Rule-driven automatic decision-making method and system

Abstract

The invention discloses an automatic decision-making method and system based on rule driving, and relates to the technical field of business decision-making automation, and the method comprises the steps: recording original decision-making data in a real decision-making scene, and extracting causal elements in the original decision-making data; creating a rule base based on causal elements of the original decision data; adding a factor element matching algorithm for the rule base, querying a rule graph corresponding to the input data based on the factor element matching algorithm, and outputting the rule graph; manual intervention is introduced to optimize the rule base, and it is ensured that each rule graph is put into a production environment after the accuracy of each rule graph reaches the standard; and the production environment outputs a final decision based on the optimized rule base. The decision-making efficiency can be improved, manual intervention is reduced through automatic processing, and the decision-making speed is increased; the interpretability and transparency are improved, the decision-making process is traceable, and audit and interpretation are facilitated; and the method is flexible and configurable, supports dynamic regulation and expansion of rules, and adapts to business change requirements.

Description

Technical Field

[0001] The present invention relates to the technical field of decision automation, and in particular to a rule-driven automated decision-making method and system. Background Art

[0002] With the rapid development of information technology, the amount of data has exploded. Traditional manual decision-making methods can no longer meet the needs of efficiency and accuracy. Although there are some automated decision-making systems in the existing technology, they mostly rely on complex machine learning models. Although these models have powerful data fitting capabilities, they are insufficient in interpretability, maintainability and rapid response. Especially when faced with scenarios that require clear decision logic and auditability, rule-based methods are more applicable. Summary of the invention

[0003] The present invention provides a rule-driven automated decision-making method, comprising: Step 1: Record the original decision data in the real decision-making scenario and extract the causal elements; Step 2: Create a rule base based on the causal elements of the original decision data; Step 3. Add the element matching algorithm to the rule base, query the rule graph corresponding to the input data based on the element matching algorithm and output it; Step 4: Introduce manual intervention to tune the rule base and ensure that each rule graph meets the accuracy standards before being put into production. Step 5: The production environment outputs the final decision based on the optimized rule base.

[0004] The above-mentioned rule-driven automated decision-making method records the original decision data in the real decision-making scenario and extracts the causal elements therein, which specifically includes the following sub-steps: Collect original decision data in real decision-making scenarios; Store the collected raw decision data into a structured database; Extract the cause and effect elements of the original decision data in the data table.

[0005] As described above, a rule-driven automated decision-making method, in which a rule base is created based on the causal elements of the original decision data, specifically includes the following sub-steps: Conduct statistical analysis on each cause element and its value to determine the rules for obtaining each effect element; Create a rule graph based on the rules for obtaining the fruit element; Set the result element that the rule graph points to and store it in the rule base.

[0006] The above-mentioned rule-driven automated decision-making method, in which each cause element and its value are statistically analyzed to determine the acquisition rule of each result element, specifically includes the following sub-steps: The extracted factor elements are labeled with the sequence number of the decision scenario; Group all the cause elements using the value of the result element as the grouping condition; Count the value range of each factor element in the same group; Compare and analyze the logical relationships between elements in different decision-making scenarios within the same group.

[0007] As described above, a rule-driven automated decision-making method is described, wherein a rule graph corresponding to input data is queried and output based on an element matching algorithm, specifically including the following sub-steps: Extract the factor elements in the input data; The rule graph containing the input data elements is queried from the rule base.

[0008] As described above, a rule-driven automated decision-making method, in which manual intervention is introduced to tune the rule base, specifically includes the following sub-steps: Prepare a test environment and ensure that all test data in the test environment are positive samples; Execute the rule base in the test environment and calculate the accuracy of each rule graph; The nodes in the rule graph that do not meet the identification accuracy standards and need to be improved will be handed over to professionals for modification; The modified rule base is retested until the accuracy of each rule map meets the requirements.

[0009] The present invention also provides an automated decision-making system based on rule-drivenness, comprising: an original decision data processing module, a rule base creation module, a rule graph query module, a rule base testing module, and a final decision return module; The decision data processing module is used to record the original decision data in the real decision scenario and extract the causal elements therein; A rule base creation module, used to create a rule base based on the causal elements of the original decision data; The rule graph query module is used to add a factor element matching algorithm to the rule base, query the rule graph corresponding to the input data based on the factor element matching algorithm, and output it; The rule base testing module is used to introduce manual intervention to tune the rule base and ensure that each rule map meets the accuracy standards before being put into production. The final decision return module is used to output the final decision using the tuned rule base.

[0010] The rule-driven automated decision-making system as described above, wherein the rule base creation module includes a factor element statistical analysis submodule and a rule graph generation submodule; The cause element statistical analysis submodule is used to perform statistical analysis on each cause element and its value and determine the acquisition rules of each effect element; The rule graph generation submodule is used to generate a rule graph according to the acquisition rules of the result elements.

[0011] A rule-driven automated decision-making system as described above, wherein the rule base test module includes a test environment building submodule, a test result verification submodule, and a rule base improvement submodule; The test environment building submodule is used to build the test environment of the rule base and prepare a positive sample test set; The test result verification submodule is used to calculate the accuracy of each rule map and verify whether the accuracy of each rule map meets the standard; The rule base improvement submodule is used to identify nodes that need to be improved in the rule graph that do not meet the accuracy standards and hand them over to professionals for modification.

[0012] The beneficial effects achieved by the present invention are as follows: improving decision-making efficiency, reducing human intervention through automated processing, and accelerating decision-making; enhancing decision-making accuracy, reducing human judgment errors through a decision-making process based on clear rules; improving explainability and transparency, making the decision-making process traceable and easy to audit and explain; being flexible and configurable, supporting dynamic adjustment and expansion of rules, and adapting to changing business needs; reducing maintenance costs, and simplifying system maintenance work through a rule management interface. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0014] Figure 1 This is a flow chart of a rule-driven automated decision-making method provided in Example 1 of the present application; Figure 2 This is a schematic diagram of a rule-driven automated decision-making system provided in Example 2 of the present application. DETAILED DESCRIPTION

[0015] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0016] Embodiment 1 like Figure 1 As shown, the first embodiment of the present application provides a rule-driven automated decision-making method, including: Step S10: Recording the original decision data in the real decision scenario and extracting the causal elements therein; The purpose of extracting causal elements is to simplify the original decision data and facilitate the subsequent rule creation operation. Specifically: Step S11: collecting original decision data in real decision scenarios; Use a variety of methods to collect original decision-making data, such as meeting minutes, questionnaires, data monitoring, etc., to ensure the accuracy and completeness of the data and avoid missing key information.

[0017] Step S12: storing the collected original decision data in a structured database; To facilitate the extraction of causal elements, each data table only stores business data generated in one decision scenario.

[0018] Step S13: extracting the cause element and effect element of the original decision data in the data table; The data table contains fields for various business data and a decision result. The field values ​​are extracted and stored from the original decision data. The decision result is the result element, and the non-empty business data is the cause element. For example, the current table contains business data fields such as turnover, customer flow, business date, and special offer, and the field values ​​are 2000, 500, null, and null respectively; the decision result is to start activity A, so starting activity A is a result element, and the cause elements that lead to this result should be turnover and customer flow.

[0019] Step S20: Creating a rule base based on the causal elements of the original decision data, specifically including the following sub-steps: Step S21: Statistically analyze each cause element and its value to determine the acquisition rule of each effect element; The rules for obtaining a result element include the logical relationship between its corresponding cause elements and the value range of each cause element, specifically: Step S211: the extracted factor elements are labeled with the sequence number of the decision scenario; This step is to distinguish the factors in different decision-making scenarios and provide a data basis for subsequent comparative analysis.

[0020] Step S212: grouping all the cause elements using the value of the effect element as a grouping condition; This step is to ensure that a rule graph ultimately points to only one decision result.

[0021] Step S213: Count the value range of each factor element in the same group; The value range statistical function is expressed as: , where the count() function is used to count the number of data in the collection. , , , They are small sets randomly intercepted from the input parameter set X, is the statistical resolution, is the minimum value in the input parameter set X, is the maximum value in the input parameter set X, if This means that the data in the input parameter set is continuous, so the output ,like This means that the data in the input set is not continuous. .

[0022] By inputting the value set of a factor element in the same group as an input parameter into the value range statistical function, the value range of the factor element can be returned.

[0023] Step S214: Compare and analyze the logical relationships between the elements in different decision-making scenarios within the same group; Logical relationships include "and", "or", and "not". If there are cause elements A, B, and C in each decision scenario in the same group, it means that the three cause elements A, B, and C must exist at the same time, which is an "and" relationship; if there are cause elements C, D, etc. in addition to A, B, and C in each decision scenario, then the relationship between C, D, etc. and the three cause elements A, B, and C should be "or"; if there is a value in the cause element that is not within its own value range, then it is necessary to add a "not" symbol in front of this cause element. By comparing and analyzing the cause elements and their indicators in the same group and different decision scenarios, the logical relationship between the cause elements can be obtained. The logical relationship and value range of the cause elements together constitute a rule for obtaining a result element.

[0024] Step S22: creating a rule graph according to the acquisition rule of the fruit element; In the rule graph, each causal element is a separate node with a value range marked on the node. For causal elements with a logical relationship of "and", the nodes are connected to form a main line; for causal elements with a logical relationship of "or", the nodes will cut out a separate branch line at the end of the main line; for causal elements with a "not" symbol, their value range is reversed and re-marked, eventually forming a complete rule graph.

[0025] Step S23: Set the result element pointed to by the rule graph and store it in the rule base.

[0026] Step S30: adding a factor element matching algorithm to the rule base, querying a rule graph corresponding to the input data based on the factor element matching algorithm and outputting it; The workflow of the element matching algorithm is: Step S31: extracting the factor element in the input data; The input data can be a manually imported business data table or real-time monitoring data of business data. The method of extracting cause elements is the same as that of the original decision data. All non-empty business fields are extracted from the data as cause elements. If it is a batch extraction, the serial number of the input data is needed to identify and distinguish it.

[0027] Step S32: querying the rule graph containing the input data element from the rule base; Take all the cause elements in an input data as a whole query condition to ensure that the queried rule graph contains all the cause elements in the input data.

[0028] Step S40: Introduce manual intervention to tune the rule base to ensure that the accuracy of each rule graph meets the standard before it is put into production environment; Run the created rule base in a simulation environment to verify whether it meets business requirements and accuracy standards. Specifically: Step S41: prepare a test environment and ensure that all test data in the test environment are positive samples; Because the applicability and accuracy of the entire process need to be tested, it is necessary to ensure that each input can go through the entire rule graph and have an accurate decision result.

[0029] Step S42: executing the rule base in the test environment and calculating the accuracy of each rule graph; After the test data is input, the cause element matching algorithm of the rule base will first query the rule graph that matches the input data, and then start to detect whether the cause elements in the input data meet the rules set in the rule graph. The detection method is divided into two steps. The first is to determine whether the cause elements of each node on the main line exist, and the second is to determine whether the value of the cause element corresponding to each node is within the value range. When the results of both steps are true, the result element pointed to by the rule graph is output, that is, the decision result. When one of the two steps is false, -1 is returned. Then the calculation formula of the accuracy of the rule graph is expressed as: ,in is the accuracy of the jth regular graph, Representation Rule Graph The i-th input data processed, It is a regular graph detection function, used to return a regular graph for The decision results, For input data The expected decision result, i takes values ​​from 1 to m, and m is the rule graph The total number of test data processed; ,in Representation Rule Graph The node set on the main line, For input data The set of elements in represents the value of the kth factor element in the P set, It represents the value range of the node corresponding to the k-th element in the P set. k ranges from 1 to L, where L is the number of elements in the P set. When true, returns the rule graph The result element pointed to, otherwise -1 is returned.

[0030] Step S43: identifying nodes in the rule graph that do not meet the accuracy standards and needing improvement, and handing them over to professionals for modification; The accuracy of the rule graph must reach 100% to be considered up to standard. For rule graphs that do not meet the standard, it is necessary to identify the points that need to be improved and then hand them over to personnel who are responsible for such business for modification. The identification function of the improved node is expressed as: , where n is the node to be identified, Q is the identification data set of n nodes, and q is the data item in the data set Q. is the value range of n node calibration, when When true, it returns 0, indicating that node n is not a node that needs to be improved. Otherwise, it returns 1, indicating that node n needs to be improved.

[0031] Input each node in the substandard rule graph into the improved node identification function, and record the nodes with a return value of 1. The identification data set is the test data processed by the substandard rule graph. Enter the value set of the factor element corresponding to the node. Finally, mark the recorded nodes in red in the rule graph and hand them over to professionals for modification.

[0032] Step S44: retesting the modified rule base until the accuracy of each rule map meets the standard; Repeat the process of testing and modifying until the accuracy of each rule map reaches 100%.

[0033] Step S50: the production environment outputs the final decision based on the tuned rule base; Filter out the information with a return value of -1, and output the remaining return values ​​to the user as the final decision. You can also set the automatic activation of decision behavior, that is, activate the preset decision behavior according to the returned decision result code.

[0034] Embodiment 2 like Figure 2 As shown, the second embodiment of the present application provides an automated decision-making system based on rule-drivenness, including: an original decision data processing module, a rule base creation module, a rule graph query module, a rule base testing module, and a final decision return module; (1) Decision data processing module, which is used to record the original decision data in the real decision scenario and extract the causal elements therein; it includes: decision data collection submodule, decision data storage submodule, and causal element extraction submodule; 1. Decision data collection submodule, used to collect original decision data in real decision scenarios; Use a variety of methods to collect original decision-making data, such as meeting minutes, questionnaires, data monitoring, etc., to ensure the accuracy and completeness of the data and avoid missing key information.

[0035] 2. Decision data storage submodule, used to store the collected original decision data into a structured database; To facilitate the extraction of causal elements, each data table only stores business data generated in one decision scenario.

[0036] 3. Cause and effect element extraction submodule, used to extract the cause and effect elements of the original decision data in the data table; The data table contains fields for various business data and a decision result. The field values ​​are extracted and stored from the original decision data. The decision result is the result element, and the non-empty business data is the cause element. For example, the current table contains business data fields such as turnover, customer flow, business date, and special offer, and the field values ​​are 2000, 500, null, and null respectively; the decision result is to start activity A, so starting activity A is a result element, and the cause elements that lead to this result should be turnover and customer flow.

[0037] (2) A rule base creation module, which is used to create a rule base based on the causal elements of the original decision data; it includes: a causal element statistical analysis submodule and a rule graph generation submodule; 1. The cause element statistical analysis submodule is used to perform statistical analysis on each cause element and its value and determine the acquisition rules of each effect element; The rules for obtaining a result element include the logical relationship between its corresponding cause elements and the value range of each cause element, specifically: I. The extracted factor elements are labeled with the sequence number of the decision scenario; This step is to distinguish the factors in different decision-making scenarios and provide a data basis for subsequent comparative analysis.

[0038] II. Group all the cause elements using the value of the result element as the grouping condition; This step is to ensure that a rule graph ultimately points to only one decision result.

[0039] III. Count the value range of each factor element in the same group; The value range statistical function is expressed as: , where the count() function is used to count the number of data in the collection. , , , They are small sets randomly intercepted from the input parameter set X, is the statistical resolution, is the minimum value in the input parameter set X, is the maximum value in the input parameter set X, if This means that the data in the input parameter set is continuous, so the output ,like This means that the data in the input set is not continuous. .

[0040] By inputting the value set of a factor element in the same group as an input parameter into the value range statistical function, the value range of the factor element can be returned.

[0041] IV. Compare and analyze the logical relationships between the elements in different decision-making scenarios within the same group; Logical relationships include "and", "or", and "not". If there are cause elements A, B, and C in each decision scenario in the same group, it means that the three cause elements A, B, and C must exist at the same time, which is an "and" relationship; if there are cause elements C, D, etc. in addition to A, B, and C in each decision scenario, then the relationship between C, D, etc. and the three cause elements A, B, and C should be "or"; if there is a value in the cause element that is not within its own value range, then it is necessary to add a "not" symbol in front of this cause element. By comparing and analyzing the cause elements and their indicators in the same group and different decision scenarios, the logical relationship between the cause elements can be obtained. The logical relationship and value range of the cause elements together constitute a rule for obtaining a result element.

[0042] 2. A rule graph generation submodule, used to generate a rule graph according to the acquisition rules of the fruit element; In the rule graph, each cause element is a separate node, and the value range is marked on the node. The nodes of the cause elements with a logical relationship of "and" are connected to form a main line; the nodes of the cause elements with a logical relationship of "or" will cut a branch line at the end of the main line; the cause elements with the "not" symbol will have their value range reversed and re-marked, and finally form a complete rule graph. Set the result element pointed to by the generated rule graph and store it in the rule base.

[0043] (3) A rule graph query module, which is used to add a factor element matching algorithm to the rule base, query the rule graph corresponding to the input data based on the factor element matching algorithm, and output it; The workflow of the element matching algorithm is: I. Extract the factor elements in the input data; The input data can be a manually imported business data table or real-time monitoring data of business data. The method of extracting cause elements is the same as that of the original decision data. All non-empty business fields are extracted from the data as cause elements. If it is a batch extraction, the serial number of the input data is needed to identify and distinguish it.

[0044] II. Query the rule graph containing the input data element from the rule base; Take all the cause elements in an input data as a whole query condition to ensure that the queried rule graph contains all the cause elements in the input data.

[0045] (4) Rule base test module, which is used to introduce manual intervention to tune the rule base and ensure that each rule map meets the accuracy standard before being put into production; it includes: test environment construction submodule, test result verification submodule, and rule base improvement submodule; 1. The test environment construction submodule is used to build the test environment of the rule library and prepare a positive sample test set; Because we need to test the applicability and accuracy of the entire process, we must ensure that each input can go through the entire rule graph and have an accurate decision result, so all the test sets used are positive samples.

[0046] 2. Test result verification submodule, used to calculate the accuracy of each rule map and verify whether the accuracy of each rule map meets the standard; After the test data is input, the cause element matching algorithm of the rule base will first query the rule graph that matches the input data, and then start to detect whether the cause elements in the input data meet the rules set in the rule graph. The detection method is divided into two steps. The first is to determine whether the cause elements of each node on the main line exist, and the second is to determine whether the value of the cause element corresponding to each node is within the value range. When the results of both steps are true, the result element pointed to by the rule graph is output, that is, the decision result. When one of the two steps is false, -1 is returned. Then the calculation formula of the accuracy of the rule graph is expressed as: ,in is the accuracy of the jth regular graph, Representation Rule Graph The i-th input data processed, It is a regular graph detection function, used to return a regular graph for The decision results, For input data The expected decision result, i takes values ​​from 1 to m, and m is the rule graph The total number of test data processed; ,in Representation Rule Graph The node set on the main line, For input data The set of elements in represents the value of the kth factor element in the P set, It represents the value range of the node corresponding to the k-th element in the P set. k ranges from 1 to L, where L is the number of elements in the P set. When true, returns the rule graph The result element pointed to, otherwise -1 is returned.

[0047] 3. The rule base improvement submodule is used to identify nodes that need to be improved in the rule graph that do not meet the accuracy standards and hand them over to professionals for modification; The accuracy of the rule graph must reach 100% to be considered up to standard. For rule graphs that do not meet the standard, it is necessary to identify the points that need to be improved and then hand them over to personnel who are responsible for such business for modification. The identification function of the improved node is expressed as: , where n is the node to be identified, Q is the identification data set of n nodes, and q is the data item in the data set Q. is the value range of n node calibration, when When true, it returns 0, indicating that node n is not a node that needs to be improved. Otherwise, it returns 1, indicating that node n needs to be improved.

[0048] Input each node in the substandard rule graph into the improved node identification function, and record the nodes with a return value of 1. The identification data set is the test data processed by the substandard rule graph. Enter the value set of the factor element corresponding to the node. Finally, mark the recorded nodes in red in the rule graph and hand them over to professionals for modification.

[0049] The modified rule base needs to be retested and verified until the accuracy of each rule map meets the requirements.

[0050] (5) Final decision return module, which is used to output the final decision using the tuned rule base; Filter out the information with a return value of -1, and output the remaining return values ​​to the user as the final decision. You can also set the automatic activation of decision behavior, that is, activate the preset decision behavior according to the returned decision result code.

[0051] Corresponding to the above embodiment, an embodiment of the present invention provides a computer storage medium, including: at least one memory and at least one processor; The memory is used to store one or more program instructions; A processor is used to run one or more program instructions to execute a rule-driven automated decision-making method.

[0052] Corresponding to the above-mentioned embodiments, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer storage medium contains one or more program instructions, and the one or more program instructions are used by a processor to execute a rule-driven automated decision-making method.

[0053] The embodiments disclosed in the present invention provide a computer-readable storage medium, in which computer program instructions are stored. When the computer program instructions are executed on a computer, the computer executes the above-mentioned rule-driven automated decision-making method.

[0054] In the embodiment of the present invention, the processor may be an integrated circuit chip having the ability to process signals. The processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0055] The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present invention can be directly embodied as a hardware decoding processor for execution, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The processor reads the information in the storage medium and completes the steps of the above method in combination with its hardware.

[0056] The storage medium may be a memory, which may be, for example, a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memory.

[0057] Among them, the non-volatile memory can be a read-only memory (ROM), a programmable ROM (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory.

[0058] The volatile memory may be a random access memory (RAM) which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDRSDRAM), enhanced synchronous DRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM bus DRAM (DRRAM).

[0059] The storage media described in the embodiments of the present invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0060] Those skilled in the art will appreciate that in one or more of the above examples, the functions described in the present invention can be implemented using a combination of hardware and software. When software is used, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any media that facilitates the transmission of computer programs from one place to another. Storage media can be any available media that can be accessed by general or special-purpose computers.

[0061] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A rule-driven automated decision-making method, characterized in that: include: Step 1: Record the original decision data in the real decision-making scenario and extract the causal elements; Step 2: Create a rule base based on the causal elements of the original decision data; Step 3. Add the element matching algorithm to the rule base, query the rule graph corresponding to the input data based on the element matching algorithm and output it; Step 4: Introduce manual intervention to tune the rule base and ensure that each rule graph meets the accuracy standards before being put into production. Step 5: The production environment outputs the final decision based on the optimized rule base.

2. The rule-driven automated decision-making method according to claim 1, characterized in that: Record the original decision data in the real decision scenario and extract the causal elements, which includes the following sub-steps: Collect original decision data in real decision-making scenarios; Store the collected raw decision data into a structured database; Extract the cause and effect elements of the original decision data in the data table.

3. The rule-driven automated decision-making method according to claim 1, characterized in that: Creating a rule base based on the causal elements of the original decision data includes the following sub-steps: Conduct statistical analysis on each cause element and its value to determine the rules for obtaining each effect element; Create a rule graph based on the rules for obtaining the fruit element; Set the result element that the rule graph points to and store it in the rule base.

4. The rule-driven automated decision-making method according to claim 3, characterized in that: Perform statistical analysis on each cause element and its value to determine the acquisition rules for each effect element, which specifically includes the following sub-steps: The extracted factor elements are labeled with the sequence number of the decision scenario; Group all the cause elements using the value of the result element as the grouping condition; Count the value range of each factor element in the same group; Compare and analyze the logical relationships between elements in different decision-making scenarios within the same group.

5. The rule-driven automated decision-making method according to claim 1, characterized in that: Based on the element matching algorithm, the rule graph corresponding to the input data is queried and output, which specifically includes the following sub-steps: Extract the factor elements in the input data; The rule graph containing the input data elements is queried from the rule base.

6. The rule-driven automated decision-making method according to claim 1, characterized in that: Manual intervention is introduced to tune the rule base, which includes the following sub-steps: Prepare a test environment and ensure that all test data in the test environment are positive samples; Execute the rule base in the test environment and calculate the accuracy of each rule graph; The nodes in the rule graph that do not meet the identification accuracy standards and need to be improved will be handed over to professionals for modification; The modified rule base is retested until the accuracy of each rule map meets the requirements.

7. A rule-driven automated decision-making system, characterized in that: include: Original decision data processing module, rule base creation module, rule graph query module, rule base testing module, and final decision return module; The decision data processing module is used to record the original decision data in the real decision scenario and extract the causal elements therein; A rule base creation module, used to create a rule base based on the causal elements of the original decision data; The rule graph query module is used to add a factor element matching algorithm to the rule base, query the rule graph corresponding to the input data based on the factor element matching algorithm, and output it; The rule base testing module is used to introduce manual intervention to tune the rule base and ensure that each rule map meets the accuracy standards before being put into production. The final decision return module is used to output the final decision using the tuned rule base.

8. The rule-driven automated decision-making system according to claim 7, characterized in that: Rule base creation module, including element statistical analysis submodule and rule graph generation submodule; The cause element statistical analysis submodule is used to perform statistical analysis on each cause element and its value and determine the acquisition rules of each effect element; The rule graph generation submodule is used to generate a rule graph according to the acquisition rules of the result elements.

9. The rule-driven automated decision-making system according to claim 7, characterized in that: Rule base test module, including test environment construction submodule, test result verification submodule, and rule base improvement submodule; The test environment building submodule is used to build the test environment of the rule base and prepare a positive sample test set; The test result verification submodule is used to calculate the accuracy of each rule map and verify whether the accuracy of each rule map meets the standard; The rule base improvement submodule is used to identify nodes that need to be improved in the rule graph that do not meet the accuracy standards and hand them over to professionals for modification.

10. A computer storage medium, characterized in that: include: at least one memory and at least one processor; A memory for storing one or more program instructions; A processor, configured to run one or more program instructions to execute a rule-driven automated decision-making method as described in any one of claims 1 to 6.

Images