between the expressions "you go to give feedback" (negative) and "help you to give feedback" (positive). At the same time, using the "before" and "after" operators, the system can capture more inappropriate expressions, such as "{(A: you are right) before (A: but)}" can identify potential negative responses such as "you are right, but..." The new system can also detect the context of keywords through the "near" operator, greatly reducing the misjudgment rate.

[0038] It can be seen that the call recording quality inspection method based on logical expressions in the embodiment of the present application can not only realize automated call recording quality inspection, but also effectively solve the problems of high misjudgment rate and poor context understanding ability in traditional methods, thereby realizing more accurate call recording quality inspection.

[0039] For ease of understanding, the following describes the process of the method provided by this implementation in combination with the above scenario. Figure 1 , which is a flow chart of a call recording quality inspection method based on logical expressions in an embodiment of the present application.

[0040] S101, receiving a logical expression input by a user and a call recording text to be detected.

[0041] Among them, the logical expression represents the rules for detecting call recordings, which are composed of logical operators and detection keywords; the call recording text refers to the text content converted from the call recording; and the user refers to the operator who uses the quality inspection system.

[0042] Before the quality inspection system conducts quality inspection on call recordings, it first needs to obtain the detection rules and the content to be detected. Specifically, the quality inspection system receives the logical expression manually entered or uploaded by the user through the user interface, and the logical expression defines the specific detection rules. At the same time, the quality inspection system receives the call recording text provided by the user, and the call recording text is converted from the call recording through speech recognition technology.

[0043] In some embodiments, receiving user input can be achieved in a variety of ways: optionally, providing a graphical user interface, including a logic expression input box and a text upload area; the user enters the logic expression in the interface and uploads the call recording text file; the system receives and stores the data entered by the user. Optionally, an API interface is provided to allow an external system to pass in the logic expression and call recording text; receive data sent by the external system through the API; verify whether the received data format is correct and perform preprocessing. It is understandable that other methods can also be used to receive user input, such as reading through a configuration file, etc., which are not limited here.

[0044] S102: extract the logical operator in the logical expression, and determine the operation logic of the logical operator.

[0045] The logical operators represent keywords used to connect and combine detection conditions, such as "and", "or", etc.; the operation logic refers to the specific calculation rules and priorities of each logical operator.

[0046] After receiving the logical expression, the quality inspection system needs to parse it to prepare for subsequent inspection. Specifically, the quality inspection system scans the entire logical expression and identifies all the logical operators contained therein. Then, the system determines the specific operation rules of each identified logical operator based on the predefined rule base, including its mode of action and operation priority.

[0047] S103: When it is detected that a custom operator exists in the logical expression, a detection logic setting of the custom operator in a custom logic database is obtained, and a custom logic of the custom operator is determined based on the detection logic setting.

[0048] Custom operators refer to custom operators used to implement specific detection logic, including "before", "after", "near", and "except". Detection logic settings refer to the set of rules that define the specific behavior of custom operators, such as detection range and matching conditions.

[0049] When processing logical expressions, the quality inspection system needs to identify and process these special custom operators. Specifically, the system identifies custom operators such as "before", "after", "near" or "except" in logical expressions. The system then queries the custom logic database to obtain the specific detection logic settings for each custom operator. For example, for the "before" and "after" operators, the system will obtain their 20-character detection range settings; for the "near" operator, the system will obtain its detection range settings for the 20 characters before and after. Finally, based on these settings, the system generates the specific operation logic for each custom operator, that is, the custom logic. The custom logic includes how to locate keywords in the text, how to determine the positional relationship between keywords, etc.

[0050] S104: Determine a detection logic tree corresponding to the call recording text according to the operation logic, the custom logic, and the lexical unit of the logic expression.

[0051] The lexical unit refers to the smallest grammatical unit in the logic expression, including standard logic operators, customized operators and detection keywords, etc. The detection logic tree refers to the logic expression represented by a tree structure, which includes standard logic nodes and customized logic nodes.

[0052] After the quality inspection system completes the parsing of logical expressions and the processing of custom operators, it needs to build a tree structure that can correctly represent the complex detection logic. Specifically, the system first decomposes the logical expression into a sequence of lexical units, including logical operators, custom operators, and keywords. Then, the system organizes these lexical units into a tree structure based on the priority and associativity of the operators. For standard logical operators, the system applies predefined operation logic; for custom operators, the system applies previously determined custom logic. For example, for the "before" operator, the system creates a special tree node that contains the logic for checking the before-and-after relationship of two keywords; for the "except" operator, the system creates a special node that contains the exclusion logic.

[0053] For example, for the logical expression: { (B: Find the leader) after (A: Cannot transfer) or (B: find the leader) after (A: unable to transfer) or (B: find a leader) after (A: useless) or (A: The leader handles it the same way) or{ (A: I help) after (B: register) and (A: The leader also handles it this way)} }, the quality inspection system can identify the logical operators "and" and "or", the custom operator "after", and the lexical units include "and", "or", "after", "B: find the leader", "A: cannot transfer"... Correspondingly, the operation logic of "and" and "or" is "and" and "or", which follow the conventional logical operation rules; and the custom logic of "after" is "afterwards", and the calculation priority of "after" is higher than that of "and" and "or", indicating the detection of the previous and next contexts; for example, (B: find a leader) after (A: cannot transfer) means that after customer B says "find a leader", customer service A replies "cannot transfer".

[0054] The quality inspection system performs parsing based on the above operation logic, customized logic and lexical units based on specific operation rules, where (B: find the leader) after (A: cannot be transferred), (B: find the leader) after (A: cannot be transferred), B: find the leader) after (A: useless), (A: the leader handles the same) are each a leaf node, connected by "or"; for leaf nodes with compound expressions, for example: { (A: I help) after (B: register) and (A: The leader also handles it this way) }, where (A: I help) after (B: register) and (A: the leader also handles it in this way) are more subdivided and are also leaf nodes, connected by "and"; according to the gradual subdivision of the levels in the braces, the quality inspection system can expand the logical expression, and the components have a tree structure with multiple branches, that is, the detection logic tree.

[0055] S105: traverse all leaf nodes of the detection logic tree to determine the node calculation results of the call recording text at the leaf nodes.

[0056] The leaf node represents the node unit at the end of the branch in the detection logic tree, corresponding to a specific detection condition. The node calculation result refers to the Boolean value obtained after applying the leaf node detection condition to the call recording text. Traversal refers to the process of visiting all leaf nodes in the tree in a specific order.

[0057] After the quality inspection system has built the detection logic tree, it needs to test the actual call recording text. Specifically, the system uses a depth-first search algorithm to traverse the detection logic tree and find all leaf nodes. For each leaf node, the system will parse the detection conditions it contains and then apply these conditions to the call recording text. The system will assign a Boolean value to each leaf node as the node calculation result based on whether the detection condition is met.

[0058] For example, for the logical expression in step S104, the quality inspection system will traverse all leaf nodes of the corresponding detection logic tree, perform text matching of the call recording text, and obtain the corresponding node calculation results. For example, for the call recording text, the quality inspection system calculates that the leaf node (B: find the leader) after (A: cannot be transferred) is true, that is, after customer B said "find the leader", customer service A answered "cannot be transferred"; correspondingly, other leaf nodes, such as (B: find the leader) after (A: cannot be transferred), B: find the leader) after (A: useless), (A: the leader handles the same), etc., can be calculated as false, true, true, etc.; false or true is the node calculation result of these leaf nodes.

[0059] S106: Determine the output result of the detection logic tree at the root node according to the node calculation results of all leaf nodes.

[0060] The root node represents the top node of the detection logic tree, which represents the final result of the entire logic expression. The output result refers to the evaluation result of the entire logic expression on the call recording text, which is a Boolean value.

[0061] After completing the calculation of all leaf nodes, the quality inspection system needs to obtain the overall test results based on the structure of the logic tree. Specifically, the system starts from the leaf nodes and propagates the calculation results upward along the tree structure. For each non-leaf node, the system will calculate the result of the node based on its corresponding operation logic or customized logic, combined with the results of its child nodes. This process will continue until it reaches the root node, and the calculation result of the root node is the output result of the entire detection logic tree.

[0062] For example, following the example in step S105 above, after calculating the node calculation results of each leaf node as true or false, the quality inspection system will calculate the root node, that is, the "or" that connects all the leaf nodes together. The corresponding quality inspection situation can be obtained through the calculation results of this node. For example, for the calculation of true, false, true, true, the root node of "or" is calculated as "true", "true" hits three times, and the corresponding index is 3, which means that there is non-compliance in the recording text, and the severity level is level 3, and the output result is true (3). Specific output results, such as severity levels, can be adjusted accordingly according to the weight settings of the user at each leaf node.

[0063] In some embodiments, the result calculation of the detection logic tree can be implemented in a variety of ways: optionally, use a post-order traversal algorithm to traverse the entire tree; for each non-leaf node, select a corresponding calculation method according to its operator type; store the calculation result in the node until the result of the root node is obtained. Optionally, convert the tree structure into a suffix expression; use a stack structure to simulate the calculation process; traverse the suffix expression, perform corresponding operations, and finally the top element of the stack is the result. It is understandable that other methods can also be used to implement the result calculation of the detection logic tree, such as using the visitor pattern, etc., which are not limited here.

[0064] S107. Generate a recording quality inspection report based on the logical expression, the call recording text and the output result.

[0065] Among them, the recording quality inspection report is a detailed document on the quality assessment of call recordings.

[0066] After completing all the detection steps, the quality inspection system needs to organize the results into a form that is easy to understand and analyze. Specifically, the system will use the original logical expression, the text content of the detected call recording, and the final output results to generate a detailed quality inspection report. This report includes a description of the detection rules, relevant fragments of the text content, an explanation of the detection results, and possible improvement suggestions.

[0067] In some embodiments, the generation of the recording quality inspection report can be achieved in a variety of ways: optionally, designing a report template, including placeholders for various results; collecting various data during the detection process; filling the collected data into the report template to generate the final report document. Optionally, establish a report object model, including various report elements; dynamically build a report object tree based on the detection results; use a report rendering engine to convert the object tree into a visual report format. It is understandable that other methods can also be used to achieve the generation of recording quality inspection reports, such as using natural language generation technology, which is not limited here.

[0068] In the above embodiment, the quality inspection system can handle a variety of complex logical relationships, such as "before", "after", "near" and "except", etc. In practical applications, these logical relationships can be used in combination to create more sophisticated and flexible detection rules to adapt to various complex call scenarios.

[0069] The following is a supplement to the scenario of this embodiment.

[0070] STQ Express subsequently introduced a machine learning algorithm to automatically generate and adjust logical expressions based on historical data. For example, the system found that a short time interval between "hello" and "goodbye" often meant that the customer service ended the call early, so it automatically generated an expression to detect this situation. In addition, the company has developed a user-friendly type interface that allows quality inspectors to easily create and modify complex logical expressions without programming knowledge, improving the accuracy and efficiency of quality inspection.

[0071] After combining the above scenarios, the following is a more detailed description of the process of the method provided by this implementation. Figure 2 , which is another flow chart of the call recording quality inspection method based on logical expressions in an embodiment of the present application.

[0072] S201. Display multiple preset lexical units and grammatical logic annotations corresponding to the preset lexical units according to the detection scenario option selected by the user.

[0073] Among them, the detection scenario options represent different types of call quality inspection scenarios, such as customer complaint handling, sales talk inspection, etc. Preset lexical units refer to commonly used logical expression components pre-defined for specific detection scenarios, such as keywords, operators, etc. Syntactic logic annotations are used to explain the usage and function of preset lexical units.

[0074] When the user starts to create a new detection rule, the quality inspection system first needs to provide relevant auxiliary information. Specifically, the system first displays a series of predefined detection scenario options for the user to choose. When the user selects a specific scenario, the system retrieves a set of preset lexical units related to the scenario from the database. The system then displays these preset lexical units on the interface, along with grammatical logic annotations for each unit. These annotations explain in detail the meaning, usage, and role of each lexical unit in logical expressions, helping users understand and use these units correctly.

[0075] In some embodiments, the display of information related to the detection scene can be achieved in a variety of ways: optionally, a drop-down menu is designed to list all available detection scene options; when the user selects a scene, an AJAX request is triggered to obtain relevant preset lexical units and annotations; a table is generated using dynamic HTML to display the preset lexical units and their corresponding grammatical logic annotations. Optionally, a scene-lexical unit mapping data structure is created to preload relevant information for all scenes; the user's scene selection event is monitored; and according to the selected scene, relevant lexical units and annotations are filtered and displayed from the preloaded data. It is understandable that other methods can also be used to achieve the display of information related to the detection scene, such as using a visualization component library to build an interactive display interface, etc., which is not limited here.

[0076] S202: Generate a preset logical formula with the preset lexical unit selected by the user as the target unit.

[0077] The target unit refers to a specific unit selected by the user from a list of preset lexical units, and the logical preset expression refers to a preliminary logical expression framework automatically generated based on the selected preset lexical unit.

[0078] After the user selects a specific preset lexical unit, the quality inspection system needs to provide the user with a preliminary logical expression structure. Specifically, the system first identifies the preset lexical unit selected by the user. Then, the system constructs a basic logical expression framework based on these selected units and predefined templates or rules. The framework contains the selected lexical units and organizes them according to a reasonable grammatical structure. The generated logical preset provides a template for the user, on which the user can make further modifications and improvements.

[0079] In some embodiments, the generation of the logical preset formula can be achieved in a variety of ways: optionally, a corresponding expression template is defined for each preset lexical unit; when the user selects multiple lexical units, these templates are spliced ​​using predefined combination rules; during the splicing process, necessary logical operators are inserted to form a complete logical preset formula. Optionally, a lexical unit combination tree is constructed, and the nodes of the tree represent different lexical unit combinations; according to the lexical unit selected by the user, the best matching path is found in the combination tree; and the corresponding logical preset formula is generated along the matching path. It is understandable that other methods can also be used to achieve the generation of logical preset formulas, such as using natural language processing technology to understand user intent and generate expressions, etc., which are not limited here.

[0080] S203: Generate a logical expression according to the user's parameter setting operation on the logical preset formula.

[0081] The parameter setting operation refers to the modification and customization of the adjustable part of the logic preset formula by the user.

[0082] After providing the logic preset, the quality inspection system needs to allow users to make personalized adjustments to meet specific quality inspection needs. Specifically, the system first displays the logic preset and identifies the part where parameters can be set. Users can modify keywords, adjust operators, set thresholds for detection values, etc. The system captures each parameter setting operation of the user in real time and applies these settings to the logic preset. Finally, the system generates a complete executable logic expression based on the modified preset.

[0083] In some embodiments, the generation of logical expressions can be achieved in a variety of ways: optionally, an interactive editor is designed to allow users to make modifications directly on the logical preset formula; real-time verification of whether the user's modifications comply with grammatical rules; when all modifications are completed and verified, the modified preset formula is converted to a standard logical expression format. Optionally, a series of parameter setting forms are provided, corresponding to the adjustable parts in the logical preset formula; all parameters entered by the user in the form are collected; using a template engine, the collected parameters are filled into the logical preset formula to generate the final logical expression. It is understandable that other methods can also be used to achieve the generation of logical expressions, such as using a visual drag-and-drop interface to build expressions, etc., which are not limited here.

[0084] It should be noted that each custom operator will set a corresponding grammar rule, which can verify the legality of the logical expression. The structure of a conventional logical expression is {() or () or ()}, that is, there are parentheses in the curly brackets, and the parentheses are detection keywords. In addition: a. Only operators such as and, or, before, after, near, and except can be used between parentheses; b. After between parentheses cannot appear after other keywords; c. Before between parentheses cannot appear after other keywords; d. Except between parentheses must be the first conjunction and there can only be one; f. The characters and text in parentheses must be connected using ':'.

[0085] S204: receiving a logical expression input by a user and a call recording text to be detected.

[0086] Referring to step S101 , the quality inspection system determines the logical expression and the call recording text.

[0087] S205: extract the logical operator in the logical expression, and determine the operation logic of the logical operator.

[0088] Referring to step S102 , the quality inspection system determines the operation logic of the logical operator.

[0089] S206: When it is detected that a custom operator exists in the logical expression, a detection logic setting of the custom operator in a custom logic database is obtained, and a custom logic of the custom operator is determined based on the detection logic setting.

[0090] Referring to step S103 , the quality inspection system determines the custom logic of the custom operator.

[0091] S207: Perform lexical segmentation on the logical expression to obtain multiple lexical units.

[0092] Lexical segmentation refers to the process of breaking down a logical expression into its smallest grammatical units. Lexical units represent the basic components of a logical expression, such as keywords and operators.

[0093] After obtaining the complete logical expression defined by the user, the quality inspection system needs to structure it for subsequent analysis. Specifically, the system first reads the entire logical expression string. Then, the system scans the expression from left to right according to the predefined lexical rules and identifies various types of lexical units. In this process, the system divides the continuous character sequence into meaningful minimum units, such as identifying "and" as a logical operator unit and "(A: Hello)" as a composite unit containing dialogue roles and keywords. Finally, the system generates a sequence containing all the recognized lexical units, which retains the structural information of the original expression.

[0094] S208: Determine the grammatical position of the lexical unit according to the operation logic or customized logic of the lexical unit.

[0095] The operation logic refers to the calculation rules and priorities of standard logical operators ("and", "or", and "not"). Customized logic refers to the special processing rules of custom operators ("before", "after", "near", and "except"). The grammatical position refers to the position and role of a lexical unit in the structure of a logical expression.

[0096] After completing the lexical segmentation, the quality inspection system needs to further determine the role of each lexical unit in the logical expression. Specifically, the system first traverses all lexical units. For each unit, the system applies the corresponding operation logic or custom logic according to its type (such as standard operator or custom operator). The system analyzes the relationship between the unit and other surrounding units, considering the priority and associativity of the operator. Through this process, the system determines the exact position of each lexical unit in the expression structure, such as as an operator or operand of a sub-expression.

[0097] In some embodiments, the determination of the grammatical position of a lexical unit can be achieved in a variety of ways: optionally, a priority table is constructed, which contains the priority information of all standard and custom operators; a stack structure is used to simulate the parsing process of an expression, and the push or pop operation is determined according to the priority; and the relative position and function of each lexical unit is determined according to the result of the stack operation. Optionally, a recursive descent parser is implemented to define the parsing function of each type of lexical unit; starting from the lexical unit with the highest priority, the entire expression is recursively parsed; during the parsing process, the grammatical position of each lexical unit is determined according to the context. It is understandable that other methods can also be used to determine the grammatical position of a lexical unit, such as using a syntax analyzer automatic generation tool, which is not limited here.

[0098] S209: Generate a detection logic tree corresponding to the call recording text according to the lexical unit and the grammatical position.

[0099] The detection logic tree refers to a logical expression represented by a tree structure, where nodes represent operators or operands and edges represent the relationship between them.

[0100] After determining the grammatical position of the lexical unit, the quality inspection system needs to build a structured representation that can effectively determine the subsequent quality inspection process. Specifically, the system first creates an empty tree structure. Then, according to the grammatical position and logical relationship of the lexical unit, the system inserts the lexical unit into the appropriate position in the tree one by one. For operators, the system creates corresponding internal nodes; for operands (such as keywords), the system creates leaf nodes. In this process, the system correctly handles complex nested relationships based on the priority and associativity of the operators. Finally, the system generates a complete detection logic tree that reflects the structure and semantics of the original logical expression.

[0101] S210, traverse all leaf nodes of the detection logic tree, and determine the node units and node logic of the leaf nodes.

[0102] The node unit refers to the specific content contained in the leaf node, such as keywords or specific detection rules.

[0103] After building the detection logic tree, the quality inspection system needs to understand the characteristics of each basic detection unit in detail. Specifically, the system first uses a depth-first search algorithm to traverse the entire detection logic tree and find all leaf nodes. For each leaf node, the system extracts the node unit information contained therein, such as specific keywords or detection rules. At the same time, the system determines the specific evaluation logic applicable to the node based on the type and context of the node, such as exact matching, fuzzy matching, or complex conditional judgment. This process enables the system to accurately understand the content and processing methods of each basic detection unit.

[0104] In some embodiments, the determination of leaf node information can be achieved in a variety of ways: optionally, a recursive tree traversal function is implemented that can identify leaf nodes; for each leaf node, its internal data structure is parsed to extract node unit information; according to the node type and the context of the parent node, appropriate node logic is selected from a predefined set of logic rules. Optionally, the detection logic tree is converted into a linear list of leaf nodes; the list is traversed, and an information extractor is applied to each node to obtain the node unit; a strategy pattern is used to dynamically select and apply the corresponding node logic according to the characteristics of the node. It is understandable that other methods can also be used to achieve the determination of leaf node information, such as using a reflection mechanism to dynamically call a node processing method, etc., which are not limited here.

[0105] S211. Generate node computing tasks corresponding to leaf nodes according to node units and node logic.

[0106] Among them, the node calculation task represents the specific operation steps for evaluating a specific leaf node.

[0107] After determining the characteristics of the leaf nodes, the quality inspection system needs to create a specific execution plan for each node. Specifically, the system first creates an independent computing task object for each leaf node. In this object, the system will include node unit information (such as keywords to be detected) and node logic (such as matching methods). At the same time, the system will determine the specific execution steps of the task based on the characteristics of the node, such as how to search for keywords in the text, how to apply matching rules, etc. This process converts abstract node information into executable specific tasks, preparing for subsequent actual inspections.

[0108] In some embodiments, the quality inspection system determines the computation priority of the node computation task corresponding to the leaf node according to the node position of the leaf node in the detection logic tree; collects all the node computation tasks, and generates a computation task queue according to the computation priority.

[0109] The leaf node position indicates the level and order of the node in the detection logic tree. The computation priority refers to the order in which the node computation tasks are executed. The node computation task is used to indicate the specific operation of evaluating a specific leaf node. The computation task queue refers to the sequence of tasks to be executed sorted by priority.

[0110] After generating the computational tasks for all leaf nodes, the quality inspection system needs to determine the execution order of these tasks to improve efficiency. Specifically, the system first analyzes the structure of the detection logic tree and determines the position of each leaf node in the tree. Based on the node position, the system assigns a priority value to each node computation task, and nodes closer to the root of the tree receive a higher priority. Then, the system collects all generated node computation tasks and sorts them according to the determined priority to form an ordered computational task queue. This queue will guide the system to execute computational tasks in the optimal order.

[0111] In some embodiments, the generation of the computing task queue can be achieved in a variety of ways: optionally, a tree traversal algorithm is implemented to calculate the depth of each leaf node and its order in the same level; a priority calculation function is designed to calculate the priority value according to the node depth and order; a priority queue data structure is used to insert all computing tasks into the queue according to priority. Optionally, a unique number is assigned to each node of the detection logic tree, and the numbering method reflects the node position; a task-priority mapping table is created to determine the task priority according to the node number; a quick sort algorithm is used to sort all tasks according to priority to generate an ordered task list. It is understandable that other methods can also be used to achieve the generation of the computing task queue, such as using a multi-level queue to manage tasks of different priorities, etc., which are not limited here.

[0112] In some embodiments, the quality inspection system determines the node unit and node logic of each leaf node; when the node units and node logic of two leaf nodes meet a preset merging condition, the node computing tasks of the two leaf nodes are merged.

[0113] The preset merging condition refers to a standard predefined by the system for determining whether two nodes can be merged. Merging means combining two independent node computing tasks into a more efficient task.

[0114] When the quality inspection system generates the computing tasks for each leaf node, it will determine whether some tasks are repeated or can be optimized. Specifically, the system first checks the node unit and node logic of each leaf node to obtain its specific detection content and evaluation method. Then, the system compares these features between different nodes to determine whether the preset merging conditions are met. The merging conditions may include similar detection content, compatible evaluation logic, etc. If two nodes are found to meet the merging conditions, the system will merge their computing tasks into a new, more efficient task to reduce repeated calculations and improve overall efficiency.

[0115] In some embodiments, the merging of node computing tasks can be achieved in a variety of ways: optionally, a task feature comparator is implemented that can analyze and compare the node units and node logic of different tasks; a set of merging rules is defined to clearly indicate which feature combinations can trigger the merge; all task pairs are traversed, and the comparator and rules are used to determine whether they can be merged, and if so, a new merged task is created. Optionally, a task feature index is constructed to group similar tasks; for each group of similar tasks, a heuristic algorithm is applied to find the optimal merging solution; the merge operation is performed, the task list is updated, and the original merged task is deleted. It can be understood that other methods can also be used to achieve the merging of node computing tasks, such as using machine learning algorithms to automatically discover mergeable task patterns, etc., which are not limited here.

[0116] S212: Determine the node calculation result of the leaf node according to the call recording text and the node calculation task.

[0117] After the quality inspection system has prepared all the necessary information and tasks, it needs to evaluate the actual call content. Specifically, the system first loads the call recording text that needs to be tested. Then, the system traverses all node calculation tasks. For each task, the system performs corresponding detection operations in the call recording text, including keyword search, pattern matching, context analysis, etc., depending on the definition of the task. After executing each task, the system determines whether the task meets the requirements based on the preset judgment criteria, and records the result (true or false) as the calculation result of the leaf node. This process applies abstract detection rules to specific call content to generate basic evaluation results.

[0118] In some embodiments, the determination of node calculation results can be achieved in a variety of ways: Optionally, create a text processing engine that can efficiently perform search and matching operations in large texts; for each node calculation task, call the text processing engine to execute the corresponding detection logic; according to the return value of the detection logic, combined with predefined judgment criteria, determine and record the calculation results of the node. Optionally, implement a task executor that can process multiple node calculation tasks in parallel; submit all tasks to the executor, and use multi-threading technology to process multiple tasks simultaneously; collect the execution results of all tasks and map them back to the corresponding leaf nodes. It can be understood that other methods can also be used to determine the node calculation results, such as using streaming processing technology to evaluate long texts in real time, etc., which are not limited here.

[0119] S213: Determine the output result of the detection logic tree at the root node according to the node calculation results of all leaf nodes.

[0120] Referring to step S106 , the quality inspection system determines the output result.

[0121] S214: Generate a recording quality inspection report based on the logical expression, the call recording text and the output result.

[0122] Referring to step S107, the quality inspection system will generate a recording quality inspection report.

[0123] In some embodiments, the quality inspection system determines the customer service terminal corresponding to the call recording text; in response to the user's quality inspection distribution instruction, sends the recording quality inspection report to the customer service terminal; in response to the appeal request returned by the customer service terminal, uses the call recording text as the review text, and determines the review voice file corresponding to the review text; collects the review feedback entered by the user after confirming the appeal request and the review voice file, and sends the review feedback to the customer service terminal.

[0124] Among them, the customer service terminal refers to the device or system (a mobile phone, computer, etc. with a designated APP or mini-program installed) used by the customer service personnel who receive the quality inspection report. The quality inspection distribution instruction refers to the command issued by the user to send the quality inspection report to the relevant customer service. The appeal request refers to the customer service’s formal application to object to the quality inspection results. The review text refers to the call recording text that needs to be re-checked. The review voice file is the original audio recording corresponding to the review text. The review feedback indicates the user’s final handling opinion on the appeal request.

[0125] After the quality inspection system completes the generation of the quality inspection report, it needs to handle the distribution of the report and possible appeals. Specifically, the system first determines the customer service terminal responsible for the call based on the relevant information of the call recording text. After the user simply confirms the quality inspection report and issues a quality inspection distribution instruction, the system will send the generated quality inspection report to the corresponding customer service terminal. If the customer service objections to the quality inspection results and submits an appeal request, the system will mark the original call recording text as a text that needs to be reviewed. Then, the system will find and extract the original voice file corresponding to the text as the basis for the review. Finally, the system will collect the handling opinions given by the user after confirming the appeal request and reviewing the voice file, and send these feedbacks back to the customer service terminal to complete the entire appeal processing process.

[0126] In some embodiments, the distribution of quality inspection reports and the handling of complaints can be achieved in a variety of ways: optionally, a customer service-call mapping database is established to quickly locate the responsible customer service according to the call identifier; a report transmission module is implemented to ensure that the quality inspection report is accurately delivered to the designated customer service terminal; a complaint processing workflow is designed, including steps such as complaint reception, text tagging, voice extraction, and feedback collection. Optionally, a centralized quality inspection management platform is developed to integrate report distribution and complaint processing functions; a real-time notification mechanism is implemented to promptly notify relevant personnel at each processing stage; a complete data tracking system is built to record the entire life cycle of each quality inspection report, including distribution, complaint, and review. It is understandable that other methods can also be used to achieve the distribution of quality inspection reports and complaint processing, such as using blockchain technology to ensure the transparency and immutability of the process, which is not limited here.

[0127] In the embodiments of the present application, since a quality inspection method based on logical expressions is introduced and multiple custom operators such as "before", "after", "near" and "except" are supported, complex semantic detection rules can be constructed, which effectively solves the limitations of traditional quality inspection methods in dealing with complex language patterns and contextual relationships, thereby achieving more accurate and intelligent call quality assessment. It not only improves the accuracy of quality inspection and reduces misjudgments and missed judgments, but also provides flexible options, allowing quality inspectors to customize detection rules according to different business needs and call scenarios. By supporting the dynamic generation and optimization of logical expressions, it is possible to continuously improve the quality inspection effect and effectively detect customer service quality.

[0128] The quality inspection system in the embodiment of the present invention is described below from the perspective of hardware processing. Figure 3 , which is a schematic diagram of a physical device structure of a quality inspection system in an embodiment of the present application.

[0129] It should be noted that Figure 3 The structure of the quality inspection system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0130] like Figure 3As shown, the quality inspection system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method described in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0131] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0132] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0133] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0134] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0135] Specifically, the quality inspection system of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the call recording quality inspection method based on logical expressions provided in the above embodiment is implemented.

[0136] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the quality inspection system described in the above embodiment; or may exist independently without being assembled into the quality inspection system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the quality inspection system, the quality inspection system implements the call recording quality inspection method based on logical expressions provided in the above embodiment.

[0137] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0138] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0139] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A call recording quality inspection method based on logical expressions, characterized in that: Applied to a quality inspection system, the method comprises: Receive the logical expression input by the user and the call recording text to be detected; Extracting the logical operator in the logical expression and determining the operation logic of the logical operator; When detecting that a custom operator exists in the logical expression, obtaining a detection logic setting of the custom operator in a custom logic database, and determining a custom logic of the custom operator based on the detection logic setting; the custom operator is an operator different from the logical operator; Determine a detection logic tree corresponding to the call recording text according to the operation logic, the custom logic and the lexical unit of the logical expression; Traversing all leaf nodes of the detection logic tree, and determining node calculation results of the call recording text at the leaf nodes; Determine the output result of the detection logic tree at the root node according to the node calculation results of all the leaf nodes; Generate a recording quality inspection report based on the logical expression, the call recording text and the output result.

2. The method according to claim 1, characterized in that Determining the detection logic tree corresponding to the call recording text according to the operation logic, the custom logic and the lexical unit of the logical expression specifically includes: Performing lexical segmentation on the logical expression to obtain a plurality of lexical units; Determining the grammatical position of the lexical unit according to the operation logic or customized logic of the lexical unit; A detection logic tree corresponding to the call recording text is generated according to the lexical unit and the grammatical position.

3. The method according to claim 2, characterized in that The traversing all leaf nodes of the detection logic tree to determine the node calculation result of the call recording text at the leaf node specifically includes: Traversing all leaf nodes of the detection logic tree to determine the node units and node logic of the leaf nodes; Generate a node computing task corresponding to the leaf node according to the node unit and the node logic; The node calculation result of the leaf node is determined according to the call recording text and the node calculation task.

4. The method according to claim 3, characterized in that After the step of generating a node computing task corresponding to the leaf node according to the node unit and the node logic, the method further comprises: Determining the calculation priority of the node calculation task corresponding to the leaf node according to the node position of the leaf node in the detection logic tree; All the node computing tasks are collected, and a computing task queue is generated according to the computing priority.

5. The method according to claim 4, characterized in that Before the step of collecting all the node computing tasks and generating a computing task queue according to the computing priority, the method further includes: Determine the node unit and node logic of each leaf node; When the node units and the node logics of two leaf nodes meet a preset merging condition, the node computing tasks of the two leaf nodes are merged.

6. The method according to claim 1 or 2, characterized in that: Before the step of receiving the logical expression input by the user and the call recording text to be detected, the method further includes: Displaying a plurality of preset lexical units and grammatical logic annotations corresponding to the preset lexical units according to the detection scenario option selected by the user; Taking the preset lexical unit selected by the user as the target unit, generating a preset logical formula; A logical expression is generated according to the user's parameter setting operation on the logical preset formula.

7. The method according to claim 1, characterized in that After the step of generating a recording quality inspection report according to the logical expression, the call recording text and the output result, the method further includes: Determine the customer service terminal corresponding to the call recording text; In response to the quality inspection distribution instruction of the user, sending the recording quality inspection report to the customer service terminal; In response to the appeal request returned by the customer service terminal, taking the call recording text as the review text, determining a review voice file corresponding to the review text; The review feedback input by the user after confirming the appeal request and the review voice file is collected, and the review feedback is sent to the customer service terminal.

8. A quality inspection system, characterized in that: The quality inspection system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the quality inspection system to execute the method described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instruction is executed on the quality inspection system, the quality inspection system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product runs on a quality inspection system, the quality inspection system is enabled to execute the method according to any one of claims 1 to 7.

Images