WO2011029306A1 - Voice-based customer evaluation system and method thereof - Google Patents

Voice-based customer evaluation system and method thereof Download PDF

Info

Publication number
WO2011029306A1
WO2011029306A1 PCT/CN2010/072214 CN2010072214W WO2011029306A1 WO 2011029306 A1 WO2011029306 A1 WO 2011029306A1 CN 2010072214 W CN2010072214 W CN 2010072214W WO 2011029306 A1 WO2011029306 A1 WO 2011029306A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
voice
module
evaluation
customer
Prior art date
Application number
PCT/CN2010/072214
Other languages
French (fr)
Chinese (zh)
Inventor
赖永森
郭潺
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2011029306A1 publication Critical patent/WO2011029306A1/en

Links

Classifications

    • 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
    • H04M3/5175—Call or contact centers supervision arrangements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M2201/00—Electronic components, circuits, software, systems or apparatus used in telephone systems
    • H04M2201/40—Electronic components, circuits, software, systems or apparatus used in telephone systems using speech recognition

Definitions

  • a voice-based customer evaluation system A voice-based customer evaluation system
  • the present invention relates to a customer service system, and more particularly to a voice-based customer evaluation system and a customer evaluation method. Background technique
  • Call center also known as customer service system, is an information system used to provide users with various access methods such as telephone, fax, and e-mail. It is mainly used to handle user requests, questions, complaints, suggestions, and consultations. For example, 1860 in the telecommunications industry, 95555 in the financial industry, etc.
  • FIG. 1 is a schematic structural diagram of a customer service system.
  • an existing customer service system includes a call access module, a Computer Telephony Integration (CTI) module, at least one agent module, and a media server.
  • the media server can provide automatic customer service for the user.
  • CTI Computer Telephony Integration
  • the CTI module can route the user call to an agent module according to a certain policy, and the agent module provides the operator with a platform for interacting with the user.
  • the object of the present invention is to provide a voice-based customer evaluation system and a customer evaluation method, to achieve objective customer evaluation, and to provide an objective basis for service and problem tracking.
  • a voice-based client evaluation system including:
  • a voice-based customer evaluation system includes a data intercepting module, a first converting module, and a voice analyzing module, wherein
  • a data intercepting module configured to intercept media stream data sent by the user end to the customer service center; a first converting module, configured to perform a converting operation on the intercepted media stream to obtain voice stream data; and a voice analyzing module, configured to analyze the obtained voice
  • the stream data is obtained by evaluating a value of the first evaluation indicator of the user.
  • the customer evaluation system further includes a second conversion module and a text analysis module, wherein the second conversion module is configured to perform a conversion operation on the voice stream data to obtain a text; a text analysis module, configured to analyze the obtained text, and obtain The value of the second evaluation indicator of the user is evaluated.
  • the customer evaluation system also includes:
  • a database for storing user information
  • An obtaining module configured to obtain caller information of the current call
  • a user querying module configured to query the database according to the calling user information of the current call, and determine a user identity
  • an update module configured to perform an update operation on the database according to the calling user information of the current call when the user cannot be queried according to the calling user information of the current call.
  • the user information is a subscriber number and/or voice information.
  • the text analysis module specifically includes a vocabulary saving unit, a statistical unit, and a calculating unit, wherein
  • a vocabulary saving unit for saving a preset vocabulary
  • a statistical unit configured to count the number of occurrences of the preset vocabulary in the text
  • a calculating unit configured to calculate a value of the second evaluation index, where the second evaluation index value The value is equal to the number of occurrences divided by the total number of words of the text.
  • the voice analyzing module specifically includes a parameter extracting unit and a sentiment judging unit, wherein the parameter extracting unit is configured to extract a value of the preset parameter according to the voice stream data; and the emotion judging unit is configured to use the value of the extracted preset parameter Determining a value of the first evaluation indicator, where the first evaluation indicator is a user emotion indicator.
  • the parameter extraction unit is specifically configured to extract a value of the preset parameter by using a principal component analysis method, a neural network method, or a Gaussian mixture model method.
  • the customer evaluation system also includes:
  • the comprehensive analysis module is configured to calculate a value for evaluating the third evaluation indicator of the user according to the historical data and the current data of the first evaluation index and the second evaluation index.
  • a voice-based customer evaluation method including:
  • the customer evaluation system intercepts media stream data sent by the client to the customer service center; the customer evaluation system performs a conversion operation on the media stream to obtain voice stream data; and the customer evaluation system analyzes the voice stream data to obtain a number for evaluating the user.
  • the value of an evaluation indicator is a value of an evaluation indicator.
  • the method also includes:
  • the customer evaluation system performs a conversion operation on the voice stream data to obtain a text; the client evaluation system analyzes the text to obtain a value for evaluating a second evaluation indicator of the user.
  • the method also includes:
  • the customer evaluation system calculates a value for evaluating a third evaluation index of the user based on historical data and current data of the first evaluation index and the second evaluation index.
  • the customer evaluation system of the present invention intercepts the media stream data sent by the client to the customer service center, converts it into voice stream data and text, and analyzes the voice stream and the text based on the voice stream and the text. Obtaining the value of the evaluation index used to evaluate the user, so the evaluation is objective and provides an objective basis for service and problem tracking.
  • Figure 1 is a schematic structural diagram of a customer service system
  • FIG. 2 is a schematic structural diagram of a first embodiment of a voice-based customer evaluation system according to the present invention
  • FIG. 3 is a schematic structural diagram of a second embodiment of a voice-based customer evaluation system according to the present invention
  • FIG. 4 is a voice-based customer evaluation of the present invention.
  • FIG. 5 is a schematic flowchart of a second embodiment of a voice-based client evaluation method according to the present invention.
  • the user's voice data is intercepted, and the voice data is analyzed, and the feature value for performing the user evaluation is obtained, and the user evaluation is performed based on the feature value.
  • FIG. 2 is a schematic structural diagram of a first embodiment of a voice-based client evaluation system according to the present invention.
  • the system includes a data intercepting module, a first converting module, and a voice analyzing module, where the data intercepting module is configured to intercept the user.
  • the media stream data sent to the customer service center.
  • the first conversion module is configured to perform a conversion operation on the intercepted media stream data to obtain voice stream data.
  • the voice analysis module is configured to analyze the obtained voice stream data, and obtain a value for evaluating the first evaluation indicator of the user.
  • FIG. 3 is a schematic structural diagram of a second embodiment of the voice-based customer evaluation system of the present invention, as shown in FIG.
  • the data intercepting module, the first converting module, the second converting module, the speech analyzing module, and the text analyzing module wherein
  • the data intercepting module is configured to intercept media stream data sent by the client to the customer service center.
  • the first conversion module is configured to perform a conversion operation on the intercepted media stream data to obtain voice stream data.
  • a second conversion module configured to perform a conversion operation on the obtained voice stream data to obtain a text
  • a voice analysis module configured to analyze the obtained voice stream data, and obtain a value of the first evaluation index
  • a text analysis module configured to analyze Text, obtaining a value for evaluating the second evaluation indicator of the user.
  • the foregoing media stream data may be media stream data in a protocol for supporting media stream transmission in any of the prior art.
  • a real-time transport protocol RTP
  • RTP real-time transport protocol
  • the customer evaluation system of the embodiment of the present invention further includes (not shown in FIG. 2 and FIG. 3 for simplification of the drawing. ):
  • the user information is a user number and/or voice information.
  • the obtaining module is configured to obtain caller information of the current call.
  • the user query module is configured to query the database according to the calling user information of the current call to determine the identity of the user.
  • the update module is configured to perform an update operation on the database according to the calling user information of the current call when the user cannot be queried according to the calling user information of the current call.
  • the user's voice information is saved in the database.
  • the obtaining module is configured to obtain a calling party number, and extract the calling user voice information from the voice stream data.
  • the user query module queries the database according to the calling user number and the calling user voice information. Determine the identity of the user.
  • the update module is configured to perform an update operation on the database according to the calling party number or the calling user voice information when the user cannot be queried according to the calling party number or the calling user voice information.
  • the acquisition module needs to use the voice stream data obtained by the first conversion module.
  • the first conversion module may notify the acquisition module to extract the voice stream data after the conversion is completed, and may also be the first conversion.
  • the module directly sends the sound stream data to the acquisition module.
  • the specific processing procedure of the user query module is as follows: the first query operation is initiated by using the calling number of the calling user as a query condition; and the voice information of the calling user is used as The query condition initiates a second query operation; in a specific embodiment of the present invention, it may be a user who returns 80% similarity with the voice information of the calling user; of course, other values such as 90% may also be used.
  • the results of the first query operation and the second query operation are compared, and the user is finally confirmed, as described in detail below.
  • the user can be queried in the first query operation and the second query operation, and when the result is the same, the user can be directly confirmed;
  • the user can be queried in the first query operation and the second query operation, but when the result is different, the result of the second query operation is taken as the standard, and the number of the user in the database is updated;
  • the result of the second query operation is taken as the standard, and the number of the user in the database is updated;
  • the result of the first query operation is taken as the standard, and the user voice information in the database is updated;
  • the user information is newly added in the database.
  • the user evaluation result corresponding to the user can be determined.
  • the user query module can directly query the database,
  • the query operation can be processed by the data update module, and the query result is returned by the data update module.
  • the sound information is a voiceprint feature
  • the user query module uses a combination of a hidden Markov model (HMM) method and a vector quantization VQ clustering method to improve the accuracy of the user identification.
  • HMM hidden Markov model
  • determining the identity of the user is not an essential feature in the embodiment of the present invention. For example, the following is explained. Assume that it is only necessary to obtain a large user evaluation result for analysis and statistics, and it is not necessary to know which user the user evaluation result is.
  • the first conversion module/second conversion module is used to implement voice and text conversion, and can be implemented based on various mature conversion tools, such as the Microsoft speech recognition system SpeechSDK 5.1, etc., belonging to the field.
  • various mature conversion tools such as the Microsoft speech recognition system SpeechSDK 5.1, etc.
  • the text analysis module specifically includes a vocabulary saving unit, a statistical unit, and a computing unit, where
  • a vocabulary saving unit configured to save a preset vocabulary
  • the preset vocabulary may be some uncivilized term, and may also be other preset vocabulary
  • a statistical unit that counts the number of occurrences of a preset vocabulary in a text.
  • a calculating unit configured to calculate a value of the second evaluation index, where the value of the second evaluation index value is equal to the number of occurrences divided by the total number of words of the text, indicating the frequency of the user using the preset vocabulary in the service process.
  • the first evaluation index may be a ratio of the number of occurrences of the preset vocabulary in the obtained text to the number of text words MGL, MGL-number of occurrences/total number of words.
  • the voice analysis module specifically includes a parameter extraction unit and a emotion evaluation unit, where
  • a parameter extraction unit configured to extract a value of the preset parameter according to the obtained voice stream data.
  • the emotion judging unit is configured to determine the emotional parameter QG of the user according to the preset parameter. Judging the user's emotion according to the preset parameters can be implemented by using the prior art. In the specific embodiment of the present invention, the relationship between the emotion and the voice parameters summarized by Murray and Arnott is determined, as shown in Table 1 below:
  • a method such as a principal component analysis method, a neural network method, or a Gaussian mixture model (GMM) may be used to derive values of user emotion parameters according to extracted parameters and against emotion and speech parameter tables. ⁇ angry, happy, sad, fearful, disgusted ⁇ .
  • GMM Gaussian mixture model
  • emotion parameter may also be other parameters.
  • the user's emotion of receiving the service can be obtained, and the result of the current user evaluation is obtained.
  • the results include two aspects, one is the evaluation of the user's terms, and the other is the evaluation of the user's emotions.
  • the method may further include:
  • the comprehensive analysis module is configured to calculate a value for evaluating the third evaluation indicator of the user according to the historical data and the current data of the first evaluation indicator and the second evaluation indicator.
  • the third evaluation index XW MAX(S, G, B, ⁇ , Y)/the total number of services in the past half year, wherein:
  • Y the number of aversions in the first 6 months *0.4 + the number of aversions in the first 5 months *0.5 + the number of aversions in the first 4 months *0.6 + the number of aversions in the first 3 months *0.7 + the number of aversions in the first 2 months *0.8 + the first 1 Monthly aversion number *0.9
  • the user's level L after calculating the user behavior evaluation result, the user evaluation result, and the MGL, the user's level L, the user's attention level D, the emotional baseline, and the like can be further obtained, as follows:
  • Level L MGL * XW
  • Emotional baseline QG. According to the above parameters such as L, D and QG, it can provide basis for user tracking, user access and the like.
  • FIG. 4 is a schematic flowchart of a first embodiment of a voice-based client evaluation method according to the present invention. As shown in FIG. 4, the method includes:
  • Step 41 The customer evaluation system intercepts the media stream data sent by the client to the customer service center.
  • Step 42 The client evaluation system performs a conversion operation on the media stream to obtain voice stream data.
  • Step 43 The customer evaluation system analyzes the voice stream data to obtain a value of the first evaluation indicator.
  • Step 44 The customer evaluation system obtains the first user evaluation result according to the value of the first evaluation indicator.
  • FIG. 5 is a schematic flowchart of a second embodiment of a voice-based client evaluation method according to the present invention. As shown in FIG. 5, the method includes:
  • Step 51 The customer evaluation system intercepts the media stream data sent by the client to the customer service center.
  • Step 52 The customer evaluation system performs a conversion operation on the media stream to obtain voice stream data and text.
  • Step 53 The customer evaluation system analyzes the voice stream data and the text, and obtains a value of the first evaluation indicator and a value of the second evaluation indicator, respectively.
  • Step 54 The customer evaluation system acquires the second user evaluation result according to the value of the first evaluation indicator and the value of the second evaluation indicator.
  • the data interception module intercepts the agent side RTP data packet and forwards it to the conversion module.
  • the conversion module acquires the data, decodes the RTP data packet and restores it into voice stream data, and converts the voice stream data into text according to the Microsoft voice recognition system.
  • the conversion module stores the voice stream data, text, and calling number information in the memory.
  • the conversion module initiates a notification message through the data communication interface, and informs the acquisition module to extract the data.
  • the obtaining module receives the data extraction notification of the conversion module, and extracts the number information from the conversion module through the data communication interface.
  • the acquisition module extracts voice voiceprint information from the voice stream data of the conversion module through a data communication interface.
  • the user query module initiates a user number and voice voiceprint information query operation through the data communication interface; the voice voiceprint information query operation returns a maximum user information with a matching rate of 80% or more.
  • the user query module compares the returned results of the two query operations, and determines the user identity based on the comparison result, and the update module performs an update operation on the database according to the comparison result.
  • the text analysis module receives the data extraction notification of the conversion module, extracts the text information from the conversion module through the data communication interface, and after processing, initiates a query operation through the data communication interface, and then sends the hit number to the comprehensive analysis module.
  • the voice analysis module extracts the voice stream data from the conversion module through the data communication interface, obtains the evaluation index, and sends the evaluation index to the comprehensive analysis module.
  • the comprehensive analysis module initiates historical data query operations through the data communication interface through the data communication interface and the determined user information.
  • the comprehensive analysis module calculates the user's level, emotional baseline, and degree of importance based on the value of the evaluation index, the number of hits, and historical data.
  • the comprehensive analysis module updates the data in the database through the data communication interface module.
  • Database processing includes the following aspects:
  • the unit receives the historical data query message sent by the comprehensive analysis module, calling the data drive of the module
  • the unit performs a database query operation and returns the result of the query.
  • the data driving unit of the module After receiving the processing completion notification message sent by the comprehensive analysis module, the data driving unit of the module is called to perform a database operation, record the current call log, update the user level, the emotional baseline, and the like, and initiate a message to the agent, the CTI, and other systems.
  • the agent and the third-party system receive the notification message, and after processing, display the current user level, the emotional baseline, the importance level, and the like for subsequent service processing; the CTI receives the notification message as the queuing priority reference.

Landscapes

  • Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Telephonic Communication Services (AREA)

Abstract

A voice-based customer evaluation system and the method thereof are provided, wherein the system includes: a data acquisition module acquiring a media stream data that is sent to a customer service center from a client side, a first conversion module converting the media stream data to obtain voice stream data, a voice analysis module analyzing the voice stream data to acquire the value of the first evaluation index for evaluating a user. The customer evaluation system acquires the media stream data that is sent to a customer service center from a client side and converts the media stream data into voice stream data and text, then acquires the value of the evaluation index for evaluating a user based on analysis of the voice stream data and text. So, the evaluation is objective, and can provide objective evidence for tracking of the service and problem.

Description

一种基于语音的客户评价*** ^户评价方法 技术领域  A voice-based customer evaluation system
本发明涉及客户服务***， 特别是一种基于语音的客户评价***及客 户评价方法。 背景技术  The present invention relates to a customer service system, and more particularly to a voice-based customer evaluation system and a customer evaluation method. Background technique
呼叫中心又称为客户服务***， 是用于向用户提供电话、 传真、 电子 邮件等多种接入手段的信息***， 主要用来处理用户对企业提出的要求、 质疑、 投诉、 建议和咨询， 如电信业中的 1860, 金融业中的 95555等。  Call center, also known as customer service system, is an information system used to provide users with various access methods such as telephone, fax, and e-mail. It is mainly used to handle user requests, questions, complaints, suggestions, and consultations. For example, 1860 in the telecommunications industry, 95555 in the financial industry, etc.
图 1为客户服务***的结构示意图， 如图 1所示， 现有客户服务*** 包括：呼叫接入模块、计算机电话集成（ CTI, Computer Telephony Integration ) 模块、 至少一个座席模块， 和媒体服务器。 其中， 媒体服务器能够为用户 提供自动客户服务， 而在需要人工服务时， CTI模块能够根据一定的策略将 用户呼叫路由到一个座席模块， 由该座席模块为话务员提供与用户进行交 互的平台。  FIG. 1 is a schematic structural diagram of a customer service system. As shown in FIG. 1, an existing customer service system includes a call access module, a Computer Telephony Integration (CTI) module, at least one agent module, and a media server. The media server can provide automatic customer service for the user. When the manual service is required, the CTI module can route the user call to an agent module according to a certain policy, and the agent module provides the operator with a platform for interacting with the user.
客服服务***中， 关注较多的是在话务员的服务评价， 而对客户在接 受服务过程中表现出来的态度、 行为的评价以及后续跟踪解决相对忽略。 特别是对于一些特殊客户， 一般是通过话务员反馈记登记。 而人为的评价 方式， 往往受到主观影响， 个体差异明显， 不能对后续更好的服务、 问题 跟踪提供客观的依据。 发明内容  In the customer service system, more attention is paid to the service evaluation of the operator, and the attitude, behavior evaluation and follow-up tracking of the customer in the process of receiving the service are relatively ignored. Especially for some special customers, it is generally registered by the attendant feedback. The artificial evaluation method is often subjectively influenced, and the individual differences are obvious. It cannot provide an objective basis for subsequent better service and problem tracking. Summary of the invention
本发明的目的是提供一种基于语音的客户评价***及客户评价方法， 实现客观的客户评价， 为服务、 问题跟踪提供客观依据。 为了实现上述目的， 本发明实施例提供了一种基于语音的客户评价系 统， 包括： The object of the present invention is to provide a voice-based customer evaluation system and a customer evaluation method, to achieve objective customer evaluation, and to provide an objective basis for service and problem tracking. In order to achieve the above object, an embodiment of the present invention provides a voice-based client evaluation system, including:
一种基于语音的客户评价***， 包括数据截获模块、 第一转换模块和 语音分析模块， 其中，  A voice-based customer evaluation system includes a data intercepting module, a first converting module, and a voice analyzing module, wherein
数据截获模块， 用于截获用户端发送到客户服务中心的媒体流数据； 第一转换模块， 用于对截获的媒体流进行转换操作， 得到语音流数据； 语音分析模块， 用于分析得到的语音流数据， 获取用于评价用户的第 一评价指标的值。  a data intercepting module, configured to intercept media stream data sent by the user end to the customer service center; a first converting module, configured to perform a converting operation on the intercepted media stream to obtain voice stream data; and a voice analyzing module, configured to analyze the obtained voice The stream data is obtained by evaluating a value of the first evaluation indicator of the user.
该客户评价***还包括第二转换模块和文本分析模块， 其中， 第二转换模块， 用于对所述语音流数据进行转换操作， 得到文本； 文本分析模块， 用于分析得到的文本， 获取用于评价用户的第二评价 指标的值。  The customer evaluation system further includes a second conversion module and a text analysis module, wherein the second conversion module is configured to perform a conversion operation on the voice stream data to obtain a text; a text analysis module, configured to analyze the obtained text, and obtain The value of the second evaluation indicator of the user is evaluated.
该客户评价***还包括：  The customer evaluation system also includes:
数据库， 用于存储用户信息；  a database for storing user information;
获取模块， 用于获取当前呼叫的主叫用户信息；  An obtaining module, configured to obtain caller information of the current call;
用户查询模块， 用于根据所述当前呼叫的主叫用户信息查询所述数据 库， 确定用户身份；  a user querying module, configured to query the database according to the calling user information of the current call, and determine a user identity;
更新模块， 用于在根据所述当前呼叫的主叫用户信息无法查询到用户 时， 根据所述当前呼叫的主叫用户信息对所述数据库执行更新操作。  And an update module, configured to perform an update operation on the database according to the calling user information of the current call when the user cannot be queried according to the calling user information of the current call.
所述用户信息为用户号码和 /或声音信息。  The user information is a subscriber number and/or voice information.
所述文本分析模块具体包括词汇保存单元、 统计单元和计算单元， 其 中，  The text analysis module specifically includes a vocabulary saving unit, a statistical unit, and a calculating unit, wherein
词汇保存单元， 用于保存预设词汇；  a vocabulary saving unit for saving a preset vocabulary;
统计单元， 用于统计所述文本中所述预设词汇的出现次数；  a statistical unit, configured to count the number of occurrences of the preset vocabulary in the text;
计算单元， 用于计算所述第二评价指标的值， 所述第二评价指标值的 值等于所述出现次数除以所述文本的总字数。 a calculating unit, configured to calculate a value of the second evaluation index, where the second evaluation index value The value is equal to the number of occurrences divided by the total number of words of the text.
所述语音分析模块具体包括参数提取单元和情绪评判单元， 其中， 参数提取单元， 用于根据所述语音流数据提取预设参数的值； 情绪评判单元， 用于根据提取的预设参数的值确定所述第一评价指标 的值， 所述第一评价指标为用户情绪指标。  The voice analyzing module specifically includes a parameter extracting unit and a sentiment judging unit, wherein the parameter extracting unit is configured to extract a value of the preset parameter according to the voice stream data; and the emotion judging unit is configured to use the value of the extracted preset parameter Determining a value of the first evaluation indicator, where the first evaluation indicator is a user emotion indicator.
所述参数提取单元， 具体用于利用主元分析方法、 神经网络方法或高 斯混合模型方法来提取所述预设参数的值。  The parameter extraction unit is specifically configured to extract a value of the preset parameter by using a principal component analysis method, a neural network method, or a Gaussian mixture model method.
该客户评价***还包括：  The customer evaluation system also includes:
综合分析模块， 用于根据所述第一评价指标和第二评价指标的历史数 据和当前数据计算用于评价用户的第三评价指标的值。  The comprehensive analysis module is configured to calculate a value for evaluating the third evaluation indicator of the user according to the historical data and the current data of the first evaluation index and the second evaluation index.
一种基于语音的客户评价方法， 包括：  A voice-based customer evaluation method, including:
客户评价***截获用户端发送到客户服务中心的媒体流数据； 客户评价***对所述媒体流进行转换操作， 得到语音流数据； 客户评价***分析所述语音流数据， 获取用于评价用户的第一评价指 标的值。  The customer evaluation system intercepts media stream data sent by the client to the customer service center; the customer evaluation system performs a conversion operation on the media stream to obtain voice stream data; and the customer evaluation system analyzes the voice stream data to obtain a number for evaluating the user. The value of an evaluation indicator.
该方法还包括：  The method also includes:
所述客户评价***对所述语音流数据进行转换操作， 得到文本； 所述客户评价***分析所述文本， 获取用于评价用户的第二评价指标 的值。  The customer evaluation system performs a conversion operation on the voice stream data to obtain a text; the client evaluation system analyzes the text to obtain a value for evaluating a second evaluation indicator of the user.
该方法还包括：  The method also includes:
所述客户评价***根据所述第一评价指标和第二评价指标的历史数据 和当前数据计算用于评价用户的第三评价指标的值。  The customer evaluation system calculates a value for evaluating a third evaluation index of the user based on historical data and current data of the first evaluation index and the second evaluation index.
本发明实施例具有以下的有益效果：  The embodiments of the present invention have the following beneficial effects:
本发明的客户评价***通过截获用户端发送到客户服务中心的媒体流 数据， 并将其转换为语音流数据和文本， 基于该语音流和文本进行分析， 获取用于评价用户的评价指标的值， 因此其评价是客观的， 为服务、 问题 跟踪提供了客观依据。 附图说明 The customer evaluation system of the present invention intercepts the media stream data sent by the client to the customer service center, converts it into voice stream data and text, and analyzes the voice stream and the text based on the voice stream and the text. Obtaining the value of the evaluation index used to evaluate the user, so the evaluation is objective and provides an objective basis for service and problem tracking. DRAWINGS
图 1为客户服务***的结构示意图；  Figure 1 is a schematic structural diagram of a customer service system;
图 2 为本发明基于语音的客户评价***的第一实施例的结构示意图； 图 3为本发明基于语音的客户评价***的第二实施例的结构示意图； 图 4为本发明基于语音的客户评价方法的第一实施例的流程示意图； 图 5为本发明基于语音的客户评价方法的第二实施例的流程示意图。 具体实施方式  2 is a schematic structural diagram of a first embodiment of a voice-based customer evaluation system according to the present invention; FIG. 3 is a schematic structural diagram of a second embodiment of a voice-based customer evaluation system according to the present invention; FIG. 4 is a voice-based customer evaluation of the present invention. A schematic flowchart of a first embodiment of the method; FIG. 5 is a schematic flowchart of a second embodiment of a voice-based client evaluation method according to the present invention. detailed description
本发明实施例中， 通过截取用户的语音数据， 并对语音数据进行分析， 得到用于进行用户评价的特征值， 并基于该特征值进行用户评价。  In the embodiment of the present invention, the user's voice data is intercepted, and the voice data is analyzed, and the feature value for performing the user evaluation is obtained, and the user evaluation is performed based on the feature value.
图 2 为本发明基于语音的客户评价***的第一实施例的结构示意图， 如图 2所示， 包括数据截获模块、 第一转换模块和语音分析模块， 其中， 数据截获模块， 用于截获用户端发送到客户服务中心的媒体流数据。 第一转换模块， 用于对截获的媒体流数据进行转换操作， 得到语音流 数据。  2 is a schematic structural diagram of a first embodiment of a voice-based client evaluation system according to the present invention. As shown in FIG. 2, the system includes a data intercepting module, a first converting module, and a voice analyzing module, where the data intercepting module is configured to intercept the user. The media stream data sent to the customer service center. The first conversion module is configured to perform a conversion operation on the intercepted media stream data to obtain voice stream data.
语音分析模块， 用于分析得到的语音流数据， 获取用于评价用户的第 一评价指标的值。  The voice analysis module is configured to analyze the obtained voice stream data, and obtain a value for evaluating the first evaluation indicator of the user.
从图 2所示的客户评价***可以看出， 其利用语音流数据进行分析， 获取第一用户评价结果， 而从语音流数据中可以分析出用户的情绪信息， 为了提高评价的准确性和全面性， 本发明第二实施例的基于语音的客户评 价***还通过用户的语言来进行评价， 图 3 为本发明基于语音的客户评价 ***的第二实施例的结构示意图， 如图 3 所示， 包括数据截获模块、 第一 转换模块、 第二转换模块、 语音分析模块和文本分析模块， 其中， 数据截获模块， 用于截获用户端发送到客户服务中心的媒体流数据。 第一转换模块， 用于对截获的媒体流数据进行转换操作， 得到语音流 数据。 It can be seen from the customer evaluation system shown in FIG. 2 that it uses voice stream data for analysis to obtain the first user evaluation result, and the user's emotion information can be analyzed from the voice stream data, in order to improve the accuracy and comprehensiveness of the evaluation. The voice-based customer evaluation system of the second embodiment of the present invention is also evaluated by the language of the user. FIG. 3 is a schematic structural diagram of a second embodiment of the voice-based customer evaluation system of the present invention, as shown in FIG. The data intercepting module, the first converting module, the second converting module, the speech analyzing module, and the text analyzing module, wherein The data intercepting module is configured to intercept media stream data sent by the client to the customer service center. The first conversion module is configured to perform a conversion operation on the intercepted media stream data to obtain voice stream data.
第二转换模块， 用于对得到的语音流数据进行转换操作， 得到文本； 语音分析模块， 用于分析得到的语音流数据， 获取第一评价指标的值； 文本分析模块， 用于分析得到的文本， 获取用于评价用户的第二评价 指标的值。  a second conversion module, configured to perform a conversion operation on the obtained voice stream data to obtain a text; a voice analysis module, configured to analyze the obtained voice stream data, and obtain a value of the first evaluation index; and a text analysis module, configured to analyze Text, obtaining a value for evaluating the second evaluation indicator of the user.
上述媒体流数据可以是现有技术中任意一种支持媒体流传输的协议下 的媒体流数据， 在本发明的具体实施例中以实时传输协议（RTP, Real-time Transport Protocol )为例进行详细说明，但并不用于限定本发明的保护范围。  The foregoing media stream data may be media stream data in a protocol for supporting media stream transmission in any of the prior art. In a specific embodiment of the present invention, a real-time transport protocol (RTP) is taken as an example for detailed description. It is intended to be illustrative, but not intended to limit the scope of the invention.
在本发明的具体实施例中， 在对单一用户进行评价时， 必然需要确定 用户信息， 因此， 本发明实施例的客户评价***还包括（为简化附图， 图 2 和图 3中未示出）：  In a specific embodiment of the present invention, when evaluating a single user, it is necessary to determine user information. Therefore, the customer evaluation system of the embodiment of the present invention further includes (not shown in FIG. 2 and FIG. 3 for simplification of the drawing. ):
数据库， 用于存储用户信息。 其中， 用户信息为用户号码和 /或声音信 息。  Database, used to store user information. The user information is a user number and/or voice information.
获取模块， 用于获取当前呼叫的主叫用户信息。  The obtaining module is configured to obtain caller information of the current call.
用户查询模块， 用于根据当前呼叫的主叫用户信息查询数据库， 确定 用户身份。  The user query module is configured to query the database according to the calling user information of the current call to determine the identity of the user.
更新模块， 用于在根据当前呼叫的主叫用户信息无法查询到用户时， 根据当前呼叫的主叫用户信息对数据库执行更新操作。  The update module is configured to perform an update operation on the database according to the calling user information of the current call when the user cannot be queried according to the calling user information of the current call.
下面以用户信息为用户号码和声音信息为例进行详细说明。 数据库中 对应保存了用户号码与所述用户的声音信息。  The following takes the user information as the user number and voice information as an example for detailed description. The user's voice information is saved in the database.
获取模块， 用于获取主叫用户号码， 并从语音流数据提取主叫用户声 音信息。  The obtaining module is configured to obtain a calling party number, and extract the calling user voice information from the voice stream data.
用户查询模块， 根据主叫用户号码和主叫用户声音信息查询数据库， 确定用户身份。 The user query module queries the database according to the calling user number and the calling user voice information. Determine the identity of the user.
更新模块， 用于在根据主叫用户号码或主叫用户声音信息无法查询到 用户时， 根据主叫用户号码或主叫用户声音信息对数据库执行更新操作。  The update module is configured to perform an update operation on the database according to the calling party number or the calling user voice information when the user cannot be queried according to the calling party number or the calling user voice information.
获取模块需要用到第一转换模块得到的语音流数据， 在本发明的具体 实施例中， 可以由第一转换模块在转换完成后通知获取模块提取该语音流 数据， 当然也可以是第一转换模块直接将该声音流数据发送给获取模块。  The acquisition module needs to use the voice stream data obtained by the first conversion module. In a specific embodiment of the present invention, the first conversion module may notify the acquisition module to extract the voice stream data after the conversion is completed, and may also be the first conversion. The module directly sends the sound stream data to the acquisition module.
在本发明的具体实施例中， 该用户查询模块的具体处理过程如下所述： 以所述主叫用户的主叫号码作为查询条件发起第一查询操作； 以所述主叫用户的声音信息作为查询条件发起第二查询操作； 在本发 明的具体实施例中， 可以是返回与主叫用户的声音信息具有 80%相似度的 用户； 当然， 也可以是 90%等其它数值。  In a specific embodiment of the present invention, the specific processing procedure of the user query module is as follows: the first query operation is initiated by using the calling number of the calling user as a query condition; and the voice information of the calling user is used as The query condition initiates a second query operation; in a specific embodiment of the present invention, it may be a user who returns 80% similarity with the voice information of the calling user; of course, other values such as 90% may also be used.
比较第一查询操作和第二查询操作的结果， 最终确认用户， 详细如下 所述。  The results of the first query operation and the second query operation are compared, and the user is finally confirmed, as described in detail below.
在第一查询操作和第二查询操作均可以查询到用户， 且结果相同时， 可以直接确认用户；  The user can be queried in the first query operation and the second query operation, and when the result is the same, the user can be directly confirmed;
在第一查询操作和第二查询操作均可以查询到用户， 但结果不同时， 以第二查询操作的结果为准， 并更新数据库中该用户的号码；  The user can be queried in the first query operation and the second query operation, but when the result is different, the result of the second query operation is taken as the standard, and the number of the user in the database is updated;
在第一查询操作无法查询到用户， 但第二查询操作能查询到用户时， 以第二查询操作的结果为准， 并更新数据库中该用户的号码；  When the first query operation cannot query the user, but the second query operation can query the user, the result of the second query operation is taken as the standard, and the number of the user in the database is updated;
在第一查询操作可以查询到用户， 但第二查询操作无法查询到用户时， 以第一查询操作的结果为准， 并更新数据库中的用户声音信息；  When the first query operation can query the user, but the second query operation cannot query the user, the result of the first query operation is taken as the standard, and the user voice information in the database is updated;
在第一查询操作和第二查询操作均无法查询到用户时， 在数据库中新 增该用户信息。  When the first query operation and the second query operation fail to query the user, the user information is newly added in the database.
确定用户身份后， 即可确定该用户对应的用户评价结果。  After determining the identity of the user, the user evaluation result corresponding to the user can be determined.
在本发明的具体实施例中， 该用户查询模块可以直接查询数据库， 也 可以通过数据更新模块来处理该查询操作， 并由数据更新模块返回查询结 果。 In a specific embodiment of the present invention, the user query module can directly query the database, The query operation can be processed by the data update module, and the query result is returned by the data update module.
在本发明具体实施例中， 声音信息为声紋特征， 用户查询模块釆用隐 式马尔可夫模型 （HMM )方法和矢量量化 VQ聚类方法相结合的方法， 提 高了用户身份识别的准确率。  In the specific embodiment of the present invention, the sound information is a voiceprint feature, and the user query module uses a combination of a hidden Markov model (HMM) method and a vector quantization VQ clustering method to improve the accuracy of the user identification.
当然， 确定用户身份并不是本发明实施例中的必要特征， 举例说明如 下。 假定当前仅需要获取一个大的用户评价结果来进行分析统计， 此时就 不需要知道该用户评价结果到底是哪个用户的。  Of course, determining the identity of the user is not an essential feature in the embodiment of the present invention. For example, the following is explained. Assume that it is only necessary to obtain a large user evaluation result for analysis and statistics, and it is not necessary to know which user the user evaluation result is.
在本发明的具体实施例中， 第一转换模块 /第二转换模块用于实现语音 和文本的转换， 可以基于各种成熟的转换工具来实现， 如微软语音识别系 统 SpeechSDK 5.1等， 属于本领域技术人员惯用技术手段， 这里不再赘述。  In a specific embodiment of the present invention, the first conversion module/second conversion module is used to implement voice and text conversion, and can be implemented based on various mature conversion tools, such as the Microsoft speech recognition system SpeechSDK 5.1, etc., belonging to the field. The technical personnel are accustomed to technical means, and will not be described here.
在本发明的具体实施例中， 文本分析模块具体包括词汇保存单元、 统 计单元和计算单元， 其中，  In a specific embodiment of the present invention, the text analysis module specifically includes a vocabulary saving unit, a statistical unit, and a computing unit, where
词汇保存单元， 用于保存预设词汇； 该预设词汇可以是一些不文明用 语， 当然也可以是其它预先设置的词汇；  a vocabulary saving unit, configured to save a preset vocabulary; the preset vocabulary may be some uncivilized term, and may also be other preset vocabulary;
统计单元， 用于统计文本中预设词汇的出现次数。  A statistical unit that counts the number of occurrences of a preset vocabulary in a text.
计算单元， 用于计算第二评价指标的值， 所述第二评价指标值的值等 于出现次数除以文本的总字数， 表明用户在服务过程中使用预设词汇的频 率。  And a calculating unit, configured to calculate a value of the second evaluation index, where the value of the second evaluation index value is equal to the number of occurrences divided by the total number of words of the text, indicating the frequency of the user using the preset vocabulary in the service process.
在本发明的具体实施例中， 第一评价指标可以是预设词汇在得到的文 本中出现的次数与文本字数的比例 MGL, MGL-出现次数 /总字数。  In a specific embodiment of the present invention, the first evaluation index may be a ratio of the number of occurrences of the preset vocabulary in the obtained text to the number of text words MGL, MGL-number of occurrences/total number of words.
在本发明的具体实施例中， 语音分析模块具体包括参数提取单元和情 绪评判单元， 其中，  In a specific embodiment of the present invention, the voice analysis module specifically includes a parameter extraction unit and a emotion evaluation unit, where
参数提取单元， 用于根据得到的语音流数据提取预设参数的值。  And a parameter extraction unit, configured to extract a value of the preset parameter according to the obtained voice stream data.
情绪评判单元， 用于根据预设参数确定用户的情绪参数 QG。 根据预设参数来判断用户情绪可以釆用现有技术实现， 在本发明具体 实施例中釆用 Murray和 Arnott总结的情感和语音参数的关系来确定，如下 表 1所示： The emotion judging unit is configured to determine the emotional parameter QG of the user according to the preset parameter. Judging the user's emotion according to the preset parameters can be implemented by using the prior art. In the specific embodiment of the present invention, the relationship between the emotion and the voice parameters summarized by Murray and Arnott is determined, as shown in Table 1 below:
Figure imgf000010_0001
Figure imgf000010_0001
表 1  Table 1
上述的非常高、 很高、 很宽等描述最后都是基于一个阔值来确定的。 在本发明的具体实施例中， 可以釆用主元分析方法、 神经网络方法或 高斯混合模型 （GMM )等方法来根据提取的参数， 并对照情感和语音参数 表得出用户情绪参数的值， {生气、 高兴、 悲伤、 恐惧、 厌恶 }。  The above descriptions of very high, very high, and very wide are finally determined based on a threshold. In a specific embodiment of the present invention, a method such as a principal component analysis method, a neural network method, or a Gaussian mixture model (GMM) may be used to derive values of user emotion parameters according to extracted parameters and against emotion and speech parameter tables. {angry, happy, sad, fearful, disgusted}.
上述仅仅是一种举例说明， 该情绪参数还可以是其它的参数。  The above is just an example, and the emotion parameter may also be other parameters.
利用上述的说明， 就可以得到用户本次接受服务的情绪， 得到本次的 用户评价的结果。 该结果包括两方面的内容， 一方面是用户的用语方面的 评价， 一方面是用户情绪方面的评价。  By using the above description, the user's emotion of receiving the service can be obtained, and the result of the current user evaluation is obtained. The results include two aspects, one is the evaluation of the user's terms, and the other is the evaluation of the user's emotions.
在本发明的具体实施例中， 还可以包括：  In a specific embodiment of the present invention, the method may further include:
综合分析模块， 用于根据所述第一评价指标和第二评价指标的历史数 据和当前数据计算用于评价用户的第三评价指标的值。 在本发明的具体实施例中，第三评价指标 XW = MAX(S, G, B, Κ, Y)/近 半年总服务数， 其中： The comprehensive analysis module is configured to calculate a value for evaluating the third evaluation indicator of the user according to the historical data and the current data of the first evaluation indicator and the second evaluation indicator. In a specific embodiment of the present invention, the third evaluation index XW = MAX(S, G, B, Κ, Y)/the total number of services in the past half year, wherein:
S = 前 6个月生气数 *0.4 +前 5个月生气数 *0.5 +前 4个月生气数 *0.6 +前 3个月生气数 *0.7 +前 2个月生气数 *0.8 +前 1个月生气数 *0.9 +当前生气数；  S = number of anger in the first 6 months *0.4 + number of anger in the first 5 months *0.5 + number of anger in the first 4 months *0.6 + number of anger in the first 3 months *0.7 + number of anger in the first 2 months *0.8 +1 Monthly anger count *0.9 + current anger count;
G =前 6个月高兴数 *0.4 +前 5个月高兴数 *0.5 +前 4个月高兴数 *0.6 +前 3个月高兴数 *0.7 +前 2个月高兴数 *0.8 +前 1个月高兴数 *0.9 +当前高兴数；  G = happy number in the first 6 months *0.4 + happy number in the first 5 months *0.5 + happy in the first 4 months *0.6 + happy in the first 3 months *0.7 + happy in the first 2 months *0.8 +1 Month happy number *0.9 + current happy number;
B =前 6个月悲伤数 *0.4 +前 5个月悲伤数 *0.5 +前 4个月悲伤数 *0.6 +前 3个月悲伤数 *0.7 +前 2个月悲伤数 *0.8 +前 1个月悲伤数 *0.9 +当前悲伤数；  B = sadness in the first 6 months *0.4 + sadness in the first 5 months *0.5 + sadness in the first 4 months *0.6 + sadness in the first 3 months *0.7 + sadness in the first 2 months *0.8 + first 1 Month sadness number *0.9 + current sad number;
K =前 6个月恐惧数 *0.4 +前 5个月恐惧数 *0.5 +前 4个月恐惧数 *0.6 +前 3个月恐惧数 *0.7 +前 2个月恐惧数 *0.8 +前 1个月恐惧数 *0.9 +当前恐惧数；  K = fear in the first 6 months *0.4 + fear in the first 5 months *0.5 + fear in the first 4 months *0.6 + fear in the first 3 months *0.7 + fear in the first 2 months *0.8 +1 Monthly fear number * 0.9 + current fear number;
Y =前 6个月厌恶数 *0.4 +前 5个月厌恶数 *0.5 +前 4个月厌恶数 *0.6 +前 3个月厌恶数 *0.7 +前 2个月厌恶数 *0.8 +前 1个月厌恶数 *0.9 Y = the number of aversions in the first 6 months *0.4 + the number of aversions in the first 5 months *0.5 + the number of aversions in the first 4 months *0.6 + the number of aversions in the first 3 months *0.7 + the number of aversions in the first 2 months *0.8 + the first 1 Monthly aversion number *0.9
+当前厌恶数。 + Current aversion.
需要说明的是， 上述的时间、 系数仅仅是举例说明， 本发明实施例中 并不受限于上述的举例说明。  It should be noted that the above-mentioned time and coefficient are merely examples, and the embodiments of the present invention are not limited to the above examples.
在本发明的具体实施例中， 计算出上述的用户行为评价结果、 用户评 价结果和 MGL之后， 即可进一步得到用户的级别 L、 用户关注程度 D、 情 感基线等， 如下：  In a specific embodiment of the present invention, after calculating the user behavior evaluation result, the user evaluation result, and the MGL, the user's level L, the user's attention level D, the emotional baseline, and the like can be further obtained, as follows:
级别 L = MGL * XW;  Level L = MGL * XW;
关注程度 D = XW;  Degree of concern D = XW;
情感基线： QG。 根据上述的 L、 D和 QG等参数， 可以为用户跟踪、 用户接入等提供依 据。 Emotional baseline: QG. According to the above parameters such as L, D and QG, it can provide basis for user tracking, user access and the like.
在本发明的具体实施例中， 并没有描述或者在附图中示出模块之间的 接口。  In the specific embodiments of the present invention, the interfaces between the modules are not described or shown in the drawings.
图 4为本发明基于语音的客户评价方法的第一实施例的流程示意图， 如图 4所示， 包括：  4 is a schematic flowchart of a first embodiment of a voice-based client evaluation method according to the present invention. As shown in FIG. 4, the method includes:
步骤 41 :客户评价***截获用户端发送到客户服务中心的媒体流数据。 步骤 42: 客户评价***对所述媒体流进行转换操作，得到语音流数据。 步骤 43: 客户评价***分析所述语音流数据， 获取第一评价指标的值。 步骤 44: 客户评价***根据所述第一评价指标的值获取第一用户评价 结果。  Step 41: The customer evaluation system intercepts the media stream data sent by the client to the customer service center. Step 42: The client evaluation system performs a conversion operation on the media stream to obtain voice stream data. Step 43: The customer evaluation system analyzes the voice stream data to obtain a value of the first evaluation indicator. Step 44: The customer evaluation system obtains the first user evaluation result according to the value of the first evaluation indicator.
图 5 为本发明基于语音的客户评价方法的第二实施例的流程示意图， 如图 5所示， 包括：  FIG. 5 is a schematic flowchart of a second embodiment of a voice-based client evaluation method according to the present invention. As shown in FIG. 5, the method includes:
步骤 51 :客户评价***截获用户端发送到客户服务中心的媒体流数据。 步骤 52: 客户评价***对所述媒体流进行转换操作， 得到语音流数据 和文本。  Step 51: The customer evaluation system intercepts the media stream data sent by the client to the customer service center. Step 52: The customer evaluation system performs a conversion operation on the media stream to obtain voice stream data and text.
步骤 53: 客户评价***分析所述语音流数据和所述文本， 分别获取第 一评价指标的值和第二评价指标的值。  Step 53: The customer evaluation system analyzes the voice stream data and the text, and obtains a value of the first evaluation indicator and a value of the second evaluation indicator, respectively.
步骤 54: 客户评价***根据所述第一评价指标的值和第二评价指标的 值获取第二用户评价结果。  Step 54: The customer evaluation system acquires the second user evaluation result according to the value of the first evaluation indicator and the value of the second evaluation indicator.
下面对本发明实施例的详细流程进行说明， 包括：  The detailed process of the embodiment of the present invention is described below, including:
数据截获模块截取座席侧 RTP数据包， 然后转发给转换模块。  The data interception module intercepts the agent side RTP data packet and forwards it to the conversion module.
转换模块获取数据， 将 RTP数据包解码后还原成语音流数据， 并按照 微软语音识别***将语音流数据转换成文字。  The conversion module acquires the data, decodes the RTP data packet and restores it into voice stream data, and converts the voice stream data into text according to the Microsoft voice recognition system.
转换模块将语音流数据、 文字、 主叫号码信息存放在内存中。 转换模块通过数据通讯接口发起通知消息， 告知获取模块提取数据。 获取模块收到转换模块的数据提取通知， 通过数据通讯接口， 从转换 模块提取号码信息。 The conversion module stores the voice stream data, text, and calling number information in the memory. The conversion module initiates a notification message through the data communication interface, and informs the acquisition module to extract the data. The obtaining module receives the data extraction notification of the conversion module, and extracts the number information from the conversion module through the data communication interface.
获取模块通过数据通讯接口， 从转换模块的语音流数据提取语音声紋 信息。  The acquisition module extracts voice voiceprint information from the voice stream data of the conversion module through a data communication interface.
用户查询模块通过数据通讯接口发起用户号码和语音声紋信息查询操 作； 该语音声紋信息查询操作返回匹配率 80 %以上的最大值用户信息。  The user query module initiates a user number and voice voiceprint information query operation through the data communication interface; the voice voiceprint information query operation returns a maximum user information with a matching rate of 80% or more.
用户查询模块比较两个查询操作的返回结果， 并根据比较结果确定用 户身份， 并由更新模块根据比较结果对数据库执行更新操作。  The user query module compares the returned results of the two query operations, and determines the user identity based on the comparison result, and the update module performs an update operation on the database according to the comparison result.
文本分析模块收到转换模块的数据提取通知， 通过数据通讯接口， 从 转换模块提取文本信息， 经处理后通过数据通讯接口， 发起查询操作， 再 将命中数发给综合分析模块。  The text analysis module receives the data extraction notification of the conversion module, extracts the text information from the conversion module through the data communication interface, and after processing, initiates a query operation through the data communication interface, and then sends the hit number to the comprehensive analysis module.
语音分析模块通过数据通讯接口， 从转换模块提取语音流数据， 获取 评价指标， 发送给综合分析模块。  The voice analysis module extracts the voice stream data from the conversion module through the data communication interface, obtains the evaluation index, and sends the evaluation index to the comprehensive analysis module.
综合分析模块通过数据通讯接口， 利用确定的用户信息， 通过数据通 讯接口发起历史数据查询操作。  The comprehensive analysis module initiates historical data query operations through the data communication interface through the data communication interface and the determined user information.
综合分析模块根据评价指标的值、 命中次数、 历史数据计算该用户的 级别、 情感基线、 重视程度等。  The comprehensive analysis module calculates the user's level, emotional baseline, and degree of importance based on the value of the evaluation index, the number of hits, and historical data.
综合分析模块通过数据通讯接口模块， 更新数据库中的数据。  The comprehensive analysis module updates the data in the database through the data communication interface module.
数据库处理包括如下几个方面：  Database processing includes the following aspects:
接收到用户查询模块发送的查询消息， 调用数据驱动单元执行数据库 查询操作， 并返回查询结果。  Receiving the query message sent by the user query module, invoking the data driving unit to perform a database query operation, and returning the query result.
接收到更新模块的用户更新消息， 调用数据驱动单元执行数据库更新 操作， 并返回更新结果。  Receiving the user update message of the update module, invoking the data drive unit to perform a database update operation, and returning the update result.
接收到综合分析模块发送的历史数据查询消息， 调用本模块的数据驱 动单元执行数据库查询操作， 并返回查询结果。 Receiving the historical data query message sent by the comprehensive analysis module, calling the data drive of the module The unit performs a database query operation and returns the result of the query.
接收到综合分析模块发送的处理完毕通知消息， 调用本模块的数据驱 动单元执行数据库操作， 记录本次通话日志， 更新用户级别、 情感基线等 参数， 并向座席、 CTI、 其他***发起消息。 座席和第三方***接收通知消 息， 处理后显示当前用户级别、 情感基线、 重视程度等信息， 以供后续服 务处理； CTI接收到通知消息， 做为排队优先级参考。  After receiving the processing completion notification message sent by the comprehensive analysis module, the data driving unit of the module is called to perform a database operation, record the current call log, update the user level, the emotional baseline, and the like, and initiate a message to the agent, the CTI, and other systems. The agent and the third-party system receive the notification message, and after processing, display the current user level, the emotional baseline, the importance level, and the like for subsequent service processing; the CTI receives the notification message as the queuing priority reference.
以上所述仅是本发明的优选实施方式， 应当指出， 对于本技术领域的 普通技术人员来说， 在不脱离本发明原理的前提下， 还可以作出若干改进 和润饰， 这些改进和润饰也应视为本发明的保护范围。  The above description is only a preferred embodiment of the present invention, and it should be noted that those skilled in the art can also make several improvements and retouchings without departing from the principles of the present invention. It is considered as the scope of protection of the present invention.

Claims

权利要求书 Claim
1. 一种基于语音的客户评价***， 其特征在于， 包括数据截获模块、 第一转换模块和语音分析模块， 其中，  A voice-based customer evaluation system, comprising: a data intercepting module, a first converting module, and a voice analyzing module, wherein
数据截获模块， 用于截获用户端发送到客户服务中心的媒体流数据； 第一转换模块， 用于对截获的媒体流进行转换操作， 得到语音流数据； 语音分析模块， 用于分析得到的语音流数据， 获取用于评价用户的第 一评价指标的值。  a data intercepting module, configured to intercept media stream data sent by the user end to the customer service center; a first converting module, configured to perform a converting operation on the intercepted media stream to obtain voice stream data; and a voice analyzing module, configured to analyze the obtained voice The stream data is obtained by evaluating a value of the first evaluation indicator of the user.
2. 根据权利要求 1所述的客户评价***， 其特征在于， 该客户评价系 统还包括第二转换模块和文本分析模块， 其中，  2. The customer evaluation system according to claim 1, wherein the customer evaluation system further comprises a second conversion module and a text analysis module, wherein
第二转换模块， 用于对所述语音流数据进行转换操作， 得到文本； 文本分析模块， 用于分析得到的文本， 获取用于评价用户的第二评价 指标的值。  a second conversion module, configured to perform a conversion operation on the voice stream data to obtain a text; and a text analysis module, configured to analyze the obtained text, and obtain a value for evaluating a second evaluation indicator of the user.
3. 根据权利要求 1或 2所述的客户评价***， 其特征在于， 该客户评 价***还包括：  3. The customer evaluation system according to claim 1 or 2, wherein the customer evaluation system further comprises:
数据库， 用于存储用户信息；  a database for storing user information;
获取模块， 用于获取当前呼叫的主叫用户信息；  An obtaining module, configured to obtain caller information of the current call;
用户查询模块， 用于根据所述当前呼叫的主叫用户信息查询所述数据 库， 确定用户身份；  a user querying module, configured to query the database according to the calling user information of the current call, and determine a user identity;
更新模块， 用于在根据所述当前呼叫的主叫用户信息无法查询到用户 时， 根据所述当前呼叫的主叫用户信息对所述数据库执行更新操作。  And an update module, configured to perform an update operation on the database according to the calling user information of the current call when the user cannot be queried according to the calling user information of the current call.
4. 根据权利要求 3所述的客户评价***， 其特征在于， 所述用户信息 为用户号码和 /或声音信息。  4. The customer evaluation system according to claim 3, wherein the user information is a user number and/or voice information.
5. 根据权利要求 2所述的客户评价***， 其特征在于， 所述文本分析 模块具体包括词汇保存单元、 统计单元和计算单元， 其中，  The customer evaluation system according to claim 2, wherein the text analysis module specifically includes a vocabulary saving unit, a statistical unit, and a calculating unit, wherein
词汇保存单元， 用于保存预设词汇； 统计单元， 用于统计所述文本中所述预设词汇的出现次数； 计算单元， 用于计算所述第二评价指标的值， 所述第二评价指标值的 值等于所述出现次数除以所述文本的总字数。 a vocabulary saving unit for saving a preset vocabulary; a statistical unit, configured to count the number of occurrences of the preset vocabulary in the text; a calculating unit, configured to calculate a value of the second evaluation index, where the value of the second evaluation index value is equal to the number of occurrences divided by The total number of words of the text.
6. 根据权利要求 5所述的客户评价***， 其特征在于， 所述语音分析 模块具体包括参数提取单元和情绪评判单元， 其中，  The customer evaluation system according to claim 5, wherein the voice analysis module specifically includes a parameter extraction unit and an emotion evaluation unit, where
参数提取单元， 用于根据所述语音流数据提取预设参数的值； 情绪评判单元， 用于根据提取的预设参数的值确定所述第一评价指标 的值， 所述第一评价指标为用户情绪指标。  a parameter extracting unit, configured to extract a value of the preset parameter according to the voice stream data; and an emotion judging unit, configured to determine a value of the first evaluation index according to the value of the extracted preset parameter, where the first evaluation index is User sentiment indicators.
7. 根据权利要求 6所述的客户评价***， 其特征在于， 所述参数提取 单元， 具体用于利用主元分析方法、 神经网络方法或高斯混合模型方法来 提取所述预设参数的值。  The customer evaluation system according to claim 6, wherein the parameter extraction unit is specifically configured to extract a value of the preset parameter by using a principal component analysis method, a neural network method, or a Gaussian mixture model method.
8. 根据权利要求 6所述的客户评价***， 其特征在于， 该客户评价系 统还包括：  8. The customer evaluation system according to claim 6, wherein the customer evaluation system further comprises:
综合分析模块， 用于根据所述第一评价指标和第二评价指标的历史数 据和当前数据计算用于评价用户的第三评价指标的值。  The comprehensive analysis module is configured to calculate a value for evaluating the third evaluation indicator of the user according to the historical data and the current data of the first evaluation index and the second evaluation index.
9. 一种基于语音的客户评价方法， 其特征在于， 包括：  9. A voice-based customer evaluation method, comprising:
客户评价***截获用户端发送到客户服务中心的媒体流数据； 客户评价***对所述媒体流进行转换操作， 得到语音流数据； 客户评价***分析所述语音流数据， 获取用于评价用户的第一评价指 标的值。  The customer evaluation system intercepts media stream data sent by the client to the customer service center; the customer evaluation system performs a conversion operation on the media stream to obtain voice stream data; and the customer evaluation system analyzes the voice stream data to obtain a number for evaluating the user. The value of an evaluation indicator.
10. 根据权利要求 9所述的客户评价方法， 其特征在于， 该方法还包 括：  10. The customer evaluation method according to claim 9, wherein the method further comprises:
所述客户评价***对所述语音流数据进行转换操作， 得到文本； 所述客户评价***分析所述文本， 获取用于评价用户的第二评价指标 的值。 The customer evaluation system performs a conversion operation on the voice stream data to obtain a text; the client evaluation system analyzes the text to obtain a value for evaluating a second evaluation indicator of the user.
11. 根据权利要求 10所述的客户评价方法， 其特征在于， 该方法还包 括： 11. The customer evaluation method according to claim 10, wherein the method further comprises:
所述客户评价***根据所述第一评价指标和第二评价指标的历史数据 和当前数据计算用于评价用户的第三评价指标的值。  The customer evaluation system calculates a value for evaluating a third evaluation index of the user based on historical data and current data of the first evaluation index and the second evaluation index.
PCT/CN2010/072214 2009-09-09 2010-04-26 Voice-based customer evaluation system and method thereof WO2011029306A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200910092843.8 2009-09-09
CN 200910092843 CN101662549B (en) 2009-09-09 2009-09-09 Customer evaluation system and customer evaluation method based on voice

Publications (1)

Publication Number Publication Date
WO2011029306A1 true WO2011029306A1 (en) 2011-03-17

Family

ID=41790320

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2010/072214 WO2011029306A1 (en) 2009-09-09 2010-04-26 Voice-based customer evaluation system and method thereof

Country Status (2)

Country Link
CN (1) CN101662549B (en)
WO (1) WO2011029306A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10708423B2 (en) 2014-12-09 2020-07-07 Alibaba Group Holding Limited Method and apparatus for processing voice information to determine emotion based on volume and pacing of the voice

Families Citing this family (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101662549B (en) * 2009-09-09 2013-02-27 中兴通讯股份有限公司 Customer evaluation system and customer evaluation method based on voice
CN103443810B (en) * 2011-03-17 2016-05-04 富士通株式会社 Operator evaluates assisting system and operator evaluates support method
CN102802114B (en) * 2012-06-20 2015-02-18 北京语言大学 Method and system for screening seat by using voices
CN103516915A (en) * 2012-06-27 2014-01-15 百度在线网络技术（北京）有限公司 Method, system and device for replacing sensitive words in call process of mobile terminal
CN105723360A (en) * 2013-09-25 2016-06-29 英特尔公司 Improving natural language interactions using emotional modulation
CN104702758B (en) * 2013-12-05 2019-05-10 中兴通讯股份有限公司 A kind of terminal and its method for managing multimedia notepad
CN103811009A (en) * 2014-03-13 2014-05-21 华东理工大学 Smart phone customer service system based on speech analysis
CN104616170A (en) * 2015-01-29 2015-05-13 北京搜狗科技发展有限公司 Interactive method and electronic equipment
CN106408106A (en) * 2015-08-03 2017-02-15 阿里巴巴集团控股有限公司 Method used for realizing high-risk right-protection early warning and apparatus thereof
CN105654250A (en) * 2016-02-01 2016-06-08 百度在线网络技术（北京）有限公司 Method and device for automatically assessing satisfaction degree
CN105933144A (en) * 2016-04-13 2016-09-07 江西掌中无限网络科技股份有限公司 Automatic intelligent quality testing method and system for callback record
CN107886951B (en) * 2016-09-29 2021-07-23 百度在线网络技术（北京）有限公司 Voice detection method, device and equipment
CN108154409A (en) * 2016-12-05 2018-06-12 天脉聚源（北京）科技有限公司 A kind of method and system of internet shopping evaluation
CN109088920B (en) * 2018-07-20 2021-09-21 北京小米移动软件有限公司 Evaluation method, device and equipment of intelligent sound box and storage medium
CN109618065A (en) * 2018-12-28 2019-04-12 合肥凯捷技术有限公司 A kind of voice quality inspection rating system
CN109819128A (en) * 2019-01-23 2019-05-28 平安科技（深圳）有限公司 A kind of quality detecting method and device of telephonograph
CN109754648A (en) * 2019-03-06 2019-05-14 北京大学第一医院 A kind of teaching method and device for analyzing clinical interview quality
CN109819122A (en) * 2019-03-12 2019-05-28 安徽赛迪信息技术有限公司 A kind of customer information analysis system
CN111311327A (en) * 2020-02-19 2020-06-19 平安科技（深圳）有限公司 Service evaluation method, device, equipment and storage medium based on artificial intelligence
CN113055537A (en) * 2021-04-13 2021-06-29 上海东普信息科技有限公司 Voice quality inspection method, device, equipment and storage medium for customer service personnel

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0910202A2 (en) * 1997-10-16 1999-04-21 Robert Bosch Gmbh Method for establishing an intercom connection and device
JP2002300259A (en) * 2001-04-03 2002-10-11 Fujitsu I-Network Systems Ltd Method and system for evaluation test of voice speech equipment
CN101056332A (en) * 2006-11-24 2007-10-17 广州市宜通世纪科技有限公司 An user experienced evaluation device and evaluation method
CN101201980A (en) * 2007-12-19 2008-06-18 北京交通大学 Remote Chinese language teaching system based on voice affection identification
CN101662549A (en) * 2009-09-09 2010-03-03 中兴通讯股份有限公司 Customer evaluation system and customer evaluation method based on voice

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100566360C (en) * 2006-01-19 2009-12-02 北京讯鸟软件有限公司 Realize the call center services method of sitting service level evaluation
CN101150435B (en) * 2007-10-24 2011-12-07 中兴通讯股份有限公司 A system and method for customer comment feedback by using multimedia
CN101415008A (en) * 2008-12-04 2009-04-22 华通科技有限公司 Method for implementing call centre based on XMPP protocol

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0910202A2 (en) * 1997-10-16 1999-04-21 Robert Bosch Gmbh Method for establishing an intercom connection and device
JP2002300259A (en) * 2001-04-03 2002-10-11 Fujitsu I-Network Systems Ltd Method and system for evaluation test of voice speech equipment
CN101056332A (en) * 2006-11-24 2007-10-17 广州市宜通世纪科技有限公司 An user experienced evaluation device and evaluation method
CN101201980A (en) * 2007-12-19 2008-06-18 北京交通大学 Remote Chinese language teaching system based on voice affection identification
CN101662549A (en) * 2009-09-09 2010-03-03 中兴通讯股份有限公司 Customer evaluation system and customer evaluation method based on voice

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10708423B2 (en) 2014-12-09 2020-07-07 Alibaba Group Holding Limited Method and apparatus for processing voice information to determine emotion based on volume and pacing of the voice

Also Published As

Publication number Publication date
CN101662549A (en) 2010-03-03
CN101662549B (en) 2013-02-27

Similar Documents

Publication Publication Date Title
WO2011029306A1 (en) Voice-based customer evaluation system and method thereof
US20180082690A1 (en) Methods and system for reducing false positive voice print matching
US7076427B2 (en) Methods and apparatus for audio data monitoring and evaluation using speech recognition
US10229676B2 (en) Phrase spotting systems and methods
US20090097634A1 (en) Method and System for Call Processing
WO2015184829A1 (en) Method and device for assisting customer service centre attendant
US11843719B1 (en) Analysis of customer interaction metrics from digital voice data in a data-communication server system
CN112885348A (en) AI-combined intelligent voice electric marketing method
CN104301554A (en) Device and method used for detecting service quality of customer service staff
CA2665055C (en) Treatment processing of a plurality of streaming voice signals for determination of responsive action thereto
CN111128241A (en) Intelligent quality inspection method and system for voice call
CN110705926A (en) Method, device and system for acquiring logistics object distribution information
CN111147669A (en) Full real-time automatic service quality inspection system and method
US8676588B2 (en) System for handling a plurality of streaming voice signals for determination of responsive action thereto
CN111835925A (en) Off-line voice quality inspection and analysis system for call center
CN116340482A (en) Multi-skill customer service auxiliary product based on enterprise WeChat combined with NLP engine
CN109891861A (en) For handling the method for user's input and with the motor vehicle of data processing equipment
US10924611B2 (en) Voice recognition system and call evaluation setting method
CN113641801B (en) Control method and system of voice scheduling system and electronic equipment
US8751222B2 (en) Recognition processing of a plurality of streaming voice signals for determination of a responsive action thereto
JP7164372B2 (en) Speech recognition system and speech recognition method
CN115731937A (en) Information processing method, information processing device, electronic equipment and storage medium
CN113743100A (en) Information processing method and device based on seat auxiliary robot
CN112463946A (en) Intelligent equity query comparison method and system based on artificial intelligence technology
CN112911071A (en) Auxiliary robot and intelligent auxiliary method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 10814908

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 10814908

Country of ref document: EP

Kind code of ref document: A1