Conflict detection method and device, computer equipment, readable storage medium and program product
By using a decision tree detection method in the interactive response strategy, and combining weight and correlation data for node splitting and pruning, the accuracy problem of conflict detection in the interactive response strategy is solved, and efficient and accurate conflict detection is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-10
- Publication Date
- 2026-04-03
AI Technical Summary
In existing technologies, conflict detection between interactive response strategies relies on human experience, resulting in low detection efficiency and low accuracy.
By acquiring interactive response strategy data, a pre-generated interactive response decision tree is used to detect strategy conflicts. Node splitting and pruning are performed by combining weight data and correlation data to generate an accurate decision tree for conflict detection.
It improves the accuracy of interactive response strategy conflict detection, reduces errors caused by manual processing, and ensures the reliability and efficiency of detection results.
Smart Images
Figure CN121786535A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of computer technology, and in particular to a collision detection method, apparatus, computer device, computer-readable storage medium, and computer program product. Background Technology
[0002] With the development of computer technology, interactive response (IVR) technology has emerged, enabling call distribution, voice navigation, and business process guidance. However, as business scenarios become increasingly complex and the number of interactive response strategies continues to increase, conflicts between these strategies are becoming increasingly prominent.
[0003] Many related technologies rely on manual experience for review, but manual experience review suffers from low detection efficiency and low accuracy. Summary of the Invention
[0004] Therefore, it is necessary to provide a collision detection method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve the accuracy of collision detection in response to the above-mentioned technical problems.
[0005] Firstly, this application provides a conflict detection method, including:
[0006] Acquire the interactive response strategy data to be detected. The interactive response strategy data to be detected includes the triggering conditions to be detected or the interactive response strategy data to be detected includes the triggering conditions to be detected and the detection results corresponding to the triggering conditions to be detected.
[0007] Input the triggering conditions to be detected into the pre-generated interactive response decision tree to obtain the interactive strategy results;
[0008] Based on the interactive strategy results, determine whether there is a conflict in the interactive response strategy data to be tested, or based on the test results and the interactive strategy results, determine whether there is a conflict in the interactive response strategy data to be tested.
[0009] In one embodiment, before inputting the triggering condition to be detected into a pre-generated interactive response decision tree to obtain the interactive policy result, the conflict detection method includes: acquiring effective interactive response policy data, which includes effective policy triggering conditions and effective policy results corresponding to the effective policy triggering conditions; determining the business type corresponding to the effective interactive response policy data and the condition type corresponding to the effective policy triggering conditions, and determining the weight data corresponding to the effective policy triggering conditions based on the business type and condition type; determining the correlation data between each effective policy triggering condition and adjacent effective policy triggering conditions; and splitting the nodes of the interactive response decision tree based on the weight data and correlation data corresponding to the effective policy triggering conditions.
[0010] In one embodiment, after splitting the nodes of the interactive response decision tree, the method further includes: determining the service precision corresponding to the service type based on the service type corresponding to the effective interactive response policy data; adjusting the pruning threshold based on the service precision; and using the pruning threshold to prune in order to generate the interactive response decision tree.
[0011] In an optional embodiment, determining whether there is a conflict in the interactive response strategy data to be detected based on the interactive strategy results includes: determining that there is a conflict in the interactive response strategy data to be detected when the interactive strategy results include at least two; determining whether there is a conflict in the interactive response strategy data to be detected based on the detection results and the interactive strategy results includes: determining that there is a conflict in the interactive response strategy data to be detected when the interactive strategy results are different from the detection results.
[0012] In an exemplary embodiment, when the interactive strategy results include at least two, after determining that there is a conflict in the interactive response strategy data to be detected, the method further includes: determining the interactive strategy corresponding to each interactive strategy result, wherein the interactive strategy includes strategy conditions and strategy results corresponding to the strategy conditions; comparing each strategy condition and strategy result to determine a first difference strategy condition and a first difference strategy result among the interactive strategies; and outputting the difference strategy condition and the difference strategy result.
[0013] In one embodiment, when the interactive strategy result is different from the result to be detected, determining that there is a conflict in the interactive response strategy data to be detected includes: determining the interactive strategy corresponding to the interactive strategy result, wherein the interactive strategy includes strategy conditions and strategy results corresponding to the strategy conditions; comparing the strategy conditions to be detected trigger conditions and the strategy results with the trigger results corresponding to the trigger conditions to be detected, to determine the second difference strategy conditions and the second difference strategy results between the interactive response strategy data to be detected and the interactive strategy; and outputting the difference strategy conditions and the difference strategy results.
[0014] Secondly, this application also provides a collision detection device, comprising:
[0015] The strategy data acquisition module is used to acquire the interactive response strategy data to be detected. The interactive response strategy data to be detected includes the triggering conditions to be detected or the interactive response strategy data to be detected includes the triggering conditions to be detected and the detection results corresponding to the triggering conditions to be detected.
[0016] The strategy result acquisition module is used to input the triggering conditions to be detected into the pre-generated interactive response decision tree to obtain the interactive strategy result;
[0017] The conflict determination module is used to determine whether there is a conflict in the interactive response strategy data to be detected based on the interactive strategy results, or based on the detection results and the interactive strategy results.
[0018] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method embodiments.
[0019] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method embodiments.
[0020] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method embodiments.
[0021] The aforementioned conflict detection method, apparatus, computer equipment, computer-readable storage medium, and computer program product utilize an interactive response decision tree to detect conflicts in interactive response strategy data. The decision tree can be traversed based on the triggering conditions to be detected, and the resulting strategy determines whether a conflict has occurred. The generation of the decision tree and the detection of the interactive response strategy data based on the triggering conditions can both be automatically generated, effectively reducing inaccuracies caused by manual processing and ensuring the accuracy of conflict detection. Attached Figure Description
[0022] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0023] Figure 1 This is a flowchart illustrating a conflict detection method in one embodiment;
[0024] Figure 2 This is a flowchart illustrating the collision detection method in another embodiment;
[0025] Figure 3 This is a flowchart illustrating the conflict detection method in another embodiment;
[0026] Figure 4 This is a block diagram of an interactive policy conflict detection system in one embodiment;
[0027] Figure 5This is a structural block diagram of a collision detection device in one embodiment;
[0028] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation
[0029] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0030] In one embodiment, such as Figure 1 As shown, a conflict detection method is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:
[0031] Step 102: Obtain the interactive response strategy data to be detected. The interactive response strategy data to be detected includes the triggering conditions to be detected, or the interactive response data to be detected may include the trigger bars to be detected and the detection results corresponding to the triggering conditions to be detected.
[0032] The interactive response strategy data is used to characterize the process of implementing interactive responses.
[0033] Optionally, the interactive response can be an interactive voice response.
[0034] Interactive Voice Response (IVR) is an automated telephone answering system in the communications field. It allows businesses to provide services to customers by inputting information via dual-tone multi-frequency (DTMF) telephones and providing pre-recorded voice messages. For example, a user dials a business service number using a terminal, triggering the IVR system. Following the prompts provided by the service hotline, the user interacts with the IVR system to obtain the necessary information.
[0035] In one embodiment, the interactive response strategy data to be detected includes the triggering conditions to be detected.
[0036] Among them, the triggering condition to be detected is used to make the process jump from one node to another.
[0037] In one embodiment, the interactive response strategy data to be detected includes the triggering conditions to be detected and the detection results corresponding to the triggering conditions.
[0038] The detection result is used to characterize the jump from one node to another node based on the detection trigger condition.
[0039] In one embodiment, the interactive response strategy data to be detected includes at least one triggering condition to be detected and at least one result to be detected.
[0040] In an optional embodiment, an IVR interface module is provided for interfacing with the IVR system and obtaining data from the IVR system during its operation.
[0041] Optionally, the data generated during the operation of the IVR system includes, but is not limited to, IVR system log data, real-time call data, and system configuration data.
[0042] In one embodiment, an IVR interface module can be used to communicate with an IVR system to obtain IVR system log data, real-time call data, and system configuration data from the IVR system.
[0043] In one embodiment, a cache is set up so that after acquiring log data, real-time call data, and system configuration data of the IVR system, the acquired data can be stored in the cache so that the data can be retrieved from the cache when needed.
[0044] Optionally, real-time call data includes, but is not limited to, call initiation time, call number, customer case operation records, and voice interaction content.
[0045] Optionally, system configuration data includes, but is not limited to, data on effective interactive response policies.
[0046] In one embodiment, after acquiring log data, real-time call data, and system configuration data from the IVR system, preprocessing is performed on these data to obtain processed log data, real-time call data, and system configuration data. For example, preprocessing may include, but is not limited to, removing duplicate records, correcting erroneous data formats, and filling in missing values.
[0047] Optionally, removing duplicate records may include: obtaining data during the operation of the IVR system and obtaining a unique identifier for each data record; if the same identifier exists, there may be two identical data records, in which case one of the data records will be deleted.
[0048] Optionally, correcting erroneous data formats may include: obtaining the original format and target format of the data during the operation of the IVR system, determining the erroneous data in the data during the operation of the IVR system based on the original format, and converting the erroneous data into the target format data.
[0049] Optionally, filling missing values may include: traversing the data during the operation of the IVR system, identifying the data in the IVR system that has missing data, determining the data type of the data in the IVR system that has missing data, and filling the missing data according to the data type. For example, if the data is numerical, the mean or median of the data in the IVR system that has missing data can be obtained, and the mean or median can be used to fill the missing values. For example, if the data is categorical, the data before and after the missing data is obtained, and the data at the missing locations is predicted based on the data before and after the missing data, and the predicted data at the missing locations is used to fill the missing values. Optionally, the prediction algorithm may include, but is not limited to, mode imputation or imputation based on machine learning algorithms.
[0050] In one embodiment, after acquiring and preprocessing the data during the operation of the IVR system, feature extraction is performed on the processed data during the operation of the IVR system to extract key interactive response features from the data during the operation of the IVR system.
[0051] In one embodiment, the extracted key features of the interactive response can be format-converted to convert the features into a format that can be used for interactive response decision tree processing.
[0052] In one embodiment, the extracted key interactive response features can be divided into basic interactive response feature data and auxiliary interactive response feature data based on their importance. Optionally, the basic interactive response feature data may include, but is not limited to, service type and customer level. Optionally, the auxiliary interactive response feature data may include, but is not limited to, the customer's call frequency within a specific time period and the average processing time for different services.
[0053] In one embodiment, different format conversion methods can be applied to the extracted key features of the interactive response, depending on the data type of the key features. For example, if the data type is categorical features, one-hot encoding, label encoding, or other methods can be used to convert the features into numerical vectors. For example, if the data type is continuous features, normalization or standardization can be used to ensure that the processed features are on the same scale for further data processing, such as enabling comparisons between data.
[0054] In one embodiment, the trigger conditions to be detected are extracted by processing data during the operation of the IVR system. For example, the log data of the IVR system records records of response failures. The trigger conditions at which the failure should have occurred are obtained through these records. These trigger conditions can be used as the trigger conditions to be detected, and interactive response strategy data to be detected is generated based on the trigger conditions to be detected.
[0055] In one embodiment, a visual interface is provided on the terminal, allowing staff to edit new interactive response strategy data by inputting trigger conditions and the corresponding strategy results. The new interactive response strategy data can then be identified as the interactive response strategy data to be tested.
[0056] Optionally, after obtaining the interactive response strategy data to be detected, the terminal sends the interactive response strategy data to be detected to the server, and the server performs conflict detection processing based on the obtained interactive response strategy data to be detected.
[0057] Step 104: Input the triggering conditions to be detected into the pre-generated interactive response decision tree, and obtain the interactive strategy results output by the interactive response decision tree.
[0058] The interactive response decision tree is used to generate strategy results based on triggering conditions.
[0059] In one embodiment, an interactive response decision tree can be pre-generated.
[0060] In one embodiment, after acquiring data from the IVR system during its operation, model training data can be determined based on this data, and an interactive response decision tree can be generated based on the determined model training data. For example, effective interactive response strategy data included in the data from the IVR system's operation is acquired, effective interactive response strategy data is used to generate model training data, and an interactive response decision tree is generated based on the generated model training data.
[0061] In one embodiment, after acquiring data from the IVR system's operation, a validation dataset can be generated based on this data. This validation dataset is used to ensure that the interactive response decision tree achieves optimal performance on the validation dataset during the training process using the model training data.
[0062] In an optional embodiment, a test sample dataset can also be generated based on data generated during the operation of the IVR system. This test sample dataset is used to ensure that the accuracy, recall, and other metrics of the trained decision tree in conflict detection on test data meet set requirements.
[0063] In one embodiment, 70% of the data from the IVR system's operation can be selected as training data for the interactive response decision tree. Optionally, 15% of the data from the IVR system's operation can be selected as a validation dataset. Optionally, 15% of the data from the IVR system's operation can be selected as a test sample dataset. In one embodiment, the model training data can be data from an already implemented interactive response strategy.
[0064] In an optional embodiment, the model training data may also include real-time operational data from the IVR system's operation that is not recorded in the already effective interactive response strategy data but yields corresponding strategy results based on given trigger conditions. Optionally, new interactive response strategy data can be generated based on the real-time operational data from the IVR system's operation that is not recorded in the already effective interactive response strategy data but yields corresponding strategy results based on given trigger conditions. This new interactive response strategy data can then be used to update the already effective interactive response strategy data, thereby expanding the amount of model training sample data generated based on the already effective interactive response strategy data and improving the accuracy of the generated interactive response decision tree, thus enhancing the accuracy of conflict detection.
[0065] Optionally, the verification dataset may include data on already effective interactive response strategies. Optionally, it may also include real-time operational data from the IVR system's operation that is not recorded in the already effective interactive response strategy data but yields corresponding strategy results based on given trigger conditions. Optionally, new interactive response strategy data may be generated based on this real-time operational data from the IVR system's operation that is not recorded in the already effective interactive response strategy data but yields corresponding strategy results based on given trigger conditions. This new interactive response strategy data is then used to update the already effective interactive response strategy data, thereby expanding the data volume of the verification dataset generated based on the already effective interactive response strategy data, improving the accuracy of the generated interactive response decision tree, and thus improving the accuracy of conflict detection.
[0066] Optionally, the test sample dataset may include data on already effective interactive response strategies. Optionally, it may also include real-time operational data from the IVR system's operation that is not recorded in the already effective interactive response strategy data but yields corresponding strategy results based on given trigger conditions. Optionally, new interactive response strategy data may be generated based on this real-time operational data from the IVR system's operation that is not recorded in the already effective interactive response strategy data but yields corresponding strategy results based on given trigger conditions. This new interactive response strategy data is then used to update the already effective interactive response strategy data, thereby expanding the data volume of the test sample dataset generated based on the already effective interactive response strategy data, improving the accuracy of the generated interactive response decision tree, and thus improving the accuracy of conflict detection.
[0067] Optionally, the effective triggering conditions may include, but are not limited to, call source, time period, customer type, etc.
[0068] Optionally, the effective policy results corresponding to the effective triggering conditions may include, but are not limited to, routing calls to a specific agent group, playing a specific voice menu, or executing specific business logic.
[0069] In an optional embodiment, the decision tree trained using the model training data can be an EnhanceTree.
[0070] In one embodiment, during the decision tree generation process, the node splitting criteria can be adjusted to ensure that the resulting EnhanceTree is more accurate, thereby further improving the accuracy of conflict detection.
[0071] For example, the method for adjusting the node splitting criteria may include: acquiring effective interactive response policy data, which includes effective policy triggering conditions and corresponding effective policy results; determining the service type corresponding to the effective interactive response policy data and the condition type corresponding to the effective policy triggering conditions, and determining the weight data corresponding to the effective policy triggering conditions based on the service type and condition type; determining the correlation data between each effective policy triggering condition and adjacent effective policy triggering conditions; and splitting the interactive response decision tree node based on the weight data and correlation data corresponding to the effective policy triggering conditions.
[0072] In an optional embodiment, a logical weight factor can be assigned to each strategy triggering condition based on factors such as the importance of the strategy in the business process and its impact on customer experience. Optionally, the value range of the weight factor can be [0.1~1].
[0073] For example, the "Customer Important Procedures" condition in the "Priority Transfer for Red and White List Customers" strategy can be assigned a weight factor of 0.8 because it is crucial to business processing. For example, auxiliary conditions such as "User Matching Rounds" can be assigned a weight factor of 0.2.
[0074] Among them, the correlation data is used to characterize the sequential order of strategy conditions in the business process and the correlation between them and adjacent strategy conditions.
[0075] In one embodiment, the effective interactive response policy data includes, but is not limited to, policy name, policy number, effective trigger conditions, and effective policy results.
[0076] In one embodiment, the association between effective policy triggering conditions can be determined based on the effective policy triggering conditions and the effective policy results corresponding to the effective policy triggering conditions included in the effective interactive response policy data. Optionally, the strength of the association can be determined based on the order of the effective policy conditions in the business process and the association between adjacent effective policy conditions.
[0077] Optionally, the correlation data corresponding to the triggering conditions of the effective strategy can be determined based on the strength of the correlation.
[0078] For example, if selecting "Voice Menu Option 1" will inevitably trigger the "Route to Technical Support Agent Group" policy, then the contextual correlation between these two policy conditions is high and can be set to 0.7.
[0079] In one embodiment, when splitting decision tree nodes, information gain, weight data, and correlation data can be comprehensively considered to achieve a comprehensive score. Optionally, the formula for calculating the comprehensive score can be as shown in formula (1):
[0080] (1)
[0081] Where IG is the information gain, W is the policy logic weight factor, C is the context relevance index, and alpha, eta, and gamma are weight coefficients.
[0082] Optionally, alpha, eta, and gamma are weighting coefficients that can be determined through experimental optimization.
[0083] Optionally, the decision tree nodes can be split based on the obtained comprehensive score to generate an interactive response decision tree.
[0084] In an optional implementation, during the generation of the decision tree, the cost complexity of the subtree can be determined. Based on the cost complexity, a pruning path can be implemented, and the decision tree can be generated based on the pruning path. Optionally, the formula for calculating the cost complexity can be as shown in formula (2):
[0085] (2)
[0086] Where R(t) is the misclassification cost of subtree t, |N(t)| is the number of leaf nodes of subtree t, and alpha is the complexity parameter.
[0087] In an optional embodiment, the decision tree generation process can also adjust the pruning threshold to achieve dynamic adjustment of the decision tree, thereby improving the accuracy of the trained decision tree and ensuring the accuracy of conflict detection using the decision tree.
[0088] For example, the pruning threshold adjustment method may include: determining the business precision corresponding to the business type based on the business type corresponding to the effective interactive response policy data; adjusting the pruning threshold based on the business precision; and using the pruning threshold to perform pruning to generate an interactive response decision tree.
[0089] In an alternative embodiment, if high business accuracy is required, the pruning threshold can be lowered to retain more subtree structures; if low business accuracy is required, the pruning threshold can be raised to simplify the decision tree model. During the pruning process, the process recursively moves upward from the leaf nodes, comparing the cost complexity before and after subtree pruning. If the cost complexity decreases after pruning and does not exceed a set threshold, then the pruning operation is performed.
[0090] In one embodiment, the pruning threshold can also be adjusted based on the data size used to generate the interactive decision tree. For example, the data size of the training sample data is obtained; if the data size is greater than or equal to a preset threshold, the pruning threshold is lowered to retain more subtree structures; if the data size is less than the preset threshold, the pruning threshold is increased to simplify the decision tree model.
[0091] In one embodiment, after obtaining the trigger condition to be detected, the trigger condition to be detected is input into the interactive response decision tree, and the interactive response decision tree is traversed based on the trigger condition to obtain the interactive strategy result.
[0092] The results of the interactive strategy can be used to determine whether there are conflicts in the interactive response strategy data to be detected.
[0093] In one embodiment, if a change is detected in the interactive response strategy data, the above steps are repeated to perform data preprocessing and extract features from the preprocessed data in order to retrain the interactive response decision tree. Optionally, changes in the interactive response strategy data may include, but are not limited to, adding, modifying, or deleting strategies.
[0094] In one embodiment, the retrained model is compared with the model before retraining to determine whether the new model performs better than the old model. When it is determined that the performance of the new model is higher than that of the old model, the old model is replaced with the new model to achieve conflict detection.
[0095] Step 106: Determine whether there is a conflict in the interactive response strategy data to be detected based on the interactive strategy results, or determine whether there is a conflict in the interactive response strategy data to be detected based on both the interactive strategy results and the interactive strategy results.
[0096] In one embodiment, the presence of conflicts in the interactive response strategy data to be detected can be determined based on the number of interactive strategy results.
[0097] Understandably, if there are at least two interactive policy results, it indicates that there are two outcomes for the same triggering condition, which may indicate a conflict. For example, a conflict determination method may include: determining that there is a conflict in the interactive response policy data to be detected when there are at least two interactive policy results.
[0098] Optionally, conflict analysis can also be performed when it is determined that there is a conflict in the interactive response strategy data.
[0099] For example, the interactive strategy corresponding to each interactive strategy result is determined, and the interactive strategy includes strategy conditions and strategy results corresponding to the strategy conditions; each strategy condition and strategy result is compared to determine the first difference strategy condition and the first difference strategy result among each interactive strategy; the first difference strategy condition and the first difference strategy result are output.
[0100] In one embodiment, a conflict can also be determined by comparing the result to be detected with the result of the interactive policy. For example, a conflict determination method may include: determining that there is a conflict in the interactive response policy data to be detected when the result of the interactive policy is different from the result to be detected.
[0101] In one embodiment of the integrated function, the difference between the interactive strategy result and the result to be detected may include: when there are multiple triggering conditions to be detected, if the interactive strategy result corresponding to any triggering condition to be detected is different from the result to be detected, it can be considered that there is a conflict in the interactive response strategy to be detected.
[0102] Optionally, when a conflict is determined in the interactive response strategy data to be detected by comparing the results of the detection with the results of the interactive strategy, conflict analysis can also be performed to determine the specific cause of the conflict.
[0103] For example, the interactive strategy corresponding to the interactive strategy result is determined, and the interactive strategy includes strategy conditions and strategy results corresponding to the strategy conditions; the strategy conditions to be detected trigger conditions and the strategy results to be detected trigger conditions are compared to determine the second difference strategy conditions and the second difference strategy results between the interactive response strategy data to be detected and the interactive strategy; the second difference strategy conditions and the second difference strategy results are output.
[0104] For example, if a call routing conflict is detected, the triggering conditions are analyzed. The analysis shows that the conflict is caused by overlapping conditions (such as two routing policies targeting "3990 Complaint Zone, VIP Customers") or by incorrect priority settings.
[0105] Optionally, after using a decision tree to detect whether there is a conflict in the interactive response strategy to be detected, if a conflict exists, the business logic and actual application scenario of the interactive response strategy to be detected are obtained, and a detailed conflict analysis report is generated. Optionally, the conflict analysis includes, but is not limited to, the specific location where the conflict occurs, the business processes that may be affected, and the affected customer groups.
[0106] In one embodiment, when a conflict is detected, the severity of the detected conflict is determined. If the severity is greater than or equal to a preset threshold, a prompt message is generated to prompt staff to handle the situation.
[0107] Optionally, determining the severity of a detected conflict may include assessing the severity of the conflict based on its frequency, scope of impact, and the type of business involved.
[0108] Specifically, when the severity level is determined to be greater than or equal to a preset threshold, a prompt message is generated to alert staff to take action. This message can be sent via SMS, email, or to relevant personnel. Optionally, the relevant personnel can be the IVR system manager, operations manager, or business manager, etc.
[0109] In the aforementioned conflict detection method, an interactive response decision tree is used to detect data conflicts in the interactive response strategy. This interactive response decision tree is trained based on weighted data determined by the type of response and the triggering condition, as well as the correlation data corresponding to the triggering condition, effectively ensuring the accuracy of the decision tree and thus the accuracy of conflict detection. Furthermore, the pruning threshold during the generation of the interactive response decision tree can be adjusted according to business precision, further guaranteeing the accuracy of the generated decision tree.
[0110] In one embodiment, such as Figure 2 As shown, a conflict detection method is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:
[0111] Step 202: Obtain the effective interactive response policy data. The effective interactive response policy data includes the effective policy trigger conditions and the effective policy results corresponding to the effective policy trigger conditions.
[0112] Step 204: Determine the business type corresponding to the effective interactive response policy data and the condition type corresponding to the effective policy triggering condition, and determine the weight data corresponding to the effective policy triggering condition based on the business type and condition type.
[0113] Step 206: Determine the correlation data between each effective policy trigger condition and adjacent effective policy trigger conditions.
[0114] Step 208: Split the interactive response decision tree node according to the weight data and correlation data corresponding to the triggering conditions of the effective strategy.
[0115] Step 210: Determine the service precision corresponding to the service type based on the service type corresponding to the effective interactive response policy data.
[0116] Step 212: Adjust the pruning threshold according to the business precision, and use the pruning threshold to prune in order to generate an interactive response decision tree.
[0117] Step 214: Obtain the interactive response strategy data to be detected. The interactive response strategy data to be detected includes the triggering conditions to be detected, or the interactive response strategy data to be detected includes the triggering conditions to be detected and the detection results corresponding to the triggering conditions to be detected.
[0118] Step 216: Input the triggering conditions to be detected into the pre-generated interactive response decision tree to obtain the interactive strategy results.
[0119] Step 218: Determine whether there is a conflict in the interactive response strategy data to be tested based on the interactive strategy results, or determine whether there is a conflict in the interactive response strategy data to be tested based on the test results and the interactive strategy results.
[0120] In the aforementioned conflict detection method, an interactive response decision tree is used to detect data conflicts in the interactive response strategy. This interactive response decision tree is trained based on weighted data determined by the type of response and the triggering condition, as well as the correlation data corresponding to the triggering condition, effectively ensuring the accuracy of the decision tree and thus the accuracy of conflict detection. Furthermore, the pruning threshold during the generation of the interactive response decision tree can be adjusted according to business precision, further guaranteeing the accuracy of the generated decision tree.
[0121] In one embodiment, such as Figure 3 As shown, a conflict detection method is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:
[0122] Step 302: Obtain the interactive response strategy data to be detected. The interactive response strategy data to be detected includes the triggering conditions to be detected or includes the triggering conditions to be detected and the corresponding detection results.
[0123] Step 304: Input the triggering conditions to be detected into the pre-generated interactive response decision tree to obtain the interactive strategy results.
[0124] Step 306: Determine whether the interactive strategy result is at least two. If yes, proceed to step 308; otherwise, proceed to step 314.
[0125] Step 308: Determine the interactive strategy corresponding to each interactive strategy result. The interactive strategy includes the strategy condition and the strategy result corresponding to the strategy condition.
[0126] Step 310: Compare the policy conditions and policy results to determine the first difference policy conditions and the first difference policy results among the interactive policies.
[0127] Step 312: Output the first difference strategy conditions and the first difference strategy results.
[0128] Step 314: Determine the interactive policy corresponding to the interactive policy result. The interactive policy includes the policy conditions and the policy results corresponding to the policy conditions.
[0129] Step 316: Compare the strategy conditions with the trigger conditions to be detected, and the strategy results with the trigger results corresponding to the trigger conditions to be detected, in order to determine the second difference strategy conditions and the second difference strategy results between the interactive response strategy data to be detected and the interactive strategy.
[0130] Step 318: Output the second difference strategy conditions and the second difference strategy results.
[0131] In this embodiment, an interactive response decision tree is used to detect data conflicts in the interactive response strategy. The interactive response decision tree is trained based on weighted data determined by the type of response and the triggering condition, as well as the correlation data corresponding to the triggering condition, effectively ensuring the accuracy of the decision tree and thus the accuracy of conflict detection. Furthermore, the pruning threshold can be adjusted during the generation of the interactive response decision tree according to business precision, further ensuring the accuracy of the generated decision tree.
[0132] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.
[0133] In one embodiment, such as Figure 4 This paper provides an interactive policy conflict detection system to enable the detection of policy conflicts.
[0134] Optionally, the interactive policy conflict detection system includes an access layer, a data processing layer, a core processing layer, a display and management layer, and a storage layer.
[0135] In one embodiment, the access layer includes an IVR system access module for communicating with the IVR system to obtain data from the IVR system during its operation.
[0136] In an optional embodiment, the IVR system access module establishes a communication connection with the IVR system via an HTTP interface. Optionally, the HTTP interface enables bidirectional data transmission, allowing the interactive policy conflict detection system to obtain data from the IVR system during its operation, and also to send instruction information sent by the interactive policy conflict detection system to the IVR system, such as sending interactive response policy adjustment instructions.
[0137] In one embodiment, the HTTP interface may include, but is not limited to, customer service IVR interfaces, marketing IVR interfaces, and debt collection IVR interfaces.
[0138] In one embodiment, the data processing layer includes a data acquisition module and a data preprocessing and feature engineering module.
[0139] Optionally, the data acquisition module is used to extract data from the data obtained from the access layer during the operation of the IVR system.
[0140] Optionally, the data acquisition module can also be used to implement data caching, temporarily storing the acquired data from the IVR system's operation in a cache area. Optionally, the cache area can employ memory caching technology (such as Redis) to achieve fast data read and write.
[0141] Optionally, the data acquisition module can also be used to transmit data temporarily stored in the buffer to the data preprocessing module according to preset rules. Optionally, the preset rules can be time intervals or data volume thresholds. For example, data in the buffer is transmitted to the data preprocessing and feature engineering module at preset time intervals; for example, when the data volume exceeds the data volume threshold, the data in the buffer is transmitted to the data preprocessing and feature engineering module.
[0142] In one embodiment, the data preprocessing and feature engineering module can be used to perform data cleaning, feature extraction and transformation, and feature filtering and combination.
[0143] Optionally, when implementing data cleaning, the data preprocessing and feature engineering modules may include performing a comprehensive cleaning of the data obtained from the cache, removing duplicate records, correcting erroneous data formats, and filling in missing values. For duplicate records, they are identified and deleted by comparing the unique identifiers of the data; for erroneous data formats, they are converted according to the definition of the data type; for missing values, appropriate filling methods are selected according to the characteristics of the data, such as using the mean or median for numerical data, and using the mode or predictive filling based on machine learning algorithms for categorical data.
[0144] Optionally, feature extraction and transformation can be used to extract key features from the cleaned data and transform them into a format suitable for EnhanceTree decision tree processing.
[0145] Optionally, features can be divided into basic features and auxiliary features based on their importance. For example, basic features can be features such as business type and customer level; auxiliary features can be features such as the customer's call frequency in a specific time period and the average processing time of different services.
[0146] Optionally, different preprocessing methods can be used depending on the classification of the extracted features. For example, for categorical features, one-hot encoding, label encoding, or other methods can be used to convert them into numerical vectors; for continuous features, normalization or standardization can be performed to make different features comparable on the same scale.
[0147] Optionally, feature selection and combination can be used to determine features that are likely to have an impact on conflict detection. For example, features obtained from feature extraction and transformation can be combined, such as combining certain features from the basic features with certain features from the auxiliary features, to obtain multiple feature combinations. These multiple feature combinations can be input into an interactive response decision tree, and the strategy results given by the interactive response decision tree can be used to determine which feature combination has the greatest impact on conflict.
[0148] In an optional embodiment, the core processing layer includes a conflict detection core module and a conflict analysis and early warning module.
[0149] Optionally, the conflict detection core module may include the EnhanceTree decision tree engine, which receives feature data processed by the data preprocessing and feature engineering modules, and uses the decision tree engine to perform inference operations on the feature data. During the inference process, the decision tree starts from the root node based on the input data and traverses downwards step by step according to the node splitting conditions until it reaches the leaf node. Each leaf node corresponds to a policy result. If, during the traversal, the same input data corresponds to multiple different policy results or the results are inconsistent with the preset business logic, a policy conflict is determined to exist.
[0150] Optionally, the conflict detection core module can implement model training and updating. For example, when the IVR policy changes (e.g., adding, modifying, or deleting a policy), this module automatically triggers the model update process. Optionally, new policy data and related runtime data are incorporated into the training dataset, which is then cleaned, preprocessed, and its features extracted. Then, the EnhanceTree decision tree model is rebuilt using the updated training dataset, repeating steps such as feature selection, node splitting, decision tree growth, and pruning to generate a new decision tree model. Optionally, after the new model is built, its performance is evaluated using methods such as cross-validation to determine if the new model outperforms the old model. If the new model outperforms the old model, it replaces the old model to ensure the accuracy and real-time performance of conflict detection.
[0151] Optionally, the conflict detection core module may also include a conflict analysis and early warning module for in-depth conflict analysis. Optionally, the severity of the conflict is assessed based on factors such as the frequency of occurrence, scope of impact, and type of business involved. Optionally, when the severity exceeds a preset threshold, relevant personnel are notified through preset prompts. For example, preset prompts may include, but are not limited to, SMS and email; optionally, relevant personnel may include, but are not limited to, the IVR system manager, operations manager, and business manager.
[0152] Optionally, the conflict analysis and early warning module can also provide a detailed description of the conflict and preliminary solution suggestions so that relevant personnel can respond quickly and take measures to resolve the conflict.
[0153] In one embodiment, the presentation and management layer includes a results presentation and analysis module and a strategy management module.
[0154] Optionally, the results display and analysis module may include a visualization interface that displays conflict detection results through visual elements. For example, visualization elements may include, but are not limited to, bar charts, line charts, pie charts, and flowcharts.
[0155] Optionally, the results display and analysis module can implement data analysis functions, providing data analysis tools so that users can filter conflicting data according to different filtering conditions and conduct in-depth analysis of conflict situations under specific conditions. Optionally, filtering conditions include, but are not limited to, business departments, time ranges, customer types, conflict types, etc.
[0156] Optionally, the results display and analysis module can also generate and export reports, producing a detection report based on the conflict detection results. For example, the report content includes a conflict overview, conflict analysis results, and suggested solutions.
[0157] Optionally, the policy management module can be used to manage interactive response policies. For example, it allows for entering new interactive response policies, editing existing interactive response policies, and deleting unused interactive response policies. Optionally, a policy change log can be generated for each policy change, so that users can identify changes that have occurred when needed. Optionally, policy changes include, but are not limited to, change time, changed personnel, and changed content.
[0158] In one embodiment, when a policy change occurs, the interactive response decision tree can be retrained to ensure the accuracy of the decision tree used to implement conflict detection.
[0159] In one embodiment, the storage layer can be a distributed storage system, for example, Ceph or GlusterFS.
[0160] In one embodiment, the storage layer can implement data encryption and backup. To ensure data security, the data stored in the distributed storage system can be encrypted. Optionally, the data stored in the distributed storage system may include, but is not limited to, IVR policy data, decision tree models, log records, etc.
[0161] Optionally, the data can be encrypted. When retrieving data, the identity and permissions of the user requesting the data are verified. If the identity and permissions are verified, the data requested by the user is sent to the user's corresponding terminal or server. For example, the encryption algorithm can be AES.
[0162] Optionally, data backup can be performed to a remote data center to prevent data loss.
[0163] Based on the same inventive concept, this application also provides a conflict detection apparatus for implementing the conflict detection method described above. The solution provided by this apparatus is similar to the implementation described in the above method; therefore, the specific limitations in one or more conflict detection apparatus embodiments provided below can be found in the limitations of the conflict detection method described above, and will not be repeated here.
[0164] In one exemplary embodiment, such as Figure 5 As shown, a conflict detection device 500 is provided, including: a strategy data acquisition module 502, a strategy result acquisition module 504, and a conflict determination module 506, wherein:
[0165] The strategy data acquisition module 502 is used to acquire interactive response strategy data to be detected. The interactive response strategy data to be detected includes the triggering conditions to be detected or includes the triggering conditions to be detected and the corresponding detection results.
[0166] The strategy result acquisition module 504 is used to input the triggering conditions to be detected into the pre-generated interactive response decision tree to obtain the interactive strategy result.
[0167] The conflict determination module 506 is used to determine whether there is a conflict in the interactive response strategy data to be detected based on the interactive strategy result, or to determine whether there is a conflict in the interactive response strategy data to be detected based on the detection result and the interactive strategy result.
[0168] In one embodiment, the conflict detection device further includes a decision tree generation module, used to acquire effective interactive response policy data, which includes effective policy triggering conditions and effective policy results corresponding to the effective policy triggering conditions; determine the service type corresponding to the effective interactive response policy data and the condition type corresponding to the effective policy triggering conditions, and determine the weight data corresponding to the effective policy triggering conditions based on the service type and condition type; determine the correlation data between each effective policy triggering condition and adjacent effective policy triggering conditions; and split the interactive response decision tree nodes based on the weight data and correlation data corresponding to the effective policy triggering conditions.
[0169] In one embodiment, the decision tree generation module is further configured to determine the business precision corresponding to the business type based on the business type corresponding to the effective interactive response strategy data; adjust the pruning threshold according to the business precision; and use the pruning threshold to perform pruning to generate an interactive response decision tree.
[0170] In an optional embodiment, the conflict determination module is further configured to determine that there is a conflict in the interactive response strategy data to be detected when the interactive strategy results include at least two; or, when the interactive strategy results are different from the results to be detected, determine that there is a conflict in the interactive response strategy data to be detected.
[0171] In an optional embodiment, the conflict determination module is further configured to determine the interactive strategy corresponding to each interactive strategy result, wherein the interactive strategy includes strategy conditions and strategy results corresponding to the strategy conditions; compare each strategy condition and strategy result to determine the first difference strategy condition and the first difference strategy result among each interactive strategy; and output the first difference strategy condition and the first difference strategy result.
[0172] In an optional embodiment, the conflict determination module is further configured to determine the interactive strategy corresponding to the interactive strategy result, wherein the interactive strategy includes strategy conditions and strategy results corresponding to the strategy conditions; compare the strategy conditions to be detected trigger conditions and the strategy results with the trigger results corresponding to the to be detected trigger conditions to determine the second difference strategy conditions and the second difference strategy results between the interactive response strategy data to be detected and the interactive strategy; and output the second difference strategy conditions and the second difference strategy results.
[0173] Each module in the aforementioned collision detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.
[0174] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6 As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When executed by the processor, the computer program implements a conflict detection method.
[0175] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0176] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method embodiments.
[0177] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method embodiments.
[0178] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the above-described method embodiments.
[0179] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0180] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.
[0181] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0182] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A collision detection method, characterized in that, The method includes: Acquire interactive response strategy data to be detected, wherein the interactive response strategy data to be detected includes a trigger condition to be detected or the interactive response strategy data to be detected includes a trigger condition to be detected and a detection result corresponding to the trigger condition to be detected; The triggering conditions to be detected are input into the pre-generated interactive response decision tree to obtain the interactive strategy results; Based on the interactive strategy result, determine whether there is a conflict in the interactive response strategy data to be detected, or determine whether there is a conflict in the interactive response strategy data to be detected based on the detection result and the interactive strategy result.
2. The method according to claim 1, characterized in that, Before inputting the triggering condition to be detected into the pre-generated interactive response decision tree to obtain the interactive strategy result, the method includes: Obtain effective interactive response policy data, which includes effective policy triggering conditions and effective policy results corresponding to the effective policy triggering conditions; Determine the service type corresponding to the effective interactive response policy data and the condition type corresponding to the effective policy triggering condition, and determine the weight data corresponding to the effective policy triggering condition based on the service type and the condition type; Determine the correlation data between each of the aforementioned effective policy triggering conditions and adjacent effective policy triggering conditions; Based on the weight data corresponding to the triggering conditions of the effective strategy and the correlation data, the nodes of the interactive response decision tree are split.
3. The method according to claim 2, characterized in that, After splitting the nodes of the interactive response decision tree, the method further includes: Based on the service type corresponding to the effective interactive response strategy data, determine the service precision corresponding to the service type; The pruning threshold is adjusted according to the business precision, and the pruning threshold is used to prune in order to generate the interactive response decision tree.
4. The method according to any one of claims 1-3, characterized in that, The step of determining whether the interactive response strategy data to be detected has a conflict based on the interactive strategy result includes: When the interactive strategy results include at least two, it is determined that there is a conflict in the interactive response strategy data to be detected; The step of determining whether there is a conflict in the interactive response strategy data to be detected based on the detection result and the interactive strategy result includes: When the result of the interactive strategy is different from the result to be detected, it is determined that there is a conflict in the interactive response strategy data to be detected.
5. The method according to claim 4, characterized in that, The step of determining that the interactive response strategy data to be detected has a conflict when the interactive strategy results include at least two is as follows: Determine the interactive strategy corresponding to each of the interactive strategy results, wherein the interactive strategy includes strategy conditions and strategy results corresponding to the strategy conditions; Compare each of the strategy conditions and the strategy results to determine the first difference strategy conditions and the first difference strategy results among the interactive strategies; Output the first difference strategy conditions and the first difference strategy results.
6. The method according to claim 4, characterized in that, When the interactive strategy result is different from the result to be detected, it is determined that there is a conflict in the interactive response strategy data to be detected: including: Determine the interactive strategy corresponding to the interactive strategy result, wherein the interactive strategy includes strategy conditions and strategy results corresponding to the strategy conditions; The strategy conditions and the trigger conditions to be detected are compared, and the strategy results and the trigger results corresponding to the trigger conditions to be detected are compared to determine the second difference strategy conditions and the second difference strategy results between the interactive response strategy data to be detected and the interactive strategy. Output the second difference strategy conditions and the second difference strategy results.
7. A collision detection device, characterized in that, The device includes: The strategy data acquisition module is used to acquire interactive response strategy data to be detected, wherein the interactive response strategy data to be detected includes a trigger condition to be detected or the interactive response strategy data to be detected includes a trigger condition to be detected and a detection result corresponding to the trigger condition to be detected. The strategy result determination module is used to input the triggering condition to be detected into a pre-generated interactive response decision tree to obtain the interactive strategy result; The conflict determination module is used to determine whether there is a conflict in the interactive response strategy data to be detected based on the interactive strategy result, or to determine whether there is a conflict in the interactive response strategy data to be detected based on the detection result and the interactive strategy result.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 4 or 5.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 4 or 5.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 4 or 5.