Call center operation analysis system and performance evaluation method based on big data

Abstract

The invention discloses a call center operation analysis system and performance evaluation method based on big data, and relates to the field of call center operation analysis systems. A score assignment mechanism and a score assignment-based utilization rate calculation mechanism are provided, the utilization rate of a call voice target is accurately calculated, when the utilization rate is smaller than or equal to a threshold value Q, a related data packet x is automatically screened and sent to a technician, and the technician can perform auxiliary judgment on a current problem reason by using description of historical customers of the same problem in historical data; technicians are assisted in solving and judging problems, communication between the technicians and clients with weak professional level and poor expression ability is reduced, and the working efficiency of the technicians is improved.

Description

Technical Field

[0001] The present invention relates to the field of call center operation analysis systems, and particularly to a call center operation analysis system and a performance evaluation method based on big data. Background Art

[0002] With the rapid development of information technology, call centers have become an important bridge for communication between enterprises and customers; traditional call center operation models mainly rely on manual records and analysis, which are not only inefficient but also difficult to deeply explore the potential value in call voices; with the rise of big data technology, call centers have begun to explore how to use big data means to improve operation efficiency and service quality;

[0003] In existing systems, call voices are often simply recorded and analyzed, lacking in-depth mining and utilization; at the same time, when technicians solve problems, they often can only rely on a single call record, lacking comprehensive data support and reference. In particular, in the prior art, technicians' judgment of problems often relies on the voices of conversations between operators and customers, and they may also communicate with customers actively to judge problems. However, there are a certain number of customers with relatively weak professional levels and poor expression abilities, so continuous voice communication has a poor effect on problem-solving. In such cases, only the communication voices between operators and these customers can be relied on for judgment, resulting in a significant reduction in judgment accuracy and thus a significant reduction in problem-solving efficiency. Summary of the Invention

[0004] To make up for the deficiencies of existing technical problems, the purpose of the present invention is to construct a database X based on historical data, automatically calculate the utilization rate LYL, and when the utilization rate LYL of the call voice 目标 is less than or equal to the threshold Q, automatically retrieve the relevant data packet x and provide it to the technician, so as to improve the accuracy of the technician's problem judgment and the efficiency of problem-solving, a call center operation analysis system and a performance evaluation method based on big data.

[0005] To solve the existing technical problems, the technical solution of the present invention is as follows:

[0006] A call center operation analysis system based on big data includes an agent terminal, a technician terminal, a storage terminal, and a processor module. The operator communicates with the customer through the agent terminal, and the call voice 目标 is recorded by the agent terminal and scored and calculated by the processor module. The scoring calculation result and the call voice are stored in the storage terminal. The operator and the technician communicate with each other through the agent terminal and the technician terminal respectively, and specifically include the following steps:

[0007] A1. The customer dials a call and accesses the agent terminal, and the call voice between the operator and the customer 目标It is collected in real time by the agent terminal, and the operator solves the problems feedback by the customer through the agent terminal;

[0008] A2. When the operator fails to solve the customer's problem, the processor module will send the call voice 目标 and the work order recorded by the operator 目标 to the technician terminal, and the technician will continue to solve the problem. The technician can view the call voice through the technician terminal 目标 to understand the problem situation and make a call to the customer through the technician terminal;

[0009] A3. The processor module analyzes the call voice 目标 , assigns a score to the call voice 目标 , and calculates the utilization rate of the call voice 目标 . When the utilization rate is less than or equal to the threshold Q, the processor module analyzes the work order 目标 to determine the customer problem type of the call voice 目标 , and then screens the relevant data packet x in the storage terminal and sends it to the technician terminal for the technician to use in problem solving;

[0010] A4. After the problem is solved, the processor module packs the call voice 目标 , the work order 目标 , the score data and the utilization rate into a data packet x and stores it in the storage terminal to construct a database X=(x 1 , x 2 ...x h ), where x h represents the hth data packet in the database X.

[0011] Preferably, the content of the work order includes at least the operator's identity information, the answering and calling time, the customer problem type, and the customer product information.

[0012] Preferably, the utilization rate of the call voice 目标 is calculated by the following formula:

[0013] LYL=(LS*a + WL*b + FH*c)*m + FK*n + FL*i;

[0014] Where, LYL represents the utilization rate, FK represents the clarity score of the call voice 目标 , FL represents the professionalism score of the call voice 目标 , m, n, and i are weights, and m + n + i = 100%, LS represents the total amount of data packets x belonging to the operator in the database X, WL represents the data packets in the database X that belong to the operator and the customer problem type in the work order 存储 is the same as that of the call voice 目标 in the work order 目标The number of data packets x with the same customer problem type, FH represents the call voice 目标 The business level score, a, b, and c are weights, and a + b + c = 100%.

[0015] Preferably, the expression of the data packet x is as follows:

[0016] x = (TH, GD, LYL, FH, FK, FL)

[0017] Where TH represents the call voice, GD represents the work order, LYL represents the utilization rate, FH represents the business level score of the call voice, FK represents the clarity score of the call voice, and FL represents the professional degree score of the call voice.

[0018] Preferably, the specific steps for screening relevant data packets x in the storage terminal in step A3 are as follows:

[0019] C1. The processor module determines the work order 目标 The customer problem type in it, and retrieves from the database X the work orders 目标 with the same customer problem type as in the work order 存储 ;

[0020] C2. Arrange all the retrieved work orders 存储 in descending order according to the LYL value of the utilization rate in the data packet x to which the work order 存储 belongs, and select the top G work orders in the arrangement 存储 for marking;

[0021] C3. Extract and send the data packet x to which the marked work order 存储 belongs to the technician terminal.

[0022] Preferably, it further includes the following steps:

[0023] A5. Set the learning period T. When the learning period T ends, the processor module calculates the average score LP of the operator, arranges all the operators in descending order according to the average score LP value, marks the top Y operators in the arrangement as excellent operators, and marks the last U operators in the arrangement as operators to be learned;

[0024] A6. Mark the data packet x of all excellent operators within the learning period T as data packet x 优等 , arrange all the data packets x 优等 in descending order according to the FH value of the business level score, select the top E data packets x in the arrangement 优 等 to pack and obtain the training package PX, and send the training package PX to the seat terminals of all operators to be learned for the operators to be learned to consult and learn.

[0025] Preferably, the steps for calculating the average score LP are as follows:

[0026] B1. The processor module retrieves the data packets x belonging to the operator within the learning cycle T in the storage terminal and extracts the service level score FH within the data packets x;

[0027] B2. Calculate the average score LP according to the following formula:

[0028]

[0029] where FH k represents the k-th extracted service level score FH, and r represents the number of extracted service level scores FH.

[0030] A performance evaluation method uses the above-mentioned call center operation analysis system based on big data to complete the performance evaluation work of the operator, specifically including the following steps:

[0031] D1. Set the evaluation cycle J. The processor module retrieves the database X and obtains the number SL of data packets x belonging to the operator in the database X within the evaluation cycle J. Calculate the performance evaluation score JX according to the following formula:

[0032]

[0033] where ZT represents the total number of calls of the operator recorded by the agent terminal within the evaluation cycle J, and Z and V are weights.

[0034] Compared with the prior art, the advantages of the present invention are as follows:

[0035] 1. By using historical data to construct the database X and providing a scoring mechanism and a utilization rate LYL calculation mechanism based on the score, accurately calculate the utilization rate LYL of the call voice. When the utilization rate LYL is less than or equal to the threshold Q, automatically screen the relevant data packets x and send them to the technician. The technician can then use the description voice of the historical customers with the same problem in the historical data to assist in judging the cause of the current problem, assist the technician in problem-solving and judgment work, reduce the communication between the technician and customers with relatively weak professional levels and poor expression abilities, and improve the work efficiency of the technician; 目标 2. By using the score, the service level of the operator can be summarized periodically, and the training package PX can be formulated using the database X to promote the learning of operators with relatively low service levels, providing convenience for the learning and communication between operators. This not only helps to improve the service level of the operators to be learned, but also promotes the knowledge sharing and common progress of the entire call center team;

[0036]

[0037] ​3. The performance evaluation method formulated by using score assignment makes the performance evaluation more fair and just, links the performance evaluation with the business level of the operator, and further achieves the effect of promoting the business level of the operator. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the system of the present invention.

[0039] Figure 2 It is a schematic diagram of the method of the present invention.

[0040] Figure 3 It is a logical schematic diagram of the present invention.

[0041] Figure 4 It is a schematic diagram of the second embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0042] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0043] Embodiment 1. Please refer to Figures 1 to 3 , this embodiment provides an operation analysis system for a call center based on big data, including an agent terminal, a technician terminal, a storage terminal, and a processor module. The operator communicates with the customer through the agent terminal. Based on communication technology and call processing technology, when a customer calls, the call is automatically assigned to a suitable agent terminal by the processor module using call processing technology;

[0044] Call voice 目标 is recorded by the agent terminal and scored by the processor module. The scoring calculation result and the call voice are stored in the storage terminal. The processing and analysis of the call voice are realized based on voice processing and recognition technology; the operator and the technician communicate with each other through the agent terminal and the technician terminal respectively;

[0045] Specifically, it includes the following steps:

[0046] A1. The customer dials a call to access the agent terminal, and the call voice between the operator and the customer 目标 is collected in real time by the agent terminal, and the operator solves the problems feedback by the customer through the agent terminal;

[0047] During the background maintenance process, a problem solution library is sorted out based on the common solutions to common problems in the call center and stored in the storage terminal. When communicating with the customer, the operator can log in to the storage terminal to consult the solution library and provide the customer with a problem solution to seek the possibility of solving the customer's problem during the call and reduce the participation of technicians;

[0048] A2. When the operator fails to solve the customer's problem, the processor module will send the call voice 目标 and the work order recorded by the operator 目标 to the technician's terminal. The technician will continue to solve the problem. The technician can view the call voice through the technician's terminal 目标 to understand the problem situation, and can make a call to the customer through the technician's terminal. The content of the work order shall at least include the operator's identity information, the call time, the type of customer problem, and the customer's product information;

[0049] A3. The processor module analyzes the call voice 目标 and assigns scores to the call voice 目标 ;

[0050] Specifically, the scores are assigned from three aspects. The processor assigns a business level score to the operator based on the operator's voice content in the call voice 目标 , assigns a clarity score and a professionalism score to the customer's statement based on the customer's voice content in the call voice 目标 , so as to obtain the professionalism score FL, clarity score FK and business level score FH of the call voice 目标 ;

[0051] The processor module can perform the score assignment work through speech recognition and natural language processing (NLP) technology, which can analyze the call voice, understand the conversation content, and respectively assign scores to the business level, statement clarity and statement professionalism of the operator's and customer's voice content according to the preset standards. Taking speech recognition technology as an example, the implementation method is as follows:

[0052] 1. Speech collection and preprocessing: Collect the call voice and preprocess the collected voice, such as denoising, enhancement, etc., to improve the accuracy of subsequent speech recognition;

[0053] 2. Speech recognition: Use the speech recognition engine to convert the preprocessed speech signal into text information. This process includes steps such as speech signal modeling, feature extraction, model training and matching;

[0054] 3. Natural language processing: Perform natural language processing on the recognized text information, including word segmentation, part-of-speech tagging, syntactic analysis, etc. Use natural language processing technology to semantically understand the conversation content and identify key information, such as conversation topic, sentiment tendency, etc.;

[0055] 4. Score assignment standard formulation: According to business requirements, formulate score assignment standards for the business level, statement clarity and statement professionalism of the operator's and customer's voice content; These standards can include multiple aspects such as the fluency, accuracy, and professionalism of language expression;

[0056] 5. Scoring implementation: Evaluate the conversation content using natural language processing technology according to the established scoring criteria; calculate the scores of various items in the voice content of the operator and the customer through algorithms;

[0057] 6. Result output.

[0058] The processor module calculates the utilization rate of the call voice 目标 When the utilization rate is less than or equal to the threshold Q, the processor module analyzes the work order 目标 to determine the customer problem type of the call voice 目标 and accordingly filters the relevant data packet x in the storage terminal and sends it to the technician terminal for the technician to use to solve the problem;

[0059] The utilization rate calculation steps of the call voice 目标 are as follows:

[0060] The processor module determines the operator identity information by analyzing the work order 目标 , and then retrieves the total amount of data packet x belonging to the operator from the database X based on the operator identity information. Continuing to analyze the work order 目标 , it determines the number of data packet x in the database that belongs to the operator and has the same customer problem type as the work order 目标 in the call voice. Substitute the data into the utilization rate of the call voice 目标 and calculate it using the following formula:

[0061] LYL = (LS * a + WL * b + FH * c) * m + FK * n + FL * i;

[0062] Among them, LYL represents the utilization rate, FK represents the clarity score of the call voice 目标 , FL represents the professionalism score of the call voice 目标 , m, n, and i are weights, and m + n + i = 100%, LS represents the total amount of data packet x belonging to the operator in database X, WL represents the number of data packet x in database X that belongs to the operator and has the same customer problem type as the work order 存储 in the call voice, 目标 FH represents the business level score of the call voice 目标 , a, b, and c are weights, and a + b + c = 100%; 目标 The part (LS * a + WL * b + FH * c) * n reflects the influence of operator-related factors on the utilization rate LYL of the call voice in the entire utilization rate LYL calculation;

[0063] (LS * a + WL * b + FH * c) * n reflects the influence of operator-related factors on the call voice utilization rate LYL in the entire utilization rate LYL calculation;

[0064] The part of FK*n + FL*i represents the contribution degree of customer clarity in the call voice utilization rate LYL and the contribution degree of customer professionalism in the call voice utilization rate LYL;

[0065] By reasonably setting weights and adjustment factors, the value and usage effect of call voice can be accurately evaluated, which is of great significance for improving customer service quality, optimizing operator training, and improving call voice processing processes, etc.;

[0066] Screening relevant data packets x specifically includes the following steps:

[0067] C1. The processor module determines the work order 目标 the customer problem type in, and retrieves from database X the work orders 目标 with the same customer problem type as in the work order 存储 ;

[0068] C2. Arrange all the retrieved work orders 存储 in descending order according to the LYL numerical value of the utilization rate within the data packet x to which the work order 存储 belongs, and select the top G work orders in the arrangement 存储 for marking;

[0069] C3. Extract and send the marked work orders and their belonging data packet x to the technician terminal.

[0070] A4. After the problem is solved, the processor module packs the call voice 目标 , work order 目标 , scoring data and utilization rate into a data packet x and stores it in the storage terminal to construct database X = (x 1 , x 2 ... x h ), where x h represents the hth data packet in database X;

[0071] The expression of the data packet x is as follows:

[0072] x = (TH, GD, LYL, FH, FK, FL);

[0073] Among them, TH represents the call voice, GD represents the work order, LYL represents the utilization rate, FH represents the business level score of the call voice, FK represents the clarity score of the call voice, and FL represents the professionalism score of the call voice.

[0074] Example illustration:

[0075] Obtain the call voice of operator 1 communicating with customer 1 at the current time 1-1 , and the call voice 1-1 is scored by the processor module to obtain FH 1-1= 85, FK 1-1 = 50, FL 1-1 = 40;

[0076] The processor module determines the total number LS of data packets x belonging to operator 1 by retrieving the work orders in all data packets x in database X 存储 , and LS = 65 for the data packets x belonging to operator 1, and for the work orders 1 where the customer problem type in the work order 存储 is the same as the call voice 1-1 in the work order 1-1 , the number WL of data packets x with the same customer problem type is WL 1-1 = 35;

[0077] Set a = 20%, b = 20%, c = 60%, m = 30%, n = 35%, i = 35%, and substitute the above data into the utilization rate LYL calculation formula to calculate as follows:

[0078] LYL 1-1 = (65 * 20% + 35 * 20% + 85 * 60%) * 30% + 50 * 35% + 40 * 35%

[0079] = 52.8

[0080] Set the threshold Q = 75, and since LYL 1-1 < Q, the processor module needs to perform screening work on relevant data packets. An example of the screening work is as follows:

[0081] Assume that the work orders retrieved from the database in database X that have the same customer problem type as the work order 1-1 and the LYL values of the data packets x to which they belong are as shown in Table 1 below: 存储 Table 1: Numerical comparison table

[0082] Table 1: Numerical comparison table

[0083]

[0084] Arrange the work orders in Table 1 above in descending order according to the LYL values to obtain the arrangement as shown in Table 2 below: 存储 Table 2: Arrangement table

[0085] Table 2: Arrangement table

[0086]

[0087] Set the G value to 5, then mark and ;

[0088] Subsequently, extract the data packets x to which the marked work orders 存储 belong and send them to complete the screening and sending work;

[0089] In summary, by using historical data to construct database X and providing a scoring mechanism and a utilization rate calculation mechanism based on scoring, the utilization rate LYL of call voice is accurately calculated, and relevant data packets x are automatically screened and sent to technicians to assist technicians in problem-solving and judgment work, reducing the communication between technicians and customers with weak professional levels and poor expression abilities, and improving the work efficiency of technicians. 目标 The utilization rate LYL of is automatically screened and relevant data packets x are sent to technicians to assist technicians in problem-solving and judgment work, reducing the communication between technicians and customers with weak professional levels and poor expression abilities, and improving the work efficiency of technicians.

[0090] Example 2, please refer to Figure 4 , this embodiment provides a further technical solution based on Embodiment 1, and further includes the following steps:

[0091] A5. The processor module periodically calculates the average score LP of operators. Set the learning period T. When the learning period T ends, the processor module calculates the average score LP of operators, sorts all operators in descending order according to the value of the average score LP, marks the top Y operators in the ranking as excellent operators, and marks the last U operators in the ranking as operators to be learned;

[0092] If the learning period T is set to 1 month, the calculation of the average score LP is performed once every 1 month;

[0093] A6. Mark the data packets x of all excellent operators within the learning period T as data packets x 优等 , and all data packets x 优等 are sorted in descending order according to the value of the business level score FH, and the top E data packets x in the ranking are selected 优等 to be packaged into a training package PX, and the training package PX is sent to the seat terminals of all operators to be learned for the operators to be learned to consult and learn.

[0094] The steps for calculating the average score LP are as follows:

[0095] B1. The processor module retrieves the data packets x belonging to the operator within the learning period T in the storage terminal and extracts the business level score FH within the data packets x;

[0096] B2. Calculate the average score LP according to the following formula:

[0097]

[0098] where FH k represents the k-th extracted business level score FH, and r represents the number of extracted business level scores FH.

[0099] Example illustration:

[0100] It is set that within the retrieval learning period T of operator 1, data packet x has data packet x 22 , data packet x 25 , data packet x 28 , data packet x 33 , data packet x 40 , data packet x 50 and data packet x 51 . The service level scores FH within each data packet x are extracted as shown in Table 3 below:

[0101] Table 3: Service level score FH

[0102] <![CDATA[Data packet x 22 > <![CDATA[Data packet x 25 > <![CDATA[Data packet x 28 > <![CDATA[Data packet x 33 > <![CDATA[Data packet x 40 > <![CDATA[Data packet x 50 > <![CDATA[Data packet x 51 > FH = 72 FH = 85 FH = 60 FH = 75 FH = 75 FH = 60 FH = 80

[0103] Substitute the data in Table 3 into the average score LP calculation formula to obtain:

[0104]

[0105] Then the average score LP of operator 1 within the learning period T is calculated 1 to be 72;

[0106] Suppose there are 6 operators, and the average scores LP of these 6 operators are as shown in Table 4 below:

[0107] Table 4: Average score LP table

[0108] Operator 1 Operator 2 Operator 3 Operator 4 Operator 5 Operator 6 <![CDATA[LP 1 > <![CDATA[LP 2 > <![CDATA[Lp 3 > <![CDATA[Lp 4 > <![CDATA[Lp 5 > <![CDATA[Lp 6 > 72 55 65 78 80 85

[0109] Arrange the operators in descending order according to all the average score LP values in Table 4 to obtain Table 5:

[0110] Table 5: Average score LP descending order table

[0111] Operator 6 Operator 5 Operator 4 Operator 1 Operator 3 Operator 2

[0112] Set Y = 2, U = 2. Then from Table 5, it can be determined that operator 6 and operator 5 are excellent operators, while operator 3 and operator 2 are operators to be trained. Subsequently, the construction and sending of training package PX can be carried out accordingly.

[0113] In summary, it is possible to periodically summarize the service levels of operators, and use database X to formulate training package PX, which promotes the learning of operators with lower service levels, provides convenience for the learning and communication among operators, not only helps to improve the service levels of operators to be trained, but also promotes the knowledge sharing and common progress of the entire call center team.

[0114] Embodiment 3. This embodiment provides a performance evaluation method, which uses the big data-based call center operation analysis system in Embodiment 1 to complete the performance evaluation of operators, and specifically includes the following steps:

[0115] D1. Set the evaluation period J. The processor module retrieves the database X to obtain the number SL of data packets x belonging to the operator within the evaluation period J in the database X, and calculates the performance evaluation score JX through the following formula:

[0116]

[0117] Among them, ZT represents the total number of calls of the operator recorded by the agent terminal within the evaluation period J, and Z and V are weights.

[0118] Example illustration:

[0119] Suppose the total number of calls ZT of operator 1 recorded by the agent terminal within the evaluation period J is 200 times. The processor retrieves the database X to determine that the number SL of data packets x belonging to operator 1 in the database X within the evaluation period J is 36. Set Z = 40% and V = 6%. Substituting the data into the calculation formula, we can get:

[0120] JX 1 =(200 - 36)*40% + 36*60% = 81 = 87.2;

[0121] The calculated performance evaluation score JX of operator 1 within the evaluation period J 1 is 87.2. The accounting personnel can then calculate the salary payment according to the enterprise's internal performance and salary comparison regulations.

[0122] In summary, the performance evaluation method formulated by score assignment makes the performance evaluation more fair and just, links the performance evaluation with the business level of the operator, and further realizes the effect of promoting the business level of the operator.

[0123] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirits of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. Call center operation analysis system based on big data, characterized by: It includes seat terminals, technician terminals, storage terminals and processor modules. The operator communicates with the customer through the seat terminals. 目标 The seat terminal records and calculates the score through the processor module. The score calculation result and the call voice are stored in the storage terminal. The operator and the technician communicate with each other through the seat terminal and the technician terminal respectively. The specific steps include the following: A1. The customer dials the phone to access the agent terminal, and the operator and the customer have a conversation. 目标 The agent terminal collects data in real time, and the operator solves the problems reported by the customer through the agent terminal; A2. When the operator fails to solve the customer's problem, the processor module will 目标 and the work orders recorded by the operator 目标 The call is sent to the technician terminal, and the technician continues to solve the problem. The technician checks the call audio through the technician terminal. 目标 Understand the problem and communicate with the customer through the technician terminal; A3. Processor module analyzes call voice 目标 , for voice calls 目标 Assign points and calculate call voice 目标 When the utilization rate is less than or equal to the threshold Q, the processor module analyzes the work order 目标 Confirm call voice 目标 The type of customer problem is selected, and the relevant data packets x are filtered in the storage terminal and sent to the technician terminal for the technician to use for problem solving; A4. After the problem is solved, the processor module will call the voice 目标 , Work Order 目标 , the scoring data and utilization rate are packaged into a data packet x and stored in the storage terminal to construct a database X = (x1, x2...xh), x h represents the hth data packet in database X.

2. The call center operation analysis system based on big data according to claim 1 is characterized in that: The content of the work order includes at least the operator's identity information, the time of answering and dialing, the type of customer problem, and the customer's product information.

3. The call center operation analysis system based on big data according to claim 2 is characterized in that: The call voice 目标 The utilization rate is calculated using the following formula: LYL=(LS*a+WL*b+FH*c)*m+FK*n+FL*i; Among them, LYL represents utilization rate, FK represents call voice 目标 The clarity score of the call is FL, which indicates the call voice. 目标 The professional level is scored, m, n and i are weights, and m+n+i=100%. LS represents the total amount of data packets x belonging to the operator in database X, and WL represents the total amount of work orders belonging to the operator in database X. 存储 Customer question types and call voices 目标 Work Order 目标 The number of packets x with the same customer problem type, FH represents the call voice 目标 The business level is scored, a, b and c are weights, and a+b+c=100%.

4. The call center operation analysis system based on big data according to claim 3 is characterized in that: The data packet x is expressed as follows: x=(TH, GD, LYL, FH, FK, FL); Among them, TH represents call voice, GD represents work order, LYL represents utilization rate, FH represents the business level score of call voice, FK represents the clarity score of call voice, and FL represents the professional level score of call voice.

5. The call center operation analysis system based on big data according to claim 4 is characterized in that: The step A3 of screening the relevant data packet x in the storage terminal specifically includes the following steps: C1. Processor module determines the work order 目标 In the customer problem type, retrieve the relevant work order from database X 目标 Tickets with the same customer issue type 存储 ; C2. All the retrieved work orders 存储 Based on work order 存储 Sort the utilization LYL values ​​in the data packet x in descending order, and select the first G work orders in the order 存储 Marking; C3. Mark the work order 存储 The data packet x belonging to it is extracted and sent to the technician terminal.

6. The call center operation analysis system based on big data according to claim 5 is characterized in that: The following steps are also included: A5. Set a learning cycle T. When the learning cycle T ends, the processor module calculates the average LP of the operators, sorts all operators in descending order according to the average LP values, marks the first Y operators as excellent operators, and marks the last U operators as operators to be learned; A6. Mark all packets x of the top operators in the learning period T as packets x. 优等 , all packets x 优等 Sort the data in descending order according to the service level score FH value, and select the top E data packets x 优等 The training package PX is packaged and sent to the seat terminals of all the operators to be trained for them to review and learn.

7. The call center operation analysis system based on big data according to claim 6 is characterized in that: The steps of calculating the average score LP are as follows: B1. The processor module retrieves the data packet x belonging to the operator within the learning cycle T in the storage terminal, and extracts the service level score FH in the data packet x; B2. Calculate the average score LP according to the following formula: Among them, FH k represents the kth extracted service level score FH, and r represents the number of extracted service level scores FH.

8. A performance evaluation method, characterized in that: The call center operation analysis system based on big data described in any one of claims 1 to 7 is used to complete the performance evaluation of the operators, specifically comprising the following steps: D1. Set the evaluation period J. The processor module retrieves the database X to obtain the number SL of data packets x belonging to the operator in the evaluation period J in the database X, and calculates the performance evaluation score JX by the following formula: Among them, ZT represents the total number of operator calls recorded by the agent terminal within the evaluation period J, and Z and V are weights.

Images