CN113628627B - Electric power industry customer service quality inspection system based on structured voice analysis - Google Patents
Electric power industry customer service quality inspection system based on structured voice analysis Download PDFInfo
- Publication number
- CN113628627B CN113628627B CN202110918237.8A CN202110918237A CN113628627B CN 113628627 B CN113628627 B CN 113628627B CN 202110918237 A CN202110918237 A CN 202110918237A CN 113628627 B CN113628627 B CN 113628627B
- Authority
- CN
- China
- Prior art keywords
- quality inspection
- analysis unit
- imf
- analysis
- service
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 125
- 238000007689 inspection Methods 0.000 title claims abstract description 75
- 238000005516 engineering process Methods 0.000 claims abstract description 5
- 238000012545 processing Methods 0.000 claims description 35
- 230000004044 response Effects 0.000 claims description 20
- 238000000354 decomposition reaction Methods 0.000 claims description 15
- 238000007781 pre-processing Methods 0.000 claims description 13
- 230000008030 elimination Effects 0.000 claims description 9
- 238000003379 elimination reaction Methods 0.000 claims description 9
- 230000008451 emotion Effects 0.000 claims description 9
- 238000001514 detection method Methods 0.000 claims description 8
- 230000002996 emotional effect Effects 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000000926 separation method Methods 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 claims description 2
- 230000006837 decompression Effects 0.000 claims description 2
- 238000012512 characterization method Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 5
- 230000009286 beneficial effect Effects 0.000 abstract description 4
- 238000000034 method Methods 0.000 description 10
- 238000007726 management method Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 4
- 238000013441 quality evaluation Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000008909 emotion recognition Effects 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
- G10L25/60—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination for measuring the quality of voice signals
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/22—Arrangements for supervision, monitoring or testing
- H04M3/2227—Quality of service monitoring
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/42—Systems providing special services or facilities to subscribers
- H04M3/50—Centralised arrangements for answering calls; Centralised arrangements for recording messages for absent or busy subscribers ; Centralised arrangements for recording messages
- H04M3/51—Centralised call answering arrangements requiring operator intervention, e.g. call or contact centers for telemarketing
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Business, Economics & Management (AREA)
- Marketing (AREA)
- Telephonic Communication Services (AREA)
Abstract
The invention provides a power industry customer service quality inspection system based on structured voice analysis, which comprises an acquisition module, a voice recognition module and a quality inspection analysis module, wherein the acquisition module is used for acquiring a voice signal of a customer; the system comprises an acquisition module, a quality inspection module and a quality inspection module, wherein the acquisition module is used for acquiring customer service data to be subjected to quality inspection, and the customer service data comprises service scene information and a call recording file between an agent and a customer; the voice recognition module is used for carrying out voice recognition on the call recording file through a voice recognition technology and acquiring a structured text file corresponding to the call recording file; the quality inspection analysis module is used for analyzing the call recording file and the corresponding structured text file according to preset quality inspection dimensions to obtain analysis results corresponding to different quality inspection dimensions; and acquiring a final service quality inspection result according to the analysis results corresponding to different quality inspection dimensions. The invention is beneficial to reducing the labor cost of the service quality management of the financial sharing service center in the power industry and improving the efficiency and the effect of the service quality management of the seat personnel.
Description
Technical Field
The invention relates to the technical field of intelligent quality inspection, in particular to a power industry customer service quality inspection system based on structured voice analysis.
Background
With the continuous development of the information era, the financial sharing service center in the power industry faces multiple tests at present, and particularly on the information construction of an auxiliary system, the requirement of the financial sharing service center cannot be met. A great number of calls are made to a financial sharing center every day to consult business problems, consult payment progress and the like, and a great number of consultations are provided with consultation services and make voice responses by seat service personnel. At present, for the service quality management of the seat service personnel, a special seat manager is usually used for monitoring the call of the seat service personnel or manually checking the call by recording, and the seat manager is used for evaluating and controlling the service quality by combining professional judgment of the seat manager, but with increasing traffic, the defects of the traditional service quality management method are obviously exposed, and because the proportion of manual check is very low, part of call contents with quality problems are not found, and the effect of checking the service quality of the seat service personnel is poor.
Disclosure of Invention
Aiming at the technical problem that the service quality inspection effect of the seat service personnel is poor due to the fact that the proportion of manual spot inspection in the traditional service quality management is low and part of call contents with quality problems are not found, the invention aims to provide a power industry customer service quality inspection system based on structured voice analysis.
The purpose of the invention is realized by adopting the following technical scheme:
the invention discloses a power industry customer service quality inspection system based on structured voice analysis, which comprises an acquisition module, a voice recognition module and a quality inspection analysis module, wherein the acquisition module is used for acquiring a voice signal; wherein,
the acquisition module is used for acquiring customer service data to be subjected to quality inspection, wherein the customer service data comprises service scene information and a call recording file between an agent and a customer;
the voice recognition module is used for carrying out voice recognition on the call recording file through a voice recognition technology and acquiring a structured text file corresponding to the call recording file;
the quality inspection analysis module is used for analyzing the call recording file and the corresponding structured text file according to preset quality inspection dimensions to obtain analysis results corresponding to different quality inspection dimensions; and acquiring a final service quality inspection result according to the analysis results corresponding to different quality inspection dimensions.
In one embodiment, the acquisition module further comprises a preprocessing unit; wherein,
the preprocessing unit is used for preprocessing the call recording file to be subjected to quality inspection, and the preprocessing includes decompression, voice segment marking, noise interference elimination and the like, so that the preprocessed call recording file is obtained.
In one embodiment, the speech recognition module includes a voiceprint recognition unit and a text recognition unit; wherein,
the voice print identification unit is used for carrying out speaker separation processing on the call recording file according to voice print characteristics and respectively dividing recording parts of seat personnel and clients in call recording;
the text recognition unit is used for performing voice recognition processing according to the obtained recording parts of the seat personnel and the clients to generate a corresponding text in a text format, and performing Chinese word segmentation, part-of-speech tagging and other processing on the text in the text format to obtain a structured text file; the acquired structured text file comprises text data respectively corresponding to the seat personnel and the clients.
In one embodiment, the predetermined quality inspection dimensions include: at least one of a business capability analysis, a violation analysis, and an attitude analysis.
In one embodiment, the quality inspection analysis module comprises a business capability analysis unit, a violation analysis unit, an attitude analysis unit and a scoring unit; wherein,
the service capability analysis unit further comprises a semantic analysis unit and a speech speed and silence analysis unit;
the semantic analysis unit is used for detecting the semantic definition of the response of the seat personnel to the questions raised by the user according to the structured text file, detecting whether the response of the seat personnel has keywords corresponding to the business scene or not, and analyzing the complete degree characteristic of the response of the seat personnel;
the speech speed and silence analysis unit is used for measuring and calculating the speech speed of the seat personnel for replying long sentences and counting the silence time of the seat personnel according to the call recording file, and analyzing the proficiency characteristics of the seat personnel for responding;
the service ability analysis unit is used for carrying out grading based on a preset service ability grading rule according to the acquired integrity degree characteristic and proficiency degree characteristic for analyzing the response of the seat personnel and acquiring the service ability grading of the seat personnel;
the violation analysis unit further comprises a tabu analysis unit;
the contraindication word analysis unit is used for matching the structured text file with a preset contraindication word library and detecting whether a contraindication word appears in the response of the seat personnel;
the violation analysis unit is used for outputting corresponding violation deductions of the seat personnel when detecting that the response of the seat personnel contains the contra-terms;
the attitude analysis unit further comprises a sensitive word analysis unit, a tone analysis unit and an emotion analysis unit;
the sensitive word analysis unit is used for matching the structured text file with a preset sensitive word library and detecting sensitive word information in the structured text file;
the tone analysis unit is used for analyzing tone and intonation of corresponding parts in the call recording file when sensitive words are detected in the structured text file, and acquiring a tone detection result;
the emotion analysis unit is used for judging the emotional state characteristics of the seat personnel and the clients according to the detected sensitive word information and the tone detection result;
the attitude analysis unit is used for scoring based on a preset service attitude scoring rule according to the emotional state characteristics of the seat personnel and the clients to obtain service attitude scores of the seat personnel;
and the evaluation unit is used for outputting a final service quality inspection result according to the acquired service capability evaluation of the seat personnel, violation deduction of the seat personnel and service attitude evaluation of the seat personnel.
The beneficial effects of the invention are as follows: calling from a storage library through an acquisition module or directly acquiring a call recording file of an agent in a service process from a financial sharing service center; converting the call recording file in the voice format into a structured text file in a text format based on a voice recognition module; and through the analysis module, service quality evaluation is carried out from different quality inspection dimensionalities to obtain a corresponding service quality inspection result. The invention can realize intelligent quality of service quality inspection based on the call recording file by means of an intelligent data processing mode, is beneficial to reducing the labor cost of service quality management of a financial sharing service center in the power industry, and improves the efficiency and effect of service quality management of seat personnel.
Drawings
The invention is further illustrated by means of the attached drawings, but the embodiments in the drawings do not constitute any limitation to the invention, and for a person skilled in the art, other drawings can be obtained on the basis of the following drawings without inventive effort.
FIG. 1 is a block diagram of an exemplary embodiment of a power industry customer service quality inspection system based on structured speech analysis according to the present invention;
FIG. 2 is a block diagram of an exemplary embodiment framework of the acquisition module of the embodiment of FIG. 1;
FIG. 3 is a block diagram of an exemplary embodiment framework of the speech recognition module of the embodiment of FIG. 1;
fig. 4 is a block diagram of an exemplary embodiment of a quality inspection module according to the embodiment of fig. 1.
Reference numerals:
the system comprises an acquisition module 1, a voice recognition module 2, a quality inspection analysis module 3, a preprocessing unit 11, a voiceprint recognition unit 21, a text recognition unit 22, a service capability analysis unit 31, a violation analysis unit 32, an attitude analysis unit 33, a scoring unit 34, a semantic analysis unit 311, a speech speed and silence analysis unit 312, a tabu word analysis unit 321, a sensitive word analysis unit 331 and a tone and intonation analysis unit 332
Detailed Description
The invention is further described in connection with the following application scenarios.
Referring to fig. 1, the embodiment of the invention provides a power industry customer service quality inspection system based on structured voice analysis, which includes an acquisition module 1, a voice recognition module 2 and a quality inspection analysis module 3; wherein,
the acquisition module 1 is used for acquiring customer service data to be subjected to quality inspection, wherein the customer service data comprises service scene information and a call recording file between an agent and a customer;
the voice recognition module 2 is used for performing voice recognition on the call recording file through a voice recognition technology to obtain a structured text file corresponding to the call recording file;
the quality inspection analysis module 3 is used for analyzing the call recording file and the corresponding structured text file according to preset quality inspection dimensions to obtain analysis results corresponding to different quality inspection dimensions; and acquiring a final service quality inspection result according to the analysis results corresponding to different quality inspection dimensions.
In the above embodiment of the present invention, the obtaining module 1 is used to call from the repository or directly obtain the call recording file of the seat staff in the service process from the financial sharing service center; converting the call recording file in the voice format into a structured text file in a text format based on the voice recognition module 2; and through the analysis module, service quality evaluation is carried out from different quality inspection dimensionalities to obtain a corresponding service quality inspection result. The invention can realize intelligent quality of service quality inspection based on the call recording file by means of an intelligent data processing mode, is beneficial to reducing the labor cost of service quality management of a financial sharing service center in the power industry, and improves the efficiency and effect of service quality management of seat personnel.
In one embodiment, the service scenario information includes service consultation, service introduction, payment progress inquiry and the like.
In one embodiment, referring to fig. 2, the acquisition module 1 further comprises a preprocessing unit 11; wherein,
the preprocessing unit 11 is configured to preprocess the call recording file to be subjected to quality inspection, including decompressing, marking a voice segment, eliminating noise interference, and the like, and obtain the preprocessed call recording file.
The method aims at the problem that the call recording file is easily affected by poor call quality or interference received in the recording process, the interference conditions that the call recording file is not clear, background noise is large and the like easily occur, and therefore the accuracy of service quality evaluation based on the call recording file is affected, therefore, the call recording file to be subjected to quality inspection, which is acquired by the acquisition module 1, is preprocessed through the preprocessing unit 11, the clarity and robustness of the call recording file to be subjected to quality inspection can be improved, and the reliability of service quality detection is indirectly improved.
In one embodiment, the preprocessing unit 11 performs noise interference elimination processing on a call recording file to be subjected to quality inspection, and specifically includes:
extracting a voice signal X in the call recording file;
to is directed atThe obtained speech signal is subjected to a Variable Mode Decomposition (VMD) process to obtain K IMF components { IMF } into which the speech signal is decomposed1,IMF2,…,IMFK}；
The first 3 IMF components to be obtained are IMF1,IMF2,IMF3Labeled low frequency IMF components; IMF the rest of IMF components4,…,IMFKLabeled high frequency IMF components;
for low frequency IMF components: reconstructing based on the obtained low-frequency IMF component to obtain a low-frequency signal Xd1；
Performing single-scale wavelet decomposition on the low-frequency component to obtain a low-frequency wavelet coefficient wd and a high-frequency wavelet coefficient wg of the low-frequency component; performing threshold processing on the obtained high-frequency wavelet coefficient wg to obtain a high-frequency wavelet coefficient wg' subjected to threshold processing;
reconstructing based on the low-frequency wavelet coefficient wd and the high-frequency wavelet coefficient wg' to obtain a preprocessed low-frequency signal X′ d1；
For the acquired high frequency IMF components: performing adaptive filtering processing based on the obtained high-frequency IMF components, including:
for the k-th1An IMF component of which k14,5, …, n, n represents a set characteristic scale characterizing variable, wherein n is greater than or equal to 5 and less than or equal to 8, for the IMF component
Performing I-scale wavelet decomposition, wherein I is n-4, and respectively acquiring low-frequency wavelet components corresponding to each scale
And high frequency wavelet components
Wherein
Representing the low frequency wavelet components acquired by the i-th scale wavelet decomposition,
representing the high-frequency wavelet component obtained by the ith scale wavelet decomposition;
based on low-frequency wavelet components
And high frequency wavelet components
Performing a reconstruction, wherein, if n-k1If < 1, high frequency wavelet component
Obtaining filtered IMF component for empty set
Based on IMF component { IMF'4,…,IMF′nReconstructing to obtain a preprocessed high-frequency signal X'g1；
According to the preprocessed low-frequency signal X'd1And a preprocessed high-frequency signal X'g1Performing superposition processing to obtain a voice signal X ' ═ X ' after noise interference elimination processing 'd1+X′g1And finishing the noise interference elimination processing of the call recording file.
In one embodiment, thresholding is performed on the acquired high-frequency wavelet coefficients wg, wherein the thresholding function is:
wherein wg' (k) represents the kth high-frequency wavelet coefficient after pre-thresholding, wg (k) represents the kth high-frequency wavelet coefficient, alpha represents a set linear adjustment factor, wherein 0 < alpha ≦ 1, and gamma represents a set exponential adjustment factor, wherein 1 < gamma ≦ 2; beta represents a set amplitude regulation factor, wherein beta is more than 0 and less than or equal to 0.3; δ represents a set threshold value.
Preferably, the linear adjustment factor α is 0.66, the exponential adjustment factor γ is 1.5, and the amplitude adjustment factor β is 0.2.
Preferably, the characteristic scale characterizing parameter n is 7.
In the above embodiment, a technical solution for performing noise interference elimination processing on a call record file is provided, where for background noise interference existing in the record file, each IMF component of a speech signal in the call record file is obtained based on a variational modal decomposition, and considering that the IMF component of the first 3 layers is usually a low-frequency component reflecting speech features, low-frequency overall filtering is performed on the low-frequency component based on a wavelet threshold processing manner to remove noise interference included in a speech part. Considering that in the remaining IMF components, some speech features are also present, but mainly the noise-dominant property, therefore, the filtering processing based on the multi-scale wavelet decomposition is provided, the voice characteristics can be attached (the reflection of the wavelet components obtained by decomposition on the voice characteristics is considered to be less and less obvious after the multi-scale decomposition, namely the leading of noise is stronger and stronger), the wavelet high-frequency components with high decomposition scale can be filtered directly in a self-adaptive manner, removing pure noise part (part not reflecting voice characteristics), and finally reconstructing based on the processed low-frequency signal and high-frequency signal to obtain voice signal processed by eliminating noise interference, the method and the device improve the definition of the voice signals in the call recording file and indirectly improve the reliability of subsequent voice recognition or other further processing according to the voice signals.
In one embodiment, referring to fig. 3, the speech recognition module 2 comprises a voiceprint recognition unit 21 and a text recognition unit 22; wherein,
the voiceprint recognition unit 21 is configured to perform speaker separation processing on the call recording file according to voiceprint characteristics, and divide recording portions of seat people and clients in the call recording respectively;
the text recognition unit 22 is configured to perform speech recognition processing according to the obtained recording portions of the seat staff and the clients, generate a corresponding text format text, perform processing such as Chinese word segmentation and part-of-speech tagging on the text format text, and obtain a structured text file; the acquired structured text file comprises text data respectively corresponding to the seat personnel and the clients.
The structured text is further classified into the text data corresponding to the seat personnel and the text data corresponding to the client for identification according to the voiceprint characteristics of the speaker in the call recording file, so that the service quality evaluation can be conveniently carried out subsequently and independently based on the recording or text data of the seat personnel or the client.
In one scenario, the text recognition unit 22 standardizes and structures text data based on a preset NLP (Natural Language Processing) technology.
In one embodiment, the predetermined quality inspection dimensions include: one or more of business capability analysis, violation analysis, attitude analysis, and the like.
The service capability analysis user checks the service capability level of the seat personnel in the service process; the attitude analysis is used for checking the service attitude of the seat personnel in the service process; violation analysis is used to check whether an agent has a violation during the service. Wherein, different quality inspection indexes and different quality inspection index grading rules are set corresponding to different quality inspection dimensions; the content of different quality inspection indexes and the grading rules of the quality inspection indexes can be correspondingly set according to different service scenes.
In one embodiment, referring to fig. 4, the quality inspection analysis module 3 includes a business capability analysis unit 31, a violation analysis unit 32, an attitude analysis unit 33, and a scoring unit 34; wherein,
the service capability analysis unit 31 further includes a semantic analysis unit 311 and a speech rate and silence analysis unit 312;
the semantic analysis unit 311 is configured to detect semantic clarity of a response of an agent to a question raised by a user according to the structured text file, detect whether a keyword corresponding to a service scene appears in the response of the agent, and analyze a completeness characteristic of the response of the agent;
the speech rate and silence analysis unit 312 is configured to measure and calculate a speech rate of the seat person for replying a long sentence and count a silence time of the seat person according to the call recording file, and analyze proficiency characteristics of the seat person for responding;
the service ability analysis unit 31 is used for carrying out scoring based on a preset service ability scoring rule according to the acquired integrity degree characteristic and proficiency degree characteristic for analyzing the response of the seat personnel and acquiring the service ability score of the seat personnel;
in one scenario, the semantic analysis unit 311 extracts a text vector based on a structured text file, obtains semantic information included in the text, and performs text similarity comparison between the obtained semantic information and a preset speech standard template to obtain semantic definition features of an agent; meanwhile, whether a specific keyword (such as an introduction keyword needed in a service introduction process) appears in the text is detected based on a specific service scene, and a score of the completeness of the response is obtained based on a preset scoring rule in combination with a semantic definition characteristic. The speech rate and silence analysis unit 312 performs speech rate identification on the call recording file by using a speech rate identification model to obtain the speech rate of the seat personnel; meanwhile, the mute duration in the call recording file is counted; judging whether the speech rate of the current seat personnel is in a threshold range or not according to a preset speech rate threshold range responded by the seat personnel; judging whether the counted continuous mute time exceeds the maximum mute time or not according to the preset maximum mute time; and combining the statistics to obtain corresponding proficiency grade according to a preset rule. The service ability analysis unit 31 combines the obtained integrity degree score and proficiency degree score and superposes the obtained integrity degree score and proficiency degree score according to a preset rule to obtain a service ability score of the seat staff.
The violation analysis unit 32 further includes a taboo word analysis unit 321;
the taboo word analysis unit 321 is configured to match the structured text file with a preset taboo word library, and detect whether a taboo word appears in a response of an attendant;
the violation analysis unit 32 is configured to output a corresponding violation deduction score for the seat staff when detecting that the response of the seat staff contains a contra word;
in one scenario, the taboo word analysis unit 321 detects whether the text data corresponding to the seat person contains preset taboos (e.g., abusive words, phrases that cannot appear according to the business scenario specification, words of negative emotion, etc.) based on the structured text file; the violation analysis unit 32 scores based on the contraindicated word detection result, and deducts the corresponding score according to a preset rule when the contraindicated word appears, so as to finally obtain the corresponding violation score of the seat personnel.
The attitude analysis unit 33 further includes a sensitive word analysis unit 331, a tone analysis unit 332, and an emotion analysis unit 333;
the sensitive word analysis unit 331 is configured to match the structured text file with a preset sensitive word library and detect sensitive word information in the structured text file;
the tone analysis unit 332 is configured to, when it is detected that a sensitive word occurs in the structured text file, perform tone and intonation analysis on a corresponding portion in the call recording file, and obtain a tone detection result;
the emotion analysis unit 333 is configured to determine emotional state characteristics of the agent and the client according to the detected sensitive word information and the detected tone detection result;
the attitude analysis unit 33 is used for scoring based on a preset service attitude scoring rule according to the emotional state characteristics of the seat personnel and the clients to obtain service attitude scores of the seat personnel;
in one scenario, the sensitive word analyzing unit 331 detects, based on the structured text file, whether the text data corresponding to the seat person includes a preset sensitive word (e.g., a sensitive tone word, a word that easily causes a complaint risk, etc.) and then, after the occurrence of the sensitive word is analyzed, further analyzes the tone and intonation of a speech segment corresponding to the occurrence of the sensitive word through the tone and intonation analyzing unit 332, where the tone and intonation of the speech segment are recognized through a tone recognition model, so as to obtain a tone feature (e.g., a statement tone, a question tone, etc.) of the seat person; tone recognition is carried out on the voice section through a tone recognition model, and tone features (such as anger, joy, sadness and the like) of the seat personnel are obtained; the emotion analysis unit 333 analyzes the emotion characteristics (e.g., positive, negative, etc.) of the agent person by the emotion recognition model based on the acquired tone characteristics and tone characteristics; and finally, the attitude analysis unit 33 performs conversion based on a preset rule according to the acquired emotional characteristics to obtain a corresponding service attitude score.
And the scoring unit 34 is used for outputting a final service quality inspection result according to the obtained service capability score of the seat personnel, violation deduction score of the seat personnel and the service attitude score of the seat personnel.
In one scenario, the scoring unit 34 generates a final service quality inspection result of the seat staff according to the obtained service capability score, the service attitude score and the violation deduction score.
In one embodiment, the system further comprises a storage module;
the storage module is also used for storing call recording files, data processing models, rule information set for different service scenes and the like. For each module to call when needed.
It should be noted that, functional units/modules in the embodiments of the present invention may be integrated into one processing unit/module, or each unit/module may exist alone physically, or two or more units/modules are integrated into one unit/module. The integrated units/modules may be implemented in the form of hardware, or may be implemented in the form of software functional units/modules.
From the above description of embodiments, it is clear for a person skilled in the art that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code or any appropriate combination thereof. For a hardware implementation, a processor may be implemented in one or more of the following units: an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a processor, a controller, a microcontroller, a microprocessor, other electronic units designed to perform the functions described herein, or a combination thereof. For a software implementation, some or all of the procedures of an embodiment may be performed by a computer program instructing associated hardware. In practice, the program may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a computer. Computer-readable media can include, but is not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.
Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the protection scope of the present invention, although the present invention is described in detail with reference to the preferred embodiments, it should be analyzed by those skilled in the art that modifications or equivalent substitutions can be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims (4)
1. A power industry customer service quality inspection system based on structured voice analysis is characterized by comprising an acquisition module, a voice recognition module and a quality inspection analysis module; wherein,
the acquisition module is used for acquiring customer service data to be subjected to quality inspection, wherein the customer service data comprises service scene information and a call recording file between an agent and a customer;
the voice recognition module is used for carrying out voice recognition on the call recording file through a voice recognition technology and acquiring a structured text file corresponding to the call recording file;
the quality inspection analysis module is used for analyzing the call recording file and the corresponding structured text file according to preset quality inspection dimensions to obtain analysis results corresponding to different quality inspection dimensions; obtaining a final service quality inspection result according to analysis results corresponding to different quality inspection dimensions;
the acquisition module further comprises a preprocessing unit; wherein,
the preprocessing unit is used for preprocessing the call recording file to be subjected to quality inspection, and comprises decompression, voice segment marking and noise interference elimination, so that the preprocessed call recording file is obtained;
the preprocessing unit carries out noise interference elimination processing on the call recording file to be subjected to quality inspection, and specifically comprises the following steps:
extracting a voice signal X in the call recording file;
performing variation modal decomposition processing on the acquired voice signal, and respectively acquiring K IMF components { IMF } decomposed from the voice signal1，IMF2，...，IMFK}；
IMF of the first 3 obtained IMF components1，IMF2，IMF3Labeled low frequency IMF components; IMF the remaining IMF components4，...，IMFKLabeled high frequency IMF components;
for low frequency IMF components: reconstructing based on the obtained low-frequency IMF component to obtain a low-frequency signal Xd1；
For low frequency signal Xd1Performing single-scale wavelet decomposition to obtain low-frequency signal Xd1Low-frequency wavelet coefficient wd and high-frequency wavelet coefficient wg; performing threshold processing on the obtained high-frequency wavelet coefficient wg to obtain a high-frequency wavelet coefficient wg' subjected to threshold processing;
reconstructing based on the low-frequency wavelet coefficient wd and the high-frequency wavelet coefficient wg 'to obtain a preprocessed low-frequency signal X'd1；
For the acquired high frequency IMF components: performing adaptive filtering processing based on the obtained high-frequency IMF components, including:
for the k-th1An IMF component of which k1K, for IMF components
Performing I-scale wavelet decomposition, wherein I is n-4, n represents a set characteristic scale characterization parameter, n is more than or equal to 5 and less than or equal to 8, and respectively acquiring low-frequency wavelet components corresponding to each scale
And high frequency wavelet components
Wherein
Representing the low frequency wavelet components acquired by the i-th scale wavelet decomposition,
representing the high-frequency wavelet component obtained by the ith scale wavelet decomposition;
based on low-frequency wavelet components
And high frequency wavelet components
Performing a reconstruction, wherein, if n-k1Less than 1, high frequency wavelet component
Obtaining filtered IMF component for empty set
Based on IMF component { IMF'4，...，IMF′KReconstructing to obtain a preprocessed high-frequency signal X'g1；
According to a preprocessed low-frequency signal X'd1And a preprocessed high-frequency signal X'g1Performing superposition processing to obtain a voice signal X ' ═ X ' after noise interference elimination processing 'd1+X′g1The noise interference elimination processing of the call recording file is completed;
wherein, the threshold processing is carried out on the obtained high-frequency wavelet coefficient wg, and the threshold processing function is as follows:
wherein wg' (k) represents the kth high-frequency wavelet coefficient after pre-thresholding, wg (k) represents the kth high-frequency wavelet coefficient, alpha represents a set linear adjustment factor, wherein 0 < alpha ≦ 1, and gamma represents a set exponential adjustment factor, wherein 1 < gamma ≦ 2; beta represents a set amplitude regulation factor, wherein beta is more than 0 and less than or equal to 0.3; δ represents a set threshold value.
2. The electric power industry customer service quality inspection system based on structured speech analysis of claim 1, wherein the speech recognition module comprises a voiceprint recognition unit and a text recognition unit; wherein,
the voice print identification unit is used for carrying out speaker separation processing on the call recording file according to voice print characteristics, and respectively dividing and acquiring recording parts of seat personnel and clients in call recording;
the text recognition unit is used for performing voice recognition processing according to the obtained recording parts of the seat personnel and the clients to generate a corresponding text in a text format, and performing Chinese word segmentation and/or part-of-speech tagging processing on the text in the text format to obtain a structured text file; the acquired structured text file comprises text data respectively corresponding to the seat personnel and the clients.
3. The electric power industry customer service quality inspection system based on structured speech analysis of claim 1, wherein the preset quality inspection dimensions comprise: at least one of service capability analysis, violation analysis and attitude analysis.
4. The electric power industry customer service quality inspection system based on structured voice analysis according to claim 3, wherein the quality inspection analysis module comprises a business capability analysis unit, a violation analysis unit, an attitude analysis unit and a scoring unit; wherein,
the service capability analysis unit further comprises a semantic analysis unit and a speech speed and silence analysis unit;
the semantic analysis unit is used for detecting the semantic definition of the response of the seat personnel to the questions raised by the user according to the structured text file, detecting whether the response of the seat personnel has keywords corresponding to the business scene or not, and analyzing the complete degree characteristic of the response of the seat personnel;
the speech speed and silence analysis unit is used for measuring and calculating the speech speed of the seat personnel for replying long sentences and counting the silence time of the seat personnel according to the call recording file, and analyzing the proficiency characteristics of the seat personnel for responding;
the service ability analysis unit is used for carrying out grading based on a preset service ability grading rule according to the acquired integrity degree characteristic and proficiency degree characteristic for analyzing the response of the seat personnel and acquiring the service ability grading of the seat personnel;
the violation analysis unit further comprises a tabu analysis unit;
the contraindication word analysis unit is used for matching the structured text file with a preset contraindication word library and detecting whether a contraindication word appears in the response of the seat personnel;
the violation analysis unit is used for outputting corresponding violation deductions of the seat personnel when detecting that the response of the seat personnel contains the contra-terms;
the attitude analysis unit further comprises a sensitive word analysis unit, a tone analysis unit and an emotion analysis unit;
the sensitive word analysis unit is used for matching the structured text file with a preset sensitive word library and detecting sensitive word information in the structured text file;
the tone analysis unit is used for analyzing tone and intonation of corresponding parts in the call recording file when sensitive words are detected in the structured text file, and acquiring a tone detection result;
the emotion analysis unit is used for judging the emotional state characteristics of the seat personnel and the clients according to the detected sensitive word information and the tone detection result;
the attitude analysis unit is used for scoring based on a preset service attitude scoring rule according to the emotional state characteristics of the seat personnel and the clients to obtain service attitude scores of the seat personnel;
and the evaluation unit is used for outputting a final service quality inspection result according to the acquired service capability evaluation of the seat personnel, violation deduction of the seat personnel and service attitude evaluation of the seat personnel.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110918237.8A CN113628627B (en) | 2021-08-11 | 2021-08-11 | Electric power industry customer service quality inspection system based on structured voice analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110918237.8A CN113628627B (en) | 2021-08-11 | 2021-08-11 | Electric power industry customer service quality inspection system based on structured voice analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113628627A CN113628627A (en) | 2021-11-09 |
| CN113628627B true CN113628627B (en) | 2022-06-14 |
Family
ID=78384378
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110918237.8A Active CN113628627B (en) | 2021-08-11 | 2021-08-11 | Electric power industry customer service quality inspection system based on structured voice analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113628627B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114418542A (en) * | 2022-01-20 | 2022-04-29 | 京东科技信息技术有限公司 | Method, device and equipment for checking business operation process and storage medium |
| CN114897499A (en) * | 2022-04-24 | 2022-08-12 | 重庆旅游云信息科技有限公司 | Method and device for judging working state of staff in scenic spot |
| CN115118818B (en) * | 2022-06-27 | 2024-06-21 | 平安银行股份有限公司 | Quality inspection method and device for call recording data, electronic equipment and storage medium |
| CN115277951A (en) * | 2022-07-26 | 2022-11-01 | 云南电网有限责任公司信息中心 | Intelligent voice outbound method, device, equipment and medium |
| CN116823069B (en) * | 2023-08-30 | 2023-12-05 | 北京中关村科金技术有限公司 | Intelligent customer service quality inspection method based on text analysis and related equipment |
| CN117640868B (en) * | 2024-01-23 | 2024-08-16 | 宁波菊风系统软件有限公司 | Intelligent double-recording system and method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102590716A (en) * | 2012-02-11 | 2012-07-18 | 广东电网公司电力科学研究院 | Power cable partial discharge signal denoising method based on wavelet analysis |
| CN102928517A (en) * | 2012-11-15 | 2013-02-13 | 河北省电力公司电力科学研究院 | Method for denoising acoustic testing data of porcelain insulator vibration based on wavelet decomposition threshold denoising |
| CN106872587A (en) * | 2017-02-10 | 2017-06-20 | 中南大学 | A kind of rail inspection instrument data de-noising method based on translation invariant GHM multi-wavelet transformations |
| CN109242799A (en) * | 2018-09-19 | 2019-01-18 | 安徽理工大学 | A kind of Wavelet noise-eliminating method of variable threshold value |
| CN112395992A (en) * | 2020-11-18 | 2021-02-23 | 云南电网有限责任公司电力科学研究院 | Electric power harmonic signal denoising method based on improved wavelet threshold |
| CN113238190A (en) * | 2021-04-12 | 2021-08-10 | 大连海事大学 | Ground penetrating radar echo signal denoising method based on EMD combined wavelet threshold |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109448730A (en) * | 2018-11-27 | 2019-03-08 | 广州广电运通金融电子股份有限公司 | A kind of automatic speech quality detecting method, system, device and storage medium |
| CN112562644A (en) * | 2020-12-03 | 2021-03-26 | 云知声智能科技股份有限公司 | Customer service quality inspection method, system, equipment and medium based on human voice separation |
| CN112885332A (en) * | 2021-01-08 | 2021-06-01 | 天讯瑞达通信技术有限公司 | Voice quality inspection method, system and storage medium |
| CN112966082B (en) * | 2021-03-05 | 2024-08-09 | 北京百度网讯科技有限公司 | Audio quality inspection method, device, equipment and storage medium |
-
2021
- 2021-08-11 CN CN202110918237.8A patent/CN113628627B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102590716A (en) * | 2012-02-11 | 2012-07-18 | 广东电网公司电力科学研究院 | Power cable partial discharge signal denoising method based on wavelet analysis |
| CN102928517A (en) * | 2012-11-15 | 2013-02-13 | 河北省电力公司电力科学研究院 | Method for denoising acoustic testing data of porcelain insulator vibration based on wavelet decomposition threshold denoising |
| CN106872587A (en) * | 2017-02-10 | 2017-06-20 | 中南大学 | A kind of rail inspection instrument data de-noising method based on translation invariant GHM multi-wavelet transformations |
| CN109242799A (en) * | 2018-09-19 | 2019-01-18 | 安徽理工大学 | A kind of Wavelet noise-eliminating method of variable threshold value |
| CN112395992A (en) * | 2020-11-18 | 2021-02-23 | 云南电网有限责任公司电力科学研究院 | Electric power harmonic signal denoising method based on improved wavelet threshold |
| CN113238190A (en) * | 2021-04-12 | 2021-08-10 | 大连海事大学 | Ground penetrating radar echo signal denoising method based on EMD combined wavelet threshold |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113628627A (en) | 2021-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113628627B (en) | Electric power industry customer service quality inspection system based on structured voice analysis | |
| CN110597964B (en) | Double-recording quality inspection semantic analysis method and device and double-recording quality inspection system | |
| CN109087670B (en) | Emotion analysis method, system, server and storage medium | |
| CN112804400A (en) | Customer service call voice quality inspection method and device, electronic equipment and storage medium | |
| CN109389971A (en) | Insurance recording quality detecting method, device, equipment and medium based on speech recognition | |
| CN109192194A (en) | Voice data mask method, device, computer equipment and storage medium | |
| CN113409824B (en) | Speech emotion recognition method | |
| CN112562736B (en) | Voice data set quality assessment method and device | |
| CN111145903A (en) | Method and device for acquiring vertigo inquiry text, electronic equipment and inquiry system | |
| CN110136726A (en) | A kind of estimation method, device, system and the storage medium of voice gender | |
| CN114612839A (en) | Short video analysis processing method, system and computer storage medium | |
| CN113450793B (en) | User emotion analysis method, device, computer readable storage medium and server | |
| CN112466284B (en) | Mask voice identification method | |
| CN117725926A (en) | Evaluation method and system based on natural language processing and violation inspection | |
| CN117116251A (en) | Repayment probability assessment method and device based on collection-accelerating record | |
| CN118035411A (en) | Customer service voice quality inspection method, customer service voice quality inspection device, customer service voice quality inspection equipment and storage medium | |
| Esmaili et al. | An automatic prolongation detection approach in continuous speech with robustness against speaking rate variations | |
| CN112885379A (en) | Customer service voice evaluation method, system, device and storage medium | |
| CN111489736A (en) | Automatic seat speech technology scoring device and method | |
| CN116150313A (en) | Data expansion processing method and device | |
| CN112052994A (en) | Customer complaint upgrade prediction method and device and electronic equipment | |
| CN112668617A (en) | Power grid employee work satisfaction evaluation method and device | |
| CN117935865B (en) | User emotion analysis method and system for personalized marketing | |
| CN118588112B (en) | Alternating current state analysis method, equipment and medium for nonverbal signals | |
| CN118333690B (en) | Intelligent advertisement auditing method, system and terminal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |