Call center order integration processing method based on data analysis

Abstract

The invention discloses a call center order integration processing method based on data analysis, and relates to the technical field of order integration processing, and the method comprises the steps: extracting call orders in a preset time window, and generating order basic information; historical data extraction is carried out on call records of the order receiving module, historical order attribute analysis is carried out in combination with a preset feature attribute set, and an order recognizer is constructed; order attribute analysis is carried out to obtain an order feature attribute set; identifying the order feature attribute set by using an identification branch of the order identifier to obtain a call order cluster; and configuring a processing unit of the order processing module according to the processing fitness, and sending the call order cluster to the processing unit for order processing. According to the method, the technical problems of poor order processing flexibility and low processing efficiency caused by difficulty in coping with complex and changeable customer requirements and changing market environments in the prior art are solved, and the technical effect of efficient and accurate processing of the orders of the call center is achieved through data analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of order integration processing, and in particular to a call center order integration processing method based on data analysis. Background Art

[0002] With the rapid development of information technology and the popularization of the Internet, call centers, as an important interactive platform between enterprises and customers, are also rapidly expanding in terms of business scale and processing capacity. Call centers need to handle a large number of customer inquiries, complaints, orders and other matters, and ensure that these matters can be processed efficiently, accurately and promptly. However, the traditional call center order processing method mainly relies on fixed computer programs for process-based operations, lacks flexibility and intelligence, and is difficult to cope with complex and changing customer needs and the ever-changing market environment. Summary of the invention

[0003] The present application provides a call center order integration processing method based on data analysis, which is used to solve the technical problem that the existing technology is difficult to cope with complex and changeable customer needs and the ever-changing market environment, resulting in poor flexibility in order processing.

[0004] In view of the above problems, the present application provides a call center order integration processing method based on data analysis.

[0005] The present application provides a method for optimizing common components based on an application market, the method comprising: Collect multiple public component information within the target enterprise through registration; retrieve the operation records of the multiple public component information, perform call analysis on the multiple public component information according to the operation records, and generate call analysis results; establish a review mechanism based on the call analysis results, review the multiple public component information, and obtain public component review results; establish a feedback channel based on the management model of the application market, synchronize the public component review results to the feedback channel, and output repair feedback data; screen the multiple public component information based on the repair feedback data, and optimize the quality of the multiple public component information according to the screening results.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: The order receiving module of the interactive call center of the present application extracts N call orders within a preset time window and generates N basic order information; extracts historical data from the call records of the order receiving module, and analyzes the historical order attributes in combination with the preset feature attribute set, and constructs an order identifier according to the analysis result, wherein the order identifier has K identification branches, and the K identification branches respectively store K historical order feature attribute sets, and have K processing fitness; analyzes the order attributes of the N basic order information with the preset feature attribute set to obtain N order feature attribute sets; uses the K identification branches of the order identifier to identify the N order feature attribute sets to obtain M call order clusters, wherein the M call order clusters have M processing fitness; configures the M processing units of the order processing module according to the M processing fitness, and sends the M call order clusters to the M processing units for order processing. The present invention solves the technical problem that the prior art is difficult to cope with complex and changeable customer needs and the ever-changing market environment, resulting in poor order processing flexibility and low processing efficiency, and achieves the technical effect of efficient and accurate processing of call center orders through data analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0008] Figure 1 A flowchart of a call center order integration processing method based on data analysis provided in an embodiment of the present application; Figure 2 A schematic diagram of a process for constructing an order identifier based on analysis results provided in an embodiment of the present application. DETAILED DESCRIPTION

[0009] This application provides a call center order integration processing method based on data analysis, aiming to solve the technical problem that the existing technology is difficult to cope with complex and changeable customer needs and the ever-changing market environment, resulting in poor order processing flexibility and low processing efficiency. Through data analysis, the technical effect of efficient and accurate processing of call center orders is achieved.

[0010] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0011] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0012] Embodiment 1 like Figure 1 As shown, the present application provides a call center order integration processing method based on data analysis, the method comprising: Step S100: The order receiving module of the interactive call center extracts N call orders within a preset time window and generates N basic order information; In the embodiment of the present application, the order receiving module first determines the preset time window according to the business needs of the enterprise or the operation strategy of the call center. The preset time window is the past hour, day, week or other specific time period. After determining the preset time window, the order receiving module of the call center records the start and end time points of the preset time window and extracts the order data within the corresponding time range from the database.

[0013] When extracting call orders from the database, the order receiving module connects to the database of the call center and prepares to perform query operations. Using SQL or other database query languages, all eligible call orders are screened from the database according to the start and end time points of the preset time window. After extracting the call order data, the order receiving module cleans and organizes the data. After cleaning and organizing, the order receiving module generates N basic order information.

[0014] Step S200: extracting historical data from the call records of the order receiving module, and analyzing historical order attributes in combination with a preset feature attribute set, and constructing an order identifier according to the analysis result, wherein the order identifier has K identification branches, and the K identification branches respectively store K historical order feature attribute sets and have K processing fitnesses; In the embodiment of the present application, historical data is first extracted from the call record of the order receiving module. The historical data includes all order information previously processed by the call center, including but not limited to order number, customer information, order time, processing time, processing results, etc.

[0015] After extracting historical data, historical orders are analyzed based on the preset feature attribute set. The preset feature attribute set is determined based on the business characteristics and needs of the call center, including order type, customer type, order time period, etc. By analyzing the attributes of historical orders, the order distribution, processing efficiency, and resource consumption under different feature attributes are determined.

[0016] According to the results of the historical order attribute analysis, the number of identification branches K of the order identifier is determined. Each identification branch focuses on identifying orders with a certain characteristic attribute. From the historical order data set, a representative set of historical order characteristic attributes is extracted for each identification branch. These historical order characteristic attribute sets are used as reference standards for the identification branches to identify new call orders.

[0017] Processing fitness is an important indicator to measure the order processing efficiency and resource consumption of the order identifier branch. It is calculated based on the processing time and resources consumed of historical orders. For each identification branch, when calculating the processing fitness, extract the processing time and resource consumption data of the historical orders processed by the branch, and calculate the average processing time of the branch to process historical orders, that is, the sum of the processing time of all historical orders of the branch divided by the number of orders. Calculate the average resource consumption of the branch to process historical orders, that is, the sum of the resources consumed by all historical orders of the branch divided by the number of orders. According to the average processing time and the average resource consumption, combined with the predetermined weights, a weighted calculation is performed to obtain the processing fitness of the branch. The weight is adjusted according to the actual situation to reflect the importance of processing time and resource consumption in evaluating processing fitness. K processing fitnesses are calculated in sequence.

[0018] Step S300: performing order attribute analysis on the N basic order information using the preset characteristic attribute set to obtain N order characteristic attribute sets; In an embodiment of the present application, a preset feature attribute set is used to perform order attribute analysis on N basic order information, and the corresponding feature values ​​in the preset feature attribute set are extracted from the basic information of each order. For certain non-numeric feature attributes, such as order type, customer category, etc., encoding is performed and converted into numerical data. For certain feature attributes that need to be calculated, such as order processing time, calculation is performed based on relevant fields in the basic order information. The extracted, encoded and calculated feature values ​​are combined to form a feature attribute set for each order.

[0019] Through the above steps, N order feature attribute sets are obtained, each set corresponds to an order and contains all feature values ​​of the order under the preset feature attribute set.

[0020] Step S400: using the K identification branches of the order identifier to identify the N order feature attribute sets to obtain M call order clusters, wherein the M call order clusters have M processing fitnesses; In the embodiment of the present application, N order feature attribute sets are loaded into the order identifier for identification processing. The order identifier matches the N order feature attribute sets with the historical order feature attribute sets in each identification branch, and calculates the similarity between the order feature attribute sets and the historical order feature attribute sets stored in the K identification branches one by one through cosine similarity.

[0021] When the similarity between an order feature attribute set and a certain identification branch, such as a historical order feature attribute set stored in the first identification branch, meets a preset similarity threshold, the order feature attribute set is stored in an internal node corresponding to the identification branch.

[0022] As the order feature attribute sets are assigned to each internal node, similar orders will gradually gather together to form a preliminary order cluster. When all N order feature attribute sets are assigned, each internal node will contain a set of order feature attribute sets with similar characteristics, forming a call order cluster.

[0023] The processing fitness of each call order cluster is determined based on the processing fitness of the identification branch corresponding to the cluster. By setting the processing fitness of each cluster to the processing fitness of the corresponding identification branch, an indicator is provided for each order cluster to measure its processing complexity and required resources.

[0024] Finally, after the above steps, M call order clusters are obtained, each cluster contains a set of order feature attributes with similar characteristics and has corresponding processing fitness.

[0025] Step S500: configuring M processing units of the order processing module according to the M processing fitnesses, and sending M call order clusters to the M processing units for order processing.

[0026] In the embodiment of the present application, the order processing module includes multiple processing units, which have the same processing capabilities and resource configurations in the initial state. According to the processing adaptability of the call order cluster, the processing requirements, efficiency requirements and resource consumption of each cluster are analyzed. According to the analysis results, a corresponding processing unit is configured for each call order cluster. The configuration includes adjusting the resource allocation of the processing unit, setting the processing priority, optimizing the processing flow, etc.

[0027] When allocating resources, the resources required by the processing unit are reasonably allocated according to the resource utilization rate and expected resource consumption in the processing fitness, ensuring that the resources are effectively utilized during the processing. When setting the processing priority, different processing priorities are set for the processing unit according to the time weight and order urgency in the processing fitness, ensuring that urgent or important orders can be processed in a timely manner. When optimizing the processing flow, the processing flow of the processing unit is optimized according to the characteristics and attributes of the call order cluster to improve the processing efficiency.

[0028] Further, such as Figure 2 As shown, step S200 in the method provided in the application embodiment also includes: Extracting historical data from the call records of the order receiving module to obtain multiple historical call order data sets, wherein the multiple historical call order data sets include multiple historical order basic information and multiple historical processing fitness levels; Extracting attributes from the plurality of historical call order data sets using the preset characteristic attribute set as an index to obtain a plurality of historical order characteristic attribute sets; Based on the multiple historical order feature attribute sets and the multiple historical processing fitness levels, similar identification is performed to construct the order identifier.

[0029] In the embodiment of the present application, historical data is extracted from the call records of the order receiving module, and historical data related to order processing is screened out from the call records, including historical order basic information and historical processing fitness. After screening and cleaning, the obtained historical data is organized into multiple historical call order data sets. Each set contains complete historical order basic information and corresponding historical processing fitness.

[0030] The preset feature attribute set is used as an index to extract the corresponding feature attributes from the historical call order data set. After the attribute extraction, multiple historical order feature attribute sets are obtained. Each set contains historical order features corresponding to the preset feature attribute set.

[0031] According to the historical order feature attribute set and historical processing fitness, select a suitable similar identification algorithm. For example, use cluster analysis, classification algorithm, etc. to identify orders with similar features and processing fitness. Apply the selected similar identification algorithm to process the historical order feature attribute set and identify the order categories with similar features. Based on the results of similar identification, build an order identifier. The order identifier contains multiple identification branches, each branch corresponds to an identified order category. Each branch stores the historical order feature attribute set and the corresponding historical processing fitness of the category.

[0032] Further, based on the plurality of historical order feature attribute sets and the plurality of historical processing fitnesses, the same type of identification is performed to construct the order identifier, and the method further includes: randomly extracting a first historical order feature attribute set from the plurality of historical order feature attribute sets without replacement and storing the set in a first identification branch, wherein the first identification branch has a first internal node and a first external node, and the first identification branch is used to perform binary classification on the plurality of historical order feature attribute sets; Using the first identification branch to identify the plurality of historical order feature attribute sets, determining whether the similarity between the plurality of historical order feature attribute sets and the first historical order feature attribute set stored in the first identification branch meets a preset similarity threshold, and if so, adding them into the first internal node, and if not, adding them into the first external node; A second historical order feature attribute set is randomly selected from the multiple historical order feature attribute sets of the first external node without replacement and stored in a second identification branch, wherein the second identification branch has a second internal node and a second external node, and the second identification branch is used to perform binary classification on the multiple historical order feature attribute sets of the first external node.

[0033] In an embodiment of the present application, a historical order feature attribute set is randomly extracted from multiple historical order feature attribute sets without replacement as an initial classification basis. The extracted historical order feature attribute set is stored in a first identification branch. The first identification branch is the starting branch of the order identifier, which is used to perform preliminary binary classification on subsequent multiple historical order feature attribute sets. Each identification branch has internal nodes and external nodes. The internal nodes store historical order feature attribute sets similar to the branch, and the external nodes store historical order feature attribute sets that are not similar to the branch.

[0034] Using the historical order feature attribute set stored in the first identification branch, the remaining multiple historical order feature attribute sets are identified by calculating the similarity using cosine similarity. For each historical order feature attribute set, determine whether its similarity with the historical order feature attribute set stored in the first identification branch meets the preset similarity threshold. The preset similarity threshold is set according to business requirements and order characteristics, and is used to determine which orders should be classified into the same branch. If the similarity meets the threshold, it means that the historical order feature attribute set is similar to the set in the first identification branch, and it is added to the first internal node. If the threshold is not met, it means that it is not similar, and it is added to the first external node.

[0035] When the number of historical order feature attribute sets of the first external node reaches a certain level, such as the number exceeds a certain preset value, one is randomly selected from the multiple historical order feature attribute sets of the first external node without replacement as a new classification basis. The extracted historical order feature attribute set is stored in a new identification branch, that is, stored in the second identification branch, and similarly processed. The second identification branch is used to further classify the multiple historical order feature attribute sets of the first external node into two categories.

[0036] This process is repeated recursively, each time extracting a new set of historical order feature attributes from the external nodes as a new classification basis, creating a new identification branch, and further classifying the remaining sets until all historical order feature attribute sets are classified into the corresponding internal nodes, or other stopping conditions are met, such as the number of branches reaches the upper limit, the classification effect is no longer improved, etc.

[0037] Furthermore, the method further comprises: Randomly extracting the K-1th historical order feature attribute set from the multiple historical order feature attribute sets of the K-2th external node and storing it in the K-1th identification branch without replacement, wherein the K-1th identification branch has the K-1th internal node and the K-1th external node, and the K-1th identification branch is used to perform binary classification on the multiple historical order feature attribute sets of the K-2th external node; storing the plurality of historical order feature attribute sets in the K-1th external node into the Kth internal node of the Kth identification branch; The order identifier is constructed according to the first identification branch, the second identification branch, the K-1th identification branch, the Kth identification branch, the first internal node, the second internal node, the K-1th internal node and the Kth internal node.

[0038] In the embodiment of the present application, when the K-1th round is reached, one of the multiple historical order feature attribute sets of the external nodes of the K-2th round is randomly selected without replacement as the K-1th historical order feature attribute set. This set is used as a new classification basis to further classify the set in the external nodes of the K-2th round.

[0039] Create a new identification branch, namely the K-1 identification branch. This branch has two nodes, the K-1 internal node and the K-1 external node. The K-1 internal node is used to store the historical order feature attribute set that is similar to the K-1 historical order feature attribute set, while the K-1 external node is used to store the dissimilar set.

[0040] Using the K-1 recognition branch, calculate the similarity between each set and the K-1 historical order feature attribute set, and classify them into the K-1 internal node or the K-1 external node according to the preset similarity threshold. Perform binary classification on multiple historical order feature attribute sets in the K-2 external node.

[0041] After a round of classification, the stopping condition is reached, such as the number of historical order feature attribute sets of the K-1th external node is less than a preset value, or the classification effect is no longer significantly improved. At this time, no new recognition branches are created, but the sets in the current external node are stored in the internal nodes of the next recognition branch.

[0042] When the stop condition is met, multiple historical order feature attribute sets in the K-1th external node are stored in the Kth internal node of the Kth identification branch. The Kth identification branch is the last identification branch and is used to store those historical order feature attribute sets that cannot be further subdivided.

[0043] Finally, the first recognition branch, the second recognition branch, ..., the K-1th recognition branch and the Kth recognition branch are combined to form an order recognition structure.

[0044] Furthermore, the method further comprises: Acquire a plurality of first processing fitnesses corresponding to a plurality of historical order feature attribute sets in the first internal node; Calculating a first mean value of the plurality of first processing fitnesses to generate a first concentration point; Calculating a first variance of the plurality of first processing fitnesses, and comparing the first variance with a preset variance, and multiplying the ratio with a preset concentration step length to obtain a first concentration step length; Counting the number of historical order feature attribute sets in an area with the first concentration point as the center and the first concentration step length as the radius to obtain a first concentration quantity; The second concentrated point is obtained by moving in any direction with the first concentrated point as a starting point according to the first concentrated step length.

[0045] In the embodiment of the present application, in each internal node of the order identifier, a set of historical order feature attributes corresponding to the node is stored, and these sets are associated with respective processing fitness. These historical order feature attribute sets are taken out from the first internal node, and their respective corresponding first processing fitness are obtained.

[0046] The mean of the first processing fitness corresponding to the characteristic attribute set of all historical orders in the first internal node is calculated to obtain a value representing the average processing effect of the orders in the node, that is, the first concentration point.

[0047] After the calculation of the first mean is completed, the deviation of each processing fitness from the mean is calculated, the square values ​​of these deviations are added together, and the sum is divided by the number of processing fitness to obtain the first variance. The calculated first variance is compared with the preset variance to obtain a ratio. This ratio reflects the relative relationship between the actual degree of discreteness of the processing fitness and the expected degree of discreteness. The preset variance is a pre-set threshold that reflects the fluctuation range of the processing fitness expected in the business scenario or order processing system. Finally, the ratio is multiplied by the preset concentration step size to obtain the first concentration step size. The preset concentration step size is a parameter for the moving or exploration step size in the quantitative analysis process.

[0048] A circular area is determined with the first concentration point as the center and the first concentration step length as the radius. Next, the number of historical order feature attribute sets in the area is counted, and the counted number is called the first concentration quantity.

[0049] The first concentrated point is taken as the starting point, and the first concentrated step is taken as the moving step to move in any direction, and the new point after the move is called the second concentrated point.

[0050] Furthermore, the method further comprises: Determine whether the first concentration amount of the first concentration point is greater than the second concentration amount of the second concentration point, and if so, update the second concentration point as the starting point; If not, a first flight factor is obtained based on a flight factor generator. When the first flight factor is greater than a preset value, the first concentrated point is taken as a starting point, and the first concentrated step length is followed to move in any direction except the second concentrated point to obtain a third concentrated point. When the first flight factor is less than or equal to a preset value, updating the second concentration point as a starting point; After the number of movements meets the preset number of times, the first processing fitness corresponding to the maximum value of the concentration during the movement is used as the processing fitness of the first internal node.

[0051] In the embodiment of the present application, first, a circular area is determined with the second concentration point as the center and the first concentration step as the radius. Next, the number of historical order feature attribute sets in the area is counted, and the counted number is called the second concentration amount. The order processing fitness concentration amounts around the first concentration point and the second concentration point are compared. If the first concentration amount of the first concentration point is greater than the second concentration amount of the second concentration point, it means that there are more orders with similar processing fitness around the first concentration point, so the second concentration point is updated as the starting point.

[0052] If the first concentration amount of the first concentration point is not greater than the second concentration amount of the second concentration point, the first flight factor is obtained based on the flight factor generator. The flight factor generator is a function specifically used to generate a flight factor. It generates a suitable flight factor based on a preset rule or algorithm, combined with the current search history, processing fitness distribution and other information.

[0053] When the first flight factor is greater than the preset value, it means that the current search state is more inclined to explore new areas. Therefore, starting from the first concentration point, move in any direction except the second concentration point according to the first concentration step length to obtain the third concentration point. Avoid repeated search of the area near the second concentration point that has been explored, and expand the search range to find other possible concentration areas.

[0054] When the first flight factor is less than or equal to the preset value, it means that the current search state is more inclined to use the existing information. Therefore, the second focus point is updated as the starting point.

[0055] To avoid infinite loops or excessive searches, a preset number of moves is set. When the number of moves reaches this preset value, the search process is stopped, and the processing fitness of the first internal node is determined based on the information recorded during the move. Specifically, the concentrations obtained during all moves are compared to find the first processing fitness corresponding to the maximum concentration. This processing fitness best represents the overall processing effect of the orders in the internal node, so it is used as the processing fitness of the first internal node.

[0056] Furthermore, the method further comprises: The flight factor generator includes a flight factor generation formula, wherein the flight factor generation formula is: ; in, is the flight factor, It is a random value generated by the rand function. .

[0057] In the embodiment of the present application, by introducing a random number By combining the sine function and the adjustment of the exponential function, the formula generates a series of flight factors with different characteristics and behaviors, which helps the search algorithm find a balance between exploration and exploitation and improves the search efficiency and effect.

[0058] Furthermore, step S500 in the method provided in the application embodiment also includes: Interactively obtain computing resources of the order processing module to obtain available computing resources; Calculate the ratio of the M processing fitnesses to the sum of the M processing fitnesses respectively, and multiply the ratio by the available computing power resources to obtain M computing power allocation results; The M processing units of the order processing module are configured based on the M computing power allocation results.

[0059] In the embodiment of the present application, the order processing module is interacted with to obtain the available computing resource information. The received computing resource information is analyzed to understand the hardware configuration and performance of the current module. Factors such as the task load of the current order processing module, the occupancy of other modules, and system maintenance requirements are considered. Taking the above factors into consideration, the current available computing power resources are calculated.

[0060] Each processing unit is evaluated for fitness to determine its efficiency and ability to process specific order tasks. The processing fitness of each processing unit is compared with the sum of the processing fitness of all processing units to obtain M ratios. The calculated ratios are multiplied by the available computing resources to obtain the amount of computing resources that the M processing units should obtain.

[0061] Finally, based on the M computing power allocation results obtained, a corresponding number of processor cores, memory, and storage resources are allocated to each processing unit to ensure that they can work according to the expected computing power requirements.

[0062] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects: The order receiving module of the interactive call center of the present application extracts N call orders within a preset time window and generates N basic order information; extracts historical data from the call records of the order receiving module, and analyzes the historical order attributes in combination with the preset feature attribute set, and constructs an order identifier according to the analysis result, wherein the order identifier has K identification branches, and the K identification branches respectively store K historical order feature attribute sets, and have K processing fitness; analyzes the order attributes of the N basic order information with the preset feature attribute set to obtain N order feature attribute sets; uses the K identification branches of the order identifier to identify the N order feature attribute sets to obtain M call order clusters, wherein the M call order clusters have M processing fitness; configures the M processing units of the order processing module according to the M processing fitness, and sends the M call order clusters to the M processing units for order processing. The present invention solves the technical problem that the prior art is difficult to cope with complex and changeable customer needs and the ever-changing market environment, resulting in poor order processing flexibility and low processing efficiency, and achieves the technical effect of efficient and accurate processing of call center orders through data analysis.

[0063] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. Other embodiments are within the scope of the attached claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0064] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0065] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A call center order integration processing method based on data analysis, characterized in that: The method comprises: The order receiving module of the interactive call center extracts N call orders within a preset time window and generates N basic order information; Extracting historical data from the call records of the order receiving module, and analyzing historical order attributes in combination with a preset feature attribute set, and constructing an order identifier according to the analysis result, wherein the order identifier has K identification branches, each of which stores K historical order feature attribute sets and has K processing fitnesses; Performing order attribute analysis on the N basic order information using the preset characteristic attribute set to obtain N order characteristic attribute sets; Using the K identification branches of the order identifier to identify the N order feature attribute sets, obtaining M call order clusters, wherein the M call order clusters have M processing fitnesses; M processing units of the order processing module are configured according to the M processing fitnesses, and the M call order clusters are sent to the M processing units for order processing.

2. The method according to claim 1, characterized in that Extracting historical data from the call records of the order receiving module, analyzing historical order attributes in combination with a preset feature attribute set, and constructing an order identifier based on the analysis results, the method includes: Extracting historical data from the call records of the order receiving module to obtain multiple historical call order data sets, wherein the multiple historical call order data sets include multiple historical order basic information and multiple historical processing fitness levels; Extracting attributes from the plurality of historical call order data sets using the preset characteristic attribute set as an index to obtain a plurality of historical order characteristic attribute sets; Based on the multiple historical order feature attribute sets and the multiple historical processing fitness levels, similar identification is performed to construct the order identifier.

3. The method according to claim 2, characterized in that Based on the multiple historical order feature attribute sets and the multiple historical processing fitnesses, the same type of identification is performed to construct the order identifier, and the method includes: randomly extracting a first historical order feature attribute set from the plurality of historical order feature attribute sets without replacement and storing the set in a first identification branch, wherein the first identification branch has a first internal node and a first external node, and the first identification branch is used to perform binary classification on the plurality of historical order feature attribute sets; Using the first identification branch to identify the plurality of historical order feature attribute sets, determining whether the similarity between the plurality of historical order feature attribute sets and the first historical order feature attribute set stored in the first identification branch meets a preset similarity threshold, and if so, adding them into the first internal node, and if not, adding them into the first external node; A second historical order feature attribute set is randomly selected from the multiple historical order feature attribute sets of the first external node without replacement and stored in a second identification branch, wherein the second identification branch has a second internal node and a second external node, and the second identification branch is used to perform binary classification on the multiple historical order feature attribute sets of the first external node.

4. The method according to claim 3, characterized in that The method comprises: Randomly extracting the K-1th historical order feature attribute set from the multiple historical order feature attribute sets of the K-2th external node and storing it in the K-1th identification branch without replacement, wherein the K-1th identification branch has the K-1th internal node and the K-1th external node, and the K-1th identification branch is used to perform binary classification on the multiple historical order feature attribute sets of the K-2th external node; storing the plurality of historical order feature attribute sets in the K-1th external node into the Kth internal node of the Kth identification branch; The order identifier is constructed according to the first identification branch, the second identification branch, the K-1th identification branch, the Kth identification branch, the first internal node, the second internal node, the K-1th internal node and the Kth internal node.

5. The method according to claim 4, characterized in that The method comprises: Acquire a plurality of first processing fitnesses corresponding to a plurality of historical order feature attribute sets in the first internal node; Calculating a first mean value of the plurality of first processing fitnesses to generate a first concentration point; Calculating a first variance of the plurality of first processing fitnesses, and comparing the first variance with a preset variance, and multiplying the ratio with a preset concentration step length to obtain a first concentration step length; Counting the number of historical order feature attribute sets in an area with the first concentration point as the center and the first concentration step length as the radius to obtain a first concentration quantity; The second concentrated point is obtained by moving in any direction with the first concentrated point as a starting point according to the first concentrated step length.

6. The method according to claim 5, characterized in that The method comprises: Determine whether the first concentration amount of the first concentration point is greater than the second concentration amount of the second concentration point, and if so, update the second concentration point as the starting point; If not, a first flight factor is obtained based on a flight factor generator. When the first flight factor is greater than a preset value, the first concentrated point is taken as a starting point, and the first concentrated step length is followed to move in any direction except the second concentrated point to obtain a third concentrated point. When the first flight factor is less than or equal to a preset value, updating the second concentration point as a starting point; After the number of movements meets the preset number of times, the first processing fitness corresponding to the maximum value of the concentration during the movement is used as the processing fitness of the first internal node.

7. The method according to claim 6, characterized in that The method comprises: The flight factor generator includes a flight factor generation formula, wherein the flight factor generation formula is: ； in, is the flight factor, It is a random value generated by the rand function. .

8. The method according to claim 1, characterized in that The method comprises: Interactively obtain computing resources of the order processing module to obtain available computing resources; Calculate the ratio of the M processing fitnesses to the sum of the M processing fitnesses respectively, and multiply the ratio by the available computing power resources to obtain M computing power allocation results; The M processing units of the order processing module are configured based on the M computing power allocation results.

Images