CN111010484A - Automatic quality inspection method for call recording - Google Patents
Automatic quality inspection method for call recording Download PDFInfo
- Publication number
- CN111010484A CN111010484A CN201911296512.6A CN201911296512A CN111010484A CN 111010484 A CN111010484 A CN 111010484A CN 201911296512 A CN201911296512 A CN 201911296512A CN 111010484 A CN111010484 A CN 111010484A
- Authority
- CN
- China
- Prior art keywords
- hit
- quality inspection
- operators
- recording
- operator
- 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.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/22—Arrangements for supervision, monitoring or testing
- H04M3/2218—Call detail recording
-
- 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
- H04M3/22—Arrangements for supervision, monitoring or testing
- H04M3/2281—Call monitoring, e.g. for law enforcement purposes; Call tracing; Detection or prevention of malicious calls
Abstract
The invention discloses an automatic quality inspection method for call recording, and relates to the technical field of recording quality inspection. The invention comprises the following steps: an external ASR access module converts the recording file into recording json with a specific format; processing the sound recording json to obtain a set object required by the system; acquiring a quality inspection rule configured by a system, taking out operators of all models in the rule, marking the operators, and classifying all the operators through the operator types; the record set object and various operators are subjected to operator hit processing to identify the hit condition of the operators; extracting hit operators for judgment; and extracting all hit operators in the model for scoring, and adding to obtain the final score of the sound recording and recording. The invention arranges and matches the quality inspection model by classifying and sorting out a plurality of basic sub-condition complementarity parameters, generates the quality inspection template for the combination of the models, automatically inspects the call record, generates the quality inspection report and quality inspection statistics, improves the quality inspection coverage rate and reduces the operation cost.
Description
Technical Field
The invention belongs to the technical field of quality inspection of recorded sound, and particularly relates to an automatic quality inspection method for call recording.
Background
With the improvement of the social civilization degree, the concept of people is changed, the citizen administrative consciousness and the legal right-maintaining consciousness are improved, and the requirement on service is higher and higher. The core of the call center service is the call between the agent and the customer, and in order to improve the service quality, the call problem can be found in time through quality inspection, which is the key to improve the service quality.
At present, the traditional call record quality inspection depends on a mode of carrying out sampling inspection by an administrator, and judges whether the service of a customer service specialist meets the specification or not by manually listening to the record. However, as the traffic volume gradually rises, the proportion of the numbers of the special customer service personnel and the quality inspection personnel is seriously unbalanced, and the defects of small coverage, strong subjectivity of analysis and judgment and fussy and repeated labor can occur in a manual quality inspection mode.
Therefore, a set of configured quality inspection rules replace artificial subjective judgment through machine intelligent matching calculation, the recording problem is accurately locked, the quality inspection coverage rate can be greatly improved, and the labor cost is reduced. The quality inspection result is more objective, and the customer service quality is improved.
In order to enable quality inspection rules to be configured more quickly and enable configured rules to be stronger in capability, the application document provides sub-conditions such as 'specific keywords', 'specific dialect', 'silent duration', 'excessive speech speed', 'verbose', 'speech robbing', 'specific semantics', 'emotional abnormity' and the like, and the rules meeting quality inspection requirements are configured through setting sub-condition parameters and mutual matching among the sub-conditions. After the rule is configured, the automatic quality inspection can be carried out on the full call records. And generating quality inspection reports and quality inspection statistics.
Disclosure of Invention
The invention aims to provide an automatic quality inspection method for call records, which arranges and matches a quality inspection model by classifying and sorting a plurality of basic sub-condition complementarity parameters, generates a quality inspection template for the combination of the models, performs automatic quality inspection on the call records, generates a quality inspection report and quality inspection statistics, and solves the problems of low coverage rate and high labor cost of the existing call record quality inspection.
In order to solve the technical problems, the invention is realized by the following technical scheme:
the invention relates to a method for automatically testing the quality of call records, which comprises the following steps:
step S1: the call center records the call records of the customer service and the customer;
step S2: an external ASR access module converts the recording file into recording json with a specific format;
step S3: processing the sound recording json to obtain a set object required by the system;
step S4: converting the role identification into a role through a role identification module;
step S5: acquiring a quality inspection rule configured by a system, taking out operators of all models in the rule, marking the operators, and classifying all the operators through the operator types;
step S6: identifying the hit condition of each operator by an operator hit processing module for the processed sound recording set object and various operators;
step S7: extracting hit operators and judging according to general conditions of all conditions;
step S8: and (4) extracting all hit operators in the model for scoring, summing up to obtain the final score of the sound recording, and recording the hit condition.
Preferably, in step S5, the quality inspection rule configured by the system classifies eight basic sub-conditions; the sub-conditions comprise specific keywords, specific dialogues, silence duration, excessive speech speed, verbose speaking, speech snatching, specific semantics and emotional abnormality.
Preferably, the model combined by the sub-condition logics is a score unit, the sum of the models is a quality inspection rule, and the final quality inspection score is obtained by calculating the hit score of the model in the rule.
Preferably, in step S6, the operator hit processing module performs overall processing on the normalized set object by using a new property Stream and a lambada expression of JAVA 8; wherein, the keyword matching operation adopts a generic hasMap; the special dialect adopts regular expression operation; judging by adopting data when the sound is muted and the speech speed is too high; the verdant matching is mainly judged by using character string matching.
Preferably, in step S7, the detection range judged by the general condition includes a full text, an absolute position, and a relative position, and if the hit operator is the full text, the full text is retained; if the hit operator is an absolute position, screening according to the parameters; and if the hit operator is in a relative position, whether the current operator is hit or not is calculated according to the hit operator.
Preferably, in step S8, the recorded hit condition further includes a dialog text and a dialog role, the hit model and the hit condition in the model can be seen in the restored dialog, the hit position and the hit reason can be accurately found by clicking the hit condition, and the currently hit recording clip can be played.
The invention has the following beneficial effects:
(1) according to the invention, full-scale automatic quality inspection can be carried out on call records through configured rules, all quality inspection scenes are realized through the collocation of operators by adopting an integrated and extracted configuration operator, the manual quality inspection is completely replaced by intelligent calculation, and the uniformity, the efficiency and the objectivity of the quality inspection are ensured;
(2) the invention accurately positions the hit position of the quality inspection by restoring the conversation scene, specifically the conversation text, the conversation time, the hit condition and the like in a way of clicking the hit condition, restores the conversation scene by playing the hit fragment record, and improves the customer service level.
(3) The invention arranges and matches the quality inspection model by classifying and sorting out a plurality of basic sub-condition complementarity parameters, generates the quality inspection template for the combination of the models, automatically inspects the call record, generates the quality inspection report and quality inspection statistics, improves the quality inspection coverage rate and reduces the operation cost.
Of course, it is not necessary for any product in which the invention is practiced to achieve all of the above-described advantages at the same time.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a step diagram of an automatic quality inspection method for call records according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, the present invention is a method for automatically inspecting quality of call records, comprising the following steps:
step S1: the call center records the call records of the customer service and the customer;
step S2: an external ASR access module converts the recording file into recording json with a specific format;
step S3: processing the sound recording json to obtain a set object required by the system;
step S4: converting the role identification into a role through a role identification module;
step S5: acquiring a quality inspection rule configured by a system, taking out operators of all models in the rule, marking the operators, and classifying all the operators through the operator types;
step S6: preliminarily identifying the hit condition of each operator by using the processed record set object and various operators through an operator hit processing module corresponding to keywords, speech speed, semantics and the like;
step S7: extracting hit operators and judging according to general conditions of all conditions;
step S8: and (4) extracting all hit operators in the model for scoring, summing up to obtain the final score of the sound recording, and recording the hit condition.
In step S2, the json of the recorded sound includes the dialog text, the character identifier, the word segmentation start time, the word segmentation end time, the speech speed, and so on.
In step S5, the quality inspection rule configured by the system classifies eight basic sub-conditions; the sub-conditions comprise specific keywords, specific dialogues, mute time, over-fast speech speed, verbose speaking, speech robbing, specific semantics and emotional abnormality; wherein, the specific keyword can be configured with the following parameters: a keyword list, a synonym list, a matching role, a matching mode and a detection range; the following parameters may be configured for a particular session: a word operation list, a matching role and a detection range; muting may be configured with the following parameters: mute duration, mute times, previous role, next role, detection range; the parameter can be configured when the speech speed is too fast: the number of words per minute, the role matching and the detection range are set; speaking treble can configure parameters: verruca littoralis list, verruca littoralis conditions, role matching and detection range; the call can be configured with the following parameters: cross duration, number of robbed words, matching roles, detection range; the specific semantics may be configured with the following parameters: reference statement list, similarity, matching role and detection range; the emotional abnormality does not need to configure parameters.
The model combined by the sub-condition logics is a score unit, the sum of the models is a quality inspection rule, and the final quality inspection score is obtained by calculating the hit score of the model in the rule.
In step S6, the operator hit processing module performs overall processing on the normalized set object mainly by using a new feature Stream and a lambada expression of JAVA8 based on the converted normalized set object (i.e., the data format required for operation) and the data rule that is lost into the cache; wherein, the keyword matching operation adopts a generic hasMap; the special dialect adopts regular expression operation; judging by adopting data when the sound is muted and the speech speed is too high; the verrucous matching is mainly judged by using character string matching; the whole operation does not deviate from the object set (list < class >), and does not deviate from the data format of the set object after each operation, and the whole operation is carried out on the object set through the stream.
Each layer hasMap may contain three attributes text, isEnd (end identifier), info (detailed information set, where keyworkruledto is a detailed information object);
if "customer your good" is matched, the specific matching process is as follows:
the first character 'customer' is obtained, get 'customer' from the first layer map, the fact that the character enters the next layer and does not directly exit matching is included, the character exits directly here, the next character 'customer' is read, the same judgment is carried out until the character 'you', the keyword includes the keyword at the beginning of the character 'you', the next character is judged, the character 'you' is judged to be matched at the moment, but the end identification isEnd is 0, and the character 'you' in matching is abandoned. Until the end is read to be 1, the matched keywords are hit keywords; whether the hit attribute is marked as hit in the keyworkruledto object and the location of the hit is recorded. The simplest matching process is described here, where info can be expanded into multiple conditions, and keywords are also divided into different attributes. Such as: and multiple attribute expansion such as keywords, synonyms, role segmentation words and the like.
In step S7, the detection range judged by the general condition includes full text, absolute position, and relative position, and if the hit operator is full text, the operator is retained; if the hit operator is an absolute position, screening according to the parameters; and if the hit operator is in a relative position, whether the current operator is hit or not is calculated according to the hit operator.
In step S8, the recorded hit condition further includes dialogue characters and dialogue roles, the hit model and the hit condition in the model can be seen in the restored dialogue, the hit position and the hit reason can be accurately found by clicking the hit condition, and the currently hit recording clip can be played.
It should be noted that, in the above system embodiment, each included unit is only divided according to functional logic, but is not limited to the above division as long as the corresponding function can be implemented; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
In addition, it is understood by those skilled in the art that all or part of the steps in the method for implementing the embodiments described above may be implemented by a program instructing associated hardware, and the corresponding program may be stored in a computer-readable storage medium.
The preferred embodiments of the invention disclosed above are intended to be illustrative only. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
Claims (6)
1. A method for automatically testing the quality of call records is characterized by comprising the following steps:
step S1: the call center records the call records of the customer service and the customer;
step S2: an external ASR access module converts the recording file into recording json with a specific format;
step S3: processing the sound recording json to obtain a set object required by the system;
step S4: converting the role identification into a role through a role identification module;
step S5: acquiring a quality inspection rule configured by a system, taking out operators of all models in the rule, marking the operators, and classifying all the operators through the operator types;
step S6: identifying the hit condition of each operator by an operator hit processing module for the processed sound recording set object and various operators;
step S7: extracting hit operators and judging according to general conditions of all conditions;
step S8: and (4) extracting all hit operators in the model for scoring, summing up to obtain the final score of the sound recording, and recording the hit condition.
2. The method as claimed in claim 1, wherein in step S5, the quality control rule configured by the system classifies eight basic sub-conditions; the sub-conditions comprise specific keywords, specific dialogues, silence duration, excessive speech speed, verbose speaking, speech snatching, specific semantics and emotional abnormality.
3. The method as claimed in claim 2, wherein the model logically combined by the sub-conditions is a score unit, the sum of the models is a quality inspection rule, and the final quality inspection score is obtained by calculating the hit score of the model in the rule.
4. The method according to claim 1, wherein in step S6, the sub hit processing module performs an overall process on the normalized set object by using a new feature Stream and lambada expression of JAVA 8; wherein, the keyword matching operation adopts a generic hasMap; the special dialect adopts regular expression operation; judging by adopting data when the sound is muted and the speech speed is too high; the verdant matching is mainly judged by using character string matching.
5. The method according to claim 1, wherein in step S7, the detection range for determining the general condition includes full text, absolute position and relative position, and if the hit operator is full text, the detection range is retained; if the hit operator is an absolute position, screening according to the parameters; and if the hit operator is in a relative position, whether the current operator is hit or not is calculated according to the hit operator.
6. The method of claim 1, wherein in step S8, the recorded hit condition further includes dialogue text and dialogue roles, the hit model and the hit condition in the model can be seen in the restored dialogue, the hit position and the hit reason can be accurately found by clicking the hit condition, and the currently hit recording clip can be played.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911296512.6A CN111010484A (en) | 2019-12-16 | 2019-12-16 | Automatic quality inspection method for call recording |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911296512.6A CN111010484A (en) | 2019-12-16 | 2019-12-16 | Automatic quality inspection method for call recording |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111010484A true CN111010484A (en) | 2020-04-14 |
Family
ID=70115344
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911296512.6A Pending CN111010484A (en) | 2019-12-16 | 2019-12-16 | Automatic quality inspection method for call recording |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111010484A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111565254A (en) * | 2020-07-14 | 2020-08-21 | 深圳追一科技有限公司 | Call data quality inspection method and device, computer equipment and storage medium |
| CN113553861A (en) * | 2021-07-30 | 2021-10-26 | 出门问问信息科技有限公司 | Information processing method and device based on dialog system and storage medium |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07107541A (en) * | 1993-10-04 | 1995-04-21 | Fujitsu Ltd | Call process debugging method for digital mobile object exchange |
| JP2007074175A (en) * | 2005-09-06 | 2007-03-22 | Fujitsu Ltd | Telephone service inspection system and program thereof |
| CN101150622A (en) * | 2007-01-30 | 2008-03-26 | 华为技术有限公司 | Quality check method, quality check device and quality check system |
| CN105975514A (en) * | 2016-04-28 | 2016-09-28 | 朱宇光 | Automatic quality testing method and system |
| CN106934000A (en) * | 2017-03-03 | 2017-07-07 | 深圳市彬讯科技有限公司 | A kind of automatic quality detecting method of the voice of calling system and system |
| CN108965620A (en) * | 2018-08-24 | 2018-12-07 | 杭州数心网络科技有限公司 | A kind of artificial intelligence call center system |
| CN109151218A (en) * | 2018-08-21 | 2019-01-04 | 平安科技（深圳）有限公司 | Call voice quality detecting method, device, computer equipment and storage medium |
| CN109327632A (en) * | 2018-11-23 | 2019-02-12 | 深圳前海微众银行股份有限公司 | Intelligent quality inspection system, method and the computer readable storage medium of customer service recording |
| 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 |
| CN109446524A (en) * | 2018-10-25 | 2019-03-08 | 第四范式（北京）技术有限公司 | A kind of voice quality detecting method and device |
| CN110197672A (en) * | 2018-02-27 | 2019-09-03 | 招商信诺人寿保险有限公司 | A kind of voice call quality detection method, server, storage medium |
| CN110334241A (en) * | 2019-07-10 | 2019-10-15 | 深圳前海微众银行股份有限公司 | Quality detecting method, device, equipment and the computer readable storage medium of customer service recording |
-
2019
- 2019-12-16 CN CN201911296512.6A patent/CN111010484A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07107541A (en) * | 1993-10-04 | 1995-04-21 | Fujitsu Ltd | Call process debugging method for digital mobile object exchange |
| JP2007074175A (en) * | 2005-09-06 | 2007-03-22 | Fujitsu Ltd | Telephone service inspection system and program thereof |
| CN101150622A (en) * | 2007-01-30 | 2008-03-26 | 华为技术有限公司 | Quality check method, quality check device and quality check system |
| CN105975514A (en) * | 2016-04-28 | 2016-09-28 | 朱宇光 | Automatic quality testing method and system |
| CN106934000A (en) * | 2017-03-03 | 2017-07-07 | 深圳市彬讯科技有限公司 | A kind of automatic quality detecting method of the voice of calling system and system |
| CN110197672A (en) * | 2018-02-27 | 2019-09-03 | 招商信诺人寿保险有限公司 | A kind of voice call quality detection method, server, storage medium |
| CN109151218A (en) * | 2018-08-21 | 2019-01-04 | 平安科技（深圳）有限公司 | Call voice quality detecting method, device, computer equipment and storage medium |
| CN108965620A (en) * | 2018-08-24 | 2018-12-07 | 杭州数心网络科技有限公司 | A kind of artificial intelligence call center system |
| CN109446524A (en) * | 2018-10-25 | 2019-03-08 | 第四范式（北京）技术有限公司 | A kind of voice quality detecting method and device |
| CN109327632A (en) * | 2018-11-23 | 2019-02-12 | 深圳前海微众银行股份有限公司 | Intelligent quality inspection system, method and the computer readable storage medium of customer service recording |
| 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 |
| CN110334241A (en) * | 2019-07-10 | 2019-10-15 | 深圳前海微众银行股份有限公司 | Quality detecting method, device, equipment and the computer readable storage medium of customer service recording |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111565254A (en) * | 2020-07-14 | 2020-08-21 | 深圳追一科技有限公司 | Call data quality inspection method and device, computer equipment and storage medium |
| CN113553861A (en) * | 2021-07-30 | 2021-10-26 | 出门问问信息科技有限公司 | Information processing method and device based on dialog system and storage medium |
| CN113553861B (en) * | 2021-07-30 | 2023-11-14 | 出门问问信息科技有限公司 | Information processing method, device and storage medium based on dialogue system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108962282B (en) | Voice detection analysis method and device, computer equipment and storage medium | |
| US11227603B2 (en) | System and method of video capture and search optimization for creating an acoustic voiceprint | |
| Schuller et al. | The INTERSPEECH 2021 computational paralinguistics challenge: COVID-19 cough, COVID-19 speech, escalation & primates | |
| US8219404B2 (en) | Method and apparatus for recognizing a speaker in lawful interception systems | |
| CN111128223B (en) | Text information-based auxiliary speaker separation method and related device | |
| WO2021068843A1 (en) | Emotion recognition method and apparatus, electronic device, and readable storage medium | |
| CN107274916B (en) | Method and device for operating audio/video file based on voiceprint information | |
| CN107093431A (en) | A kind of method and device that quality inspection is carried out to service quality | |
| CN110135879A (en) | Customer service quality automatic scoring method based on natural language processing | |
| CN111639484A (en) | Method for analyzing seat call content | |
| CN110136696B (en) | Audio data monitoring processing method and system | |
| CN111010484A (en) | Automatic quality inspection method for call recording | |
| JP6208794B2 (en) | Conversation analyzer, method and computer program | |
| CN113129866A (en) | Voice processing method, device, storage medium and computer equipment | |
| CN115063155A (en) | Data labeling method and device, computer equipment and storage medium | |
| CN116150313A (en) | Data expansion processing method and device | |
| CN109635151A (en) | Establish the method, apparatus and computer equipment of audio retrieval index | |
| CN110797032B (en) | Voiceprint database establishing method and voiceprint identification method | |
| CN114449105A (en) | Voice-based electric power customer service telephone traffic quality inspection system | |
| CN114356982A (en) | Marketing compliance checking method and device, computer equipment and storage medium | |
| CN113744742A (en) | Role identification method, device and system in conversation scene | |
| CN111666469B (en) | Statement library construction method, device, equipment and storage medium | |
| US20230238002A1 (en) | Signal processing device, signal processing method and program | |
| CN113810548A (en) | Intelligent call quality inspection method and system based on IOT | |
| CN114333784A (en) | Information processing method, information processing device, computer equipment and storage medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200414 |
|
| RJ01 | Rejection of invention patent application after publication |