WO2019019778A1 - 通话数据处理方法、装置、存储介质和计算机设备 - Google Patents

通话数据处理方法、装置、存储介质和计算机设备 Download PDF

Info

Publication number
WO2019019778A1
WO2019019778A1 PCT/CN2018/087327 CN2018087327W WO2019019778A1 WO 2019019778 A1 WO2019019778 A1 WO 2019019778A1 CN 2018087327 W CN2018087327 W CN 2018087327W WO 2019019778 A1 WO2019019778 A1 WO 2019019778A1
Authority
WO
WIPO (PCT)
Prior art keywords
word
feature word
feature
occurrences
agent
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.)
Ceased
Application number
PCT/CN2018/087327
Other languages
English (en)
French (fr)
Inventor
陆诗
项同德
杨辛未
张旭
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen Co Ltd
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 Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Publication of WO2019019778A1 publication Critical patent/WO2019019778A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/26—Speech to text systems
    • 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
    • H04M3/523—Centralised call answering arrangements requiring operator intervention, e.g. call or contact centers for telemarketing with call distribution or queueing
    • H04M3/5232—Call distribution algorithms

Definitions

  • the present application relates to the field of computer technologies, and in particular, to a call data processing method, apparatus, storage medium, and computer device.
  • the inventor realizes that in the traditional human interaction, the user needs to actively inform the telephone customer service staff of their own needs after talking with the telephone customer service personnel, and the customer service personnel passively handle the related business according to the user's voice content. Making the phone interaction less efficient.
  • a call data processing method, apparatus, storage medium, and computer apparatus are provided.
  • a call data processing method includes:
  • a corresponding dedicated call channel is assigned to the selected feature words.
  • a call data processing device includes:
  • a first acquiring module configured to acquire user voice data in a historical voice session according to a time period
  • a word segmentation module configured to obtain a feature word set for the session text segmentation obtained by identifying the user voice data
  • a second obtaining module configured to obtain a proportion of occurrences of each feature word in the feature word set to a sum of occurrence times of each feature word
  • a determining module configured to determine a difference between a proportion of each of the feature words acquired in a current time period and a percentage obtained in a historical time period before the current time period;
  • a selection module configured to select a characteristic word whose corresponding difference is greater than a preset difference and is inconsistent with the preset word
  • an allocating module configured to allocate a corresponding dedicated call channel for the selected feature words.
  • One or more non-volatile storage media storing computer readable instructions, when executed by one or more processors, cause one or more processors to perform the following steps:
  • a corresponding dedicated call channel is assigned to the selected feature words.
  • a computer device comprising a memory and one or more processors having stored therein computer readable instructions, the computer readable instructions being executable by the processor to cause the one or more processors to execute The following steps:
  • a corresponding dedicated call channel is assigned to the selected feature words.
  • FIG. 1 is an application environment diagram of a call data processing method in accordance with one or more embodiments.
  • FIG. 2 is a block diagram of a computer device in accordance with one or more embodiments.
  • FIG. 3 is a flow diagram of a method of processing call data in accordance with one or more embodiments.
  • FIG. 4 is a flow chart showing a method of processing call data according to another embodiment.
  • FIG. 5 is a block diagram of a call data processing apparatus in accordance with one or more embodiments.
  • FIG. 6 is a block diagram of a call data processing apparatus in accordance with another embodiment.
  • FIG. 1 is an application environment diagram of a call data processing method in an embodiment.
  • the call data processing method is applied to a call data processing system.
  • the call data processing system includes a user terminal 110, a server 120, and an agent terminal 130, and both the user terminal 110 and the agent terminal 130 can communicate with the server 120 through a network.
  • the user terminal 110 may be a mobile terminal used by the user or a fixed call terminal.
  • the server 120 may be an independent physical server or a server cluster composed of a plurality of physical servers.
  • the agent terminal 130 is a fixed call terminal, and each agent terminal corresponds to a fixed artificial seat. Multiple agent terminals correspond to a dedicated call channel.
  • the server 120 extracts the feature words representing the user's needs from the user voice data in the historical voice session by the user, and obtains the proportion of the occurrence times of the feature words respectively to the sum of the occurrence times of the feature words, and is the current
  • the ratio of the time period acquisition, the difference between the proportion obtained in the historical time period before the current time period is greater than the preset difference, and the feature words that are inconsistent with the preset words are allocated a dedicated call channel, so that the user enjoys Consultation, business handling and other services provided by the agents.
  • the computer device can be the server 120 of FIG.
  • the computer device includes a processor memory and a network interface connected by a system bus.
  • the memory includes a nonvolatile storage medium and an internal memory.
  • the non-volatile storage medium of the computer device can store an operating system and computer readable instructions that, when executed, cause the processor to perform a method of processing call data.
  • the processor of the computer device is used to provide computing and control capabilities to support the operation of the entire computer device.
  • the computer readable instructions are used to implement a call data processing method provided by the following embodiments when the computer readable instructions are executed by the processor.
  • the computer device can also connect to the user terminal and/or the agent terminal through the network to receive voice data sent by the user terminal and/or the agent terminal.
  • FIG. 2 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the terminal to which the solution of the present application is applied.
  • the specific terminal may include a ratio. More or fewer components are shown in the figures, or some components are combined, or have different component arrangements.
  • a call data processing method is provided. This embodiment is mainly illustrated by the method being applied to the server 120 in FIG. 1 described above. Referring to FIG. 3, the call data processing method specifically includes the following steps:
  • a voice conversation is a conversation between a person and a user.
  • User voice data is voice data of a user generated during a voice session.
  • a historical voice conversation is a voice conversation that has taken place.
  • the time period refers to the period in which data is acquired, such as one or more days, weeks, or months.
  • the server determines whether the current time reaches a periodic time point that meets the time period. Periodic time points such as daily preset time or monthly preset date.
  • the server acquires user voice data in the historical voice session when it is determined that the current time reaches the periodic time point. User voice data is stored in the server's database, file or cache.
  • the user terminal can establish a voice session with the server through the telephone network by dialing the service number, collect voice data generated by the user in the voice session, and send the collected user voice data to the server through the telephone network.
  • the server can store the received user voice data in a database, cache or file for reading when needed.
  • the user terminal may also establish an internet connection with the server by initiating a network request for the voice session, thereby establishing an internet-based voice session, collecting voice data generated by the user in the voice session, and collecting the user voice data. Sent to the server via the internet.
  • the server can store the received user voice data in a database, cache or file for reading when needed.
  • Word segmentation refers to the division of a sequence of consecutive characters into separate characters or sequences of characters.
  • a feature word refers to a character or sequence of characters having a semantic expression function.
  • the server may perform feature extraction on the user voice data, obtain the user voice feature data to be identified, and then perform voice segmentation processing on the user voice feature data to be identified based on the acoustic model to obtain a plurality of phonemes, according to the candidate words in the candidate font library.
  • the correspondence between the phonemes converts the processed plurality of phonemes into a sequence of characters, and then uses the language model to adjust the transformed character sequence, thereby obtaining a conversation text conforming to the natural language mode.
  • the server may perform word segmentation on the session text by using a preset word segmentation method to obtain a plurality of characters or character sequences, and select a character or a character sequence having actual semantics as a feature word from the obtained character sequence to form a feature word set.
  • the feature set may include one or more feature words.
  • the default word segmentation method can be based on character matching, semantic based understanding or statistical based word segmentation.
  • a stop word is a functional character or sequence of characters included in a natural language. Such functional characters or sequences of characters have no actual semantics, including a tone or sequence of characters representing a tone and a sequence of characters or characters representing a logical relationship. Wait. Specifically, a modal character such as "?" or "?", etc., a connection character such as "of” or “at”, etc., a sequence of tempo characters such as "only” or “is”, etc., a sequence of connected characters such as “as for” or “ Then” wait.
  • the number of occurrences of the feature word is the number of times the feature word appears in the conversation text recognized by the user's voice data.
  • the server may count the number of occurrences of each feature word in the feature word set in the conversation text recognized by the user voice data, and obtain the appearance of each feature word. The number of times, and then the ratio of the number of occurrences of each feature word to the sum of the number of occurrences of each feature word is calculated.
  • the feature word set includes three characteristics of “health insurance”, “one card” and “medical insurance”.
  • the feature word "health insurance” is obtained 185 times in the session text obtained by the user voice data in the current time period, and the feature word "one card” is obtained in the session text recognized by the user voice data in the current time period.
  • the number of times is 260, and the feature word "medical insurance” gets the number of occurrences in the conversation text recognized by the user's voice data in the current time period as 97 times.
  • the appearance of the feature word "health insurance” accounts for the appearance of each feature word.
  • the server may obtain, for each feature word in the feature word set obtained in the current time period, a proportion of the number of occurrences of the number of occurrences of each feature word in the current time period, and before the current time period.
  • the proportion of occurrences to the sum of the occurrences of each feature word and then calculate the difference between the two ratios.
  • the historical time period before the current time period may be the previous time period adjacent to the current time period, or may be any time period before the current time period.
  • S310 Select a feature word whose corresponding difference is greater than a preset difference and is inconsistent with the preset word.
  • the preset difference is a difference between a pre-set feature word acquired in the current time period and a percentage obtained in a historical time period before the current time period.
  • the preset difference value is configured to select a condition for assigning a feature word of the dedicated call channel.
  • the preset word is a preset feature word, and the server has assigned a corresponding dedicated call channel for the preset word.
  • the server may determine, for each feature word in the feature word set obtained in the current time period, a ratio of the feature word acquired in the current time period, and a ratio obtained in the historical time period before the current time period. After the difference, the determined differences are compared with the preset difference one by one, and the corresponding determined feature words whose difference value is greater than the preset difference are selected. The server may further compare the filtered feature words with the preset words, and select feature words that are inconsistent with the preset words.
  • a dedicated call channel is a call channel dedicated to a particular service.
  • a call channel dedicated to health insurance or a call channel dedicated to health insurance.
  • the feature word selected by the server represents the focus of the user's recent attention, and the server may allocate a corresponding channel for the feature word after selecting the feature word, and the corresponding exclusive service has been provided.
  • the above call data processing method extracts, according to a time period, a feature word representing a user's demand from user voice data in a historical voice session, and obtains a proportion of occurrences of each feature word to a sum of occurrences of the feature words, and The difference between the proportion of each feature word acquired in the current time period and the proportion of the historical time period before the current time period is determined.
  • the corresponding difference of the feature words is greater than the preset difference and is inconsistent with the preset words, it is determined that the traffic volume related to the feature words is significantly increased in the near future, and a special call channel is automatically allocated for the feature words, so that the dedicated call channel can be allocated.
  • the call channel actively provides the user with information about the feature word related services, thereby improving the efficiency of the telephone interaction.
  • step S304 includes: obtaining a separate word for the session text segmentation obtained by identifying the user voice data; classifying the individual words into individual word subsets according to the semantics of the individual word belonging; and selecting the highest number of occurrences in each individual word subset A single word is used as a feature word to obtain a feature word set.
  • Step S306 includes: obtaining, for each feature word in the feature word set, a sum of occurrence times of each individual word in the individual word subset selected from the current feature word as the number of occurrences of the current feature word; determining the occurrence times of each feature word The proportion of the sum of the occurrences of feature words.
  • a single word is a sequence of characters or characters that represent a single semantic.
  • the server may perform word segmentation processing based on character matching-based word segmentation, and separate a single character from the conversation text in a front-to-back or back-to-front order, and then the single character and the standard thesaurus. Make a match. If the match is successful, the character is obtained as a single word; if the match fails, the match is continued by adding a character until all the characters included in the session text match.
  • the server can also perform forward matching segmentation and reverse matching segmentation for the conversation text at the same time.
  • a plurality of individual characters or character sequences obtained by the word segmentation are taken as separate words.
  • the word segmentation results of the two word segmentation methods are different, the number of individual characters or character sequences obtained by the two word segmentation methods is separately calculated, and the individual characters or character sequences obtained by the segmentation method with a small number of calculations are selected as the individual words.
  • the server may determine the semantics of the obtained individual words, and divide the individual words with the same semantics into the same class to obtain a plurality of individual word subsets. For each individual word subset obtained, the server can select the individual words with the highest number of occurrences as feature words to represent the corresponding individual word subsets, thereby obtaining the feature word set. For each feature word in the feature word set, since the individual words in each individual word subset belong to the same semantics, the server can select the sum of the occurrence times of each individual word in the individual word subsets of each feature word as the feature words. The number of occurrences; recalculate the proportion of occurrences of each feature word as a percentage of the number of occurrences of each feature word.
  • the individual words belonging to the same semantics are divided into the same class, and the individual words with the highest number of occurrences are selected as representatives, and then the individual word subsets selected as the representative feature words are presented, and the individual words appear.
  • the sum of the times is used as the number of occurrences of the feature words, so that the statistics on the number of occurrences of the feature words are more reasonable, so that the selected feature words can better reflect the focus of the user's recent attention.
  • the call data processing method further includes: collecting historical service data corresponding to each agent identifier; determining a correlation between the historical service data of each agent identifier and the selected feature word; The corresponding agent identifiers are sorted; the agent identifier of the preset proportion is selected from the first position of the sorted agent identifier; and the agent terminal corresponding to the selected agent identifier is associated with the dedicated call channel.
  • the agent identification is used to uniquely identify an agent.
  • the agent identification may be a character string including at least one of a number, a letter, and a symbol.
  • Historical service data refers to the online call records that agents provide online business consulting and business processing for users during historical time.
  • the historical service data includes data such as the number of online calls connected, the corresponding call history of each online call, and user feedback.
  • the server may obtain historical service data corresponding to each agent identifier, calculate a correlation between historical service data of each agent identifier and the selected feature words, and then sort the corresponding agent identifiers according to the correlation degree. When sorting, sorting in descending order of relevance, the correlation is high and the correlation is low.
  • the server may further select the agent identifier from the sorted agent identifiers, starting from the most relevant agent identifier, according to a preset percentage of the total number of the sorted agent identifiers. For example, if there are a total of X seat identification sorts, the top X*10% agent logo is taken, and the preset ratio may be 10%.
  • the server may further associate the agent terminal corresponding to the selected agent identifier with a dedicated call channel corresponding to the selected feature word.
  • the dedicated call channel associated with the selected feature word is associated with the agent terminal, and the familiar agent provides services for the user. Thereby improving the efficiency of telephone interaction.
  • the call data processing method further includes: transmitting a voice guiding information to the user terminal after establishing a communication connection with the user terminal; receiving a dedicated call channel selected by the user terminal according to the guiding information in real time; establishing the user terminal and the selected dedicated The communication connection between the agent terminals corresponding to the call channel.
  • Voice guidance information is information that guides a user to select a business service. For example, for health insurance consultation, please perform XX operation, please perform XX operation for general insurance account inquiry.
  • the server may invoke the IVR system to send voice guidance information to the user terminal, the voice guidance information including voice prompt information for selection of a service type.
  • the user terminal can play the voice prompt information, and send the acquired operation information of the user to the server.
  • the server determines the corresponding feature word according to the correspondence between the preset operation information and the feature word, and establishes a seat terminal corresponding to the dedicated call channel corresponding to the determined feature word by the user terminal. Communication between the connections.
  • the server obtains a dedicated call channel selected by the user through the voice guidance information, and establishes a communication connection between the user terminal and the corresponding agent terminal of the dedicated call channel in real time, thereby real-time connecting to the dedicated call channel.
  • Professional agents interact with the user to improve the efficiency of telephone interaction.
  • the call data processing method further includes: determining a service type to which the selected feature word belongs; acquiring service knowledge information from a service database corresponding to the service type; and generating a response corresponding to the service type according to the service knowledge information
  • the template is sent to the agent terminal corresponding to the selected agent identifier.
  • the types of services include credit card business types, banking business types, securities business types, and insurance business types.
  • the insurance business type may include sub-business types such as life insurance business type, health insurance business type, and investment-linked insurance business type.
  • the feature words of the plurality of service types are stored in the agent business database. And the corresponding relationship between each service type and one or more feature words is set, and the feature words corresponding to the service type are feature words subordinate to the service type.
  • the server may determine the service type to which the feature word belongs according to the corresponding relationship of the corresponding service type.
  • the server further sets a calling interface corresponding to the service database for different service types, and allocates the same or different number of interface calling resources for different service types.
  • the server may establish a connection with the service database corresponding to the service type according to the interface corresponding to the service type.
  • business knowledge information of the corresponding service type of the service database is stored in the service database.
  • Business knowledge information includes business-related knowledge and knowledge information about questions and answers.
  • Knowledge information includes information such as a professional terminology, calculation formula, and business process.
  • the service knowledge information and the response template related to the feature words are automatically provided to the agent terminal, so that the agent can improve the problem when the related business knowledge information needs to be retrieved to solve the problem of the user consultation. The efficiency of the answer.
  • the call data processing method includes the following steps:
  • S404 obtaining a separate word for the session text segmentation obtained by identifying the user voice data; classifying the individual words into individual word subsets according to the semantics of the individual words; selecting the individual words with the highest number of occurrences in each individual word subset as the feature words to obtain the feature word set .
  • S410 Select a feature word whose corresponding difference is greater than a preset difference and is inconsistent with the preset word; and allocate a corresponding dedicated call channel for the selected feature word.
  • S412 Collect historical service data corresponding to each agent identifier; determine correlation degree between historical service data of each agent identifier and selected feature words; sort the corresponding agent identifiers according to relevance degree descending; The agent identifier of the preset proportion is selected; the agent terminal corresponding to the selected agent identifier is associated with the dedicated call channel.
  • S414 Determine a service type to which the selected feature word belongs, obtain service knowledge information from a service database corresponding to the service type, generate a response template corresponding to the service type according to the service knowledge information, and send the response template to the selected agent identifier.
  • Agent terminal Determine a service type to which the selected feature word belongs, obtain service knowledge information from a service database corresponding to the service type, generate a response template corresponding to the service type according to the service knowledge information, and send the response template to the selected agent identifier.
  • Agent terminal Determine a service type to which the selected feature word belongs, obtain service knowledge information from a service database corresponding to the service type, generate a response template corresponding to the service type according to the service knowledge information, and send the response template to the selected agent identifier.
  • the feature words representing the user's needs are extracted from the user voice data of the user in the historical voice session according to a time period, and the proportion of occurrences of the feature words to the sum of the number of occurrences of the feature words is obtained, and The difference between the proportion of each feature word acquired in the current time period and the proportion of the historical time period before the current time period is determined.
  • the corresponding difference of the feature words is greater than the preset difference and is inconsistent with the preset words, it is determined that the traffic volume related to the feature words is significantly increased in the near future, and a special call channel is automatically allocated for the feature words, so that the dedicated call channel can be allocated.
  • the call channel actively provides the user with information about the feature word related services, thereby improving the efficiency of the telephone interaction.
  • the dedicated call channel assigned to the feature word is associated with the agent terminal corresponding to the familiar agent, and provides the queryed business knowledge information and response template related to the feature word, so that the agent needs to retrieve relevant business knowledge.
  • the efficiency of obtaining the answers to the questions is improved, and the efficiency of the telephone interaction is further improved.
  • a call data processing apparatus 500 includes: a first acquiring module 501, a word segmentation module 502, a second obtaining module 503, and a determining module 504. The module 505 and the allocation module 506 are selected.
  • the first obtaining module 501 is configured to acquire user voice data in a historical voice session according to a time period.
  • the word segmentation module 502 is configured to obtain a feature word set for the session text segmentation obtained by identifying the user voice data.
  • the second obtaining module 503 is configured to obtain a proportion of the number of occurrences of each feature word in the feature word set to the sum of the number of occurrences of each feature word.
  • the determining module 504 is configured to determine a difference between the proportion of each feature word acquired in the current time period and the proportion obtained in the historical time period before the current time period.
  • the selecting module 505 is configured to select a feature word whose corresponding determined difference is greater than the preset difference and is inconsistent with the preset word.
  • the distribution module 506 is configured to allocate a corresponding dedicated call channel for the selected feature words.
  • the call data processing device 500 extracts feature words representing the user's needs from the user voice data in the historical voice session by the user, and obtains the proportion of the number of occurrences of the feature words in the sum of the number of occurrences of the feature words. And determining the difference between the proportion of each feature word acquired in the current time period and the proportion obtained in the historical time period before the current time period.
  • the corresponding difference of the feature words is greater than the preset difference and is inconsistent with the preset words, it is determined that the traffic volume related to the feature words is significantly increased in the near future, and a special call channel is automatically allocated for the feature words, so that the dedicated call channel can be allocated.
  • the call channel actively provides the user with information about the feature word related services, thereby improving the efficiency of the telephone interaction.
  • the word segmentation module 502 is further configured to obtain a separate word for the session text segmentation obtained by identifying the user voice data; classify the individual words into individual word subsets according to the semantics of the individual word attribution; and select the number of occurrences in each individual word subset. The highest individual word is used as a feature word to obtain a feature word set.
  • the second obtaining module 503 is further configured to obtain, for each feature word in the feature word set, a sum of occurrence times of each individual word in the individual word subset selected from the current feature word, as the number of occurrences of the current feature word; and determine each feature word The percentage of occurrences as a percentage of the number of occurrences of each feature word.
  • the call data processing apparatus 500 further includes an association module 507.
  • the association module 507 is configured to collect historical service data corresponding to each agent identifier, determine a correlation degree between the historical service data of each agent identifier and the selected feature word, and sort the corresponding agent identifiers according to the descending degree of relevance;
  • the agent identifier of the selected seat is selected from the first place; the agent terminal corresponding to the selected agent identifier is associated with the dedicated call channel.
  • the association module 507 is further configured to send voice guidance information to the user terminal after establishing a communication connection with the user terminal; receive a dedicated call channel selected by the user terminal according to the guidance information in real time; and establish a user terminal and the selected dedicated call channel. Corresponding communication connection between the agent terminals.
  • the association module 507 is further configured to: determine a service type to which the selected feature word belongs; obtain service knowledge information from a service database corresponding to the service type; and generate a response template corresponding to the service type according to the service knowledge information; Send the response template to the agent terminal corresponding to the selected agent ID.
  • the various modules in the data processing apparatus described above may be implemented in whole or in part by software, hardware, and combinations thereof. Each of the above modules may be embedded in or independent of the processor in the computer device, or may be stored in a memory in the computer device in a software form, so that the processor invokes the operations corresponding to the above modules.
  • One or more non-volatile storage media storing computer readable instructions, when executed by one or more processors, cause one or more processors to perform the steps of: acquiring a historical voice session over a time period User voice data; obtaining a feature word set for the session text segmentation obtained by identifying the user voice data; obtaining the proportion of the occurrence times of each feature word in the feature word set as a sum of the occurrence times of each feature word; determining each feature word at the current The difference between the ratio of the time period acquisition and the proportion of the historical time period before the current time period; selecting the characteristic word whose corresponding difference is greater than the preset difference and is inconsistent with the preset word; The selected feature words are assigned corresponding dedicated call channels.
  • the session text segmentation obtained by identifying the user voice data obtains the feature word set, including: obtaining a separate word for the session text segmentation obtained by identifying the user voice data; and classifying the individual words according to the semantics of the individual word to obtain the individual words. Set; select the individual words with the highest number of occurrences in each individual word subset as the feature words to obtain the feature word set.
  • Obtaining the proportion of occurrences of each feature word in the feature word set to the sum of the occurrence times of each feature word including: for each feature word in the feature word set, obtaining the occurrence of each individual word in the individual word subset selected from the current feature word The sum of the times is taken as the number of occurrences of the current feature words; the ratio of the number of occurrences of each feature word to the sum of the number of occurrences of each feature word is determined.
  • the processor when the computer executable instructions are executed by the processor, the processor further causes the processor to perform the steps of: collecting historical service data corresponding to each agent identifier; determining historical service data of each agent identifier and the selected feature word Correlation degree; sort the corresponding agent identifiers according to the descending degree of relevance; select the seat identifier of the preset proportion from the first position of the sorted agent identification; associate the agent terminal corresponding to the selected agent identifier to the dedicated call channel.
  • the processor when the computer executable instructions are executed by the processor, the processor is further caused to perform the steps of: transmitting a voice guidance information to the user terminal after establishing a communication connection with the user terminal; receiving a dedicated call selected by the user terminal according to the guidance information in real time. Channel; establishing a communication connection between the user terminal and the agent terminal corresponding to the selected dedicated call channel.
  • the processor when the computer executable instructions are executed by the processor, the processor further causes the processor to: determine a service type to which the selected feature word belongs; obtain business knowledge information from a service database corresponding to the service type; The knowledge information generates a response template corresponding to the service type, and sends the response template to the agent terminal corresponding to the selected agent identifier.
  • a computer device comprising a memory and one or more processors having stored therein computer readable instructions, the computer readable instructions being executed by the processor such that the one or more processors perform the steps of: acquiring the history by time period User voice data in a voice conversation; a feature word set obtained by identifying a conversation text segmentation obtained by the user voice data; obtaining a proportion of occurrence times of each feature word in the feature word set as a sum of occurrence times of each feature word; determining each feature word The difference between the ratio obtained in the current time period and the ratio obtained in the historical time period before the current time period; selecting the characteristic word whose corresponding difference is greater than the preset difference and is inconsistent with the preset word ; Assign a corresponding dedicated call channel to the selected feature words.
  • the session text segmentation obtained by identifying the user voice data obtains the feature word set, including: obtaining a separate word for the session text segmentation obtained by identifying the user voice data; and classifying the individual words according to the semantics of the individual word to obtain the individual words. Set; select the individual words with the highest number of occurrences in each individual word subset as the feature words to obtain the feature word set.
  • Obtaining the proportion of occurrences of each feature word in the feature word set to the sum of the occurrence times of each feature word including: for each feature word in the feature word set, obtaining the occurrence of each individual word in the individual word subset selected from the current feature word The sum of the times is taken as the number of occurrences of the current feature words; the ratio of the number of occurrences of each feature word to the sum of the number of occurrences of each feature word is determined.
  • the processor when the computer executable instructions are executed by the processor, the processor further causes the processor to perform the steps of: collecting historical service data corresponding to each agent identifier; determining historical service data of each agent identifier and the selected feature word Correlation degree; sort the corresponding agent identifiers according to the descending degree of relevance; select the seat identifier of the preset proportion from the first position of the sorted agent identification; associate the agent terminal corresponding to the selected agent identifier to the dedicated call channel.
  • the processor when the computer executable instructions are executed by the processor, the processor is further caused to perform the steps of: transmitting a voice guidance information to the user terminal after establishing a communication connection with the user terminal; receiving a dedicated call selected by the user terminal according to the guidance information in real time. Channel; establishing a communication connection between the user terminal and the agent terminal corresponding to the selected dedicated call channel.
  • the processor when the computer executable instructions are executed by the processor, the processor further causes the processor to: determine a service type to which the selected feature word belongs; obtain business knowledge information from a service database corresponding to the service type; The knowledge information generates a response template corresponding to the service type, and sends the response template to the agent terminal corresponding to the selected agent identifier.
  • Non-volatile memory can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory.
  • Volatile memory can include random access memory (RAM) or external cache memory.
  • RAM is available in a variety of formats, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronization chain.
  • SRAM static RAM
  • DRAM dynamic RAM
  • SDRAM synchronous DRAM
  • DDRSDRAM double data rate SDRAM
  • ESDRAM enhanced SDRAM
  • Synchlink DRAM SLDRAM
  • Memory Bus Radbus
  • RDRAM Direct RAM
  • DRAM Direct Memory Bus Dynamic RAM
  • RDRAM Memory Bus Dynamic RAM

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Signal Processing (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Telephonic Communication Services (AREA)

Abstract

一种通话数据处理方法，包括：按时间周期获取历史语音会话中的用户语音数据；对识别所述用户语音数据得到的会话文本分词得到特征词集；获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；确定各所述特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；选取相应确定的差值大于预设差值、且与预设词不一致的特征词；为选取的所述特征词分配相应的专用通话通道。

Description

通话数据处理方法、装置、存储介质和计算机设备
相关申请的交叉引用
本申请要求于2017年07月25日提交中国专利局，申请号为2017106129912，发明名称为“通话数据处理方法、装置、存储介质和计算机设备”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及计算机技术领域，特别是涉及一种通话数据处理方法、装置、存储介质和计算机设备。
背景技术
随着计算机技术的发展，使用电话与电话客服人员进行交互，通过远程通话来实现各种类型的业务办理和咨询越来越常用。
然而，发明人意识到，在传统的人工交互中，用户需要在与电话客服人员进行通话后，主动告知电话客服人员自己的需求，客服人员再被动根据用户的语音内容为用户办理相关的业务，使得电话交互的效率较低。
发明内容
根据本申请公开的各种实施例，提供一种通话数据处理方法、装置、存储介质和计算机设备。
一种通话数据处理方法，包括：
按时间周期获取历史语音会话中的用户语音数据；
对识别所述用户语音数据得到的会话文本分词得到特征词集；
获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；
确定各所述特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；
选取相应确定的差值大于预设差值、且与预设词不一致的特征词；及
为选取的所述特征词分配相应的专用通话通道。
一种通话数据处理装置，包括：
第一获取模块，用于按时间周期获取历史语音会话中的用户语音数据；
分词模块，用于对识别所述用户语音数据得到的会话文本分词得到特征词集；
第二获取模块，用于获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；
确定模块，用于确定各所述特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；
选取模块，用于选取相应确定的差值大于预设差值、且与预设词不一致的特征词；及
分配模块，用于为选取的所述特征词分配相应的专用通话通道。
一个或多个存储有计算机可读指令的非易失性存储介质，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行以下步骤：
按时间周期获取历史语音会话中的用户语音数据；
对识别所述用户语音数据得到的会话文本分词得到特征词集；
获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；
确定各所述特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；
选取相应确定的差值大于预设差值、且与预设词不一致的特征词；及
为选取的所述特征词分配相应的专用通话通道。
一种计算机设备，包括存储器和一个或多个处理器，所述存储器中储存有计算机可读指令，所述计算机可读指令被所述处理器执行时，使得所述一个或多个处理器执行以下步骤：
按时间周期获取历史语音会话中的用户语音数据；
对识别所述用户语音数据得到的会话文本分词得到特征词集；
获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；
确定各所述特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；
选取相应确定的差值大于预设差值、且与预设词不一致的特征词；及
为选取的所述特征词分配相应的专用通话通道。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本申请的其它特征、 目的和优点将从说明书、附图以及权利要求书变得明显。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它的附图。
图1为根据一个或多个实施例中通话数据处理方法的应用环境图。
图2为根据一个或多个实施例中计算机设备的框图。
图3为根据一个或多个实施例中通话数据处理方法的流程示意图。
图4为根据另一个实施例中通话数据处理方法的流程示意图。
图5为根据一个或多个实施例中通话数据处理装置的框图。
图6为根据另一个实施例中通话数据处理装置的框图。
具体实施方式
为了使本申请的技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
图1为一个实施例中通话数据处理方法的应用环境图。参照图1，该通话数据处理方法应用于通话数据处理系统。通话数据处理系统包括用户终端110、服务器120和坐席终端130，用户终端110和坐席终端130均可通过网络与服务器120通信。用户终端110可以是用户使用的移动终端，也可以是固定通话终端。服务器120可以是独立的物理服务器，也可以是多个物理服务器构成的服务器集群。坐席终端130为固定通话终端，每个坐席终端对应着一个固定的人工坐席。多个坐席终端对应一条专用通话通道。服务器120按时间周期，从用户在历史语音会话中的用户语音数据提取代表用户需求的特征词，获取这些特征词各自的出现次数占这些特征词的出现次数之和的占比，并为在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值大于预设差值、且与预设词不一致的特征词分配专用通话通道，以使用户享受坐席人员提供的咨询、业务办理等服务。
图2为一个实施例中计算机设备的框图。该计算机设备可以是图1中服务器120。如 图2所示，该计算机设备包括通过系统总线连接的处理器存储器和网络接口。存储器包括非易失性存储介质和内存储器。该计算机设备的非易失性存储介质可存储操作系统和计算机可读指令，该计算机可读指令被执行时，可使得处理器执行一种通话数据处理方法。该计算机设备的处理器用于提供计算和控制能力，支撑整个计算机设备的运行。该内存储器中可储存有计算机可读指令，该计算机可读指令被处理器执行时，该计算机可读指令用于实现以下各实施例所提供的一种通话数据处理方法。该计算机设备还可通过网络与用户终端和/或坐席终端连接，接收用户终端和/或坐席终端发送的语音数据。本领域技术人员可以理解，图2中示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的终端的限定，具体的终端可以包括比图中所示更多或更少的部件，或者组合某些部件，或者具有不同的部件布置。
如图3所示，在其中一个实施例中，提供了一种通话数据处理方法。本实施例主要以该方法应用于上述图1中的服务器120来举例说明。参照图3，该通话数据处理方法具体包括如下步骤：
S302，按时间周期获取历史语音会话中的用户语音数据。
语音会话是坐席人员与用户之间进行语音对答的会话。用户语音数据是在语音会话过程中产生的用户的语音数据。历史语音会话是已经发生的语音会话。时间周期是指进行数据获取的周期，如一个或多个日、周或者月。
具体地，服务器在制定时间周期后，判断当下时间是否达到符合时间周期的周期性时间点。周期性时间点如每日预设时刻或者每月预设日。服务器在判定当下时间达到周期性时间点时，获取历史语音会话中的用户语音数据。用户语音数据存储在服务器的数据库、文件或者缓存等。
在其中一个实施例中，用户终端可通过拨打服务号码的方式，与服务器通过电话网络建立语音会话，采集用户在语音会话中产生的语音数据，将采集的用户语音数据通过电话网络发送至服务器。服务器可将接收到的用户语音数据存储在数据库、缓存或者文件中，以在需要的时候读取。
在其中一个实施例中，用户终端也可以通过发起语音会话的网络请求，与服务器建立互联网连接，从而建立基于互联网的语音会话，采集用户在语音会话中产生的语音数据，将采集的用户语音数据通过互联网发送至服务器。服务器可将接收到的用户语音数据存储在数据库、缓存或者文件中，以在需要的时候读取。
S304，对识别用户语音数据得到的会话文本分词得到特征词集。
分词是指将一个连续的字符序列切分成多个单独的字符或者字符序列。特征词是指具有语义表达功能的字符或者字符序列。
具体地，服务器可对用户语音数据进行特征提取，获得待识别的用户语音特征数据，然后基于声学模型对待识别的用户语音特征数据进行语音分帧处理得到多个音素，根据候选字库中候选字与音素的对应关系，将处理得到的多个音素转化为字符序列，再利用语言模型调整转化得到的字符序列，从而得到符合自然语言模式的会话文本。
进一步地，服务器可采用预设的分词方式对会话文本进行分词处理，得到多个字符或者字符序列，从得到的字符序列中筛选出具有实际语义的字符或者字符序列作为特征词，形成特征词集。特征词集可以包括一个或多个特征词。预设的分词方式可以是基于字符匹配、基于语义理解或者基于统计的分词方式。
更进一步地，服务器从得到的字符或字符序列中筛选出具有实际语义的字符或者字符序列作为特征词时，具体可从得到的字符或字符序列中过滤掉停用词。停用词是指自然语言中包括的一种功能字符或者字符序列，这类功能字符或者字符序列并无实际语义，包括代表语气的语气字符或字符序列和表示某种逻辑关系连接字符或字符序列等。具体地，语气字符比如“吗”或者“呢”等，连接字符比如“的”或“在”等，语气字符序列比如“而已”或者“就是了”等，连接字符序列比如“至于”或“然后”等。
S306，获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比。
特征词的出现次数是特征词在由用户语音数据识别得到的会话文本中出现的次数。具体地，服务器在对识别用户语音数据得到的会话文本分词得到特征词集后，可统计特征词集中各特征词在由用户语音数据识别得到的会话文本中出现的次数，得到各特征词的出现次数，再计算得到各特征词的出现次数占各特征词的出现次数之和的占比。
举例说明，特征词集中包括“健康险”、“一卡通”和“医疗保险”三个特征词。特征词“健康险”在当前时间周期获取由用户语音数据识别得到的会话文本中出现的次数为185次，特征词“一卡通”在当前时间周期获取由用户语音数据识别得到的会话文本中出现的次数为260次，特征词“医疗保险”在当前时间周期获取由用户语音数据识别得到的会话文本中出现的次数为97次.那么，特征词“健康险”的出现次数占各特征词的出现次数之和的占比为：185/(185+160+97)＝41.86％。
S308，确定各特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时 间周期获取的占比之间的差值。
具体地，服务器可对于在当前时间周期得到的特征词集中的每个特征词，获取在当前时间周期时，出现次数占各特征词的出现次数之和的占比，以及在当前时间周期之前的历史时间周期时，出现次数占各特征词的出现次数之和的占比，再计算两个占比之间的差值。在当前时间周期之前的历史时间周期可以是与当前时间周期相邻的上一时间周期，也可以是时间在当前时间周期之前的任一时间周期。
S310，选取相应确定的差值大于预设差值、且与预设词不一致的特征词。
预设差值是预先设置的特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值。预设差值被配置为选取分配专用通话通道的特征词的条件。当特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值大于预设差值，则判定该特征词满足被分配专用通话通道的条件。预设词是预先设置的特征词，且服务器已为预设词分配相应的专用通话通道。
具体地，服务器可在对于在当前时间周期得到的特征词集中的每个特征词确定该特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值后，将确定的各差值逐一与预设差值进行比较，筛选出相应确定的差值大于预设差值的特征词。服务器可再将筛选出的特征词与预设词比较，选取与预设词不一致的特征词。
S312，为选取的特征词分配相应的专用通话通道。
专用通话通道是专用于某一特定业务的通话通道。比如，专用于健康险的通话通道，或者专用于医疗保险的通话通道。具体地，服务器选取的特征词代表用户近期关注的重点，服务器可在选取特征词后为该特征词分配相应的通道，已提供相应的专属服务。
上述通话数据处理方法，按时间周期从用户在历史语音会话中的用户语音数据提取代表用户需求的特征词，获取这些特征词各自的出现次数占这些特征词的出现次数之和的占比，并确定各特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值。当特征词相应的差值大于预设差值、且与预设词不一致时，判定近期该特征词相关的业务量明显增多，自动为该特征词分配专用通话通道，这样即可通过分配的专用通话通道主动为用户提供特征词相关业务的信息，从而提高电话交互的效率。
在其中一个实施例中，步骤S304包括：对识别用户语音数据得到的会话文本分词得到单独词；按单独词所属语义对单独词分类得到单独词子集；挑选各单独词子集中出现次 数最高的单独词作为特征词得到特征词集。步骤S306包括：对于特征词集中每个特征词，获取当前特征词挑选自的单独词子集中各单独词的出现次数之和，作为当前特征词的出现次数；确定各特征词的出现次数占各特征词的出现次数之和的占比。
单独词是表示单一语义的字符或者字符序列。
在其中一个实施例中，服务器可采用基于字符匹配的分词方式进行分词处理，将会话文本按照从前到后或者从后到前的顺序逐一切分出单个字符，再将该单个字符与标准词库进行匹配。如果匹配成功，则获取该字符作为一个单独词；若匹配失败，则通过增加一个字符继续进行匹配，直至会话文本中包括的字符全部匹配完成。
在其中一个实施例中，服务器也可同时对会话文本进行正向匹配分词和逆向匹配分词。在两种分词方式的分词结果相同时，将分词得到的多个单独的字符或者字符序列作为单独词。在两种分词方式的分词结果不相同时，分别计算两种分词方式得到的单独的字符或者字符序列的数量，选取计算的数量少的分词方式得到的单独的字符或者字符序列为作为单独词。
进一步地，服务器可确定得到的各单独词所属的语义，将所属语义相同的单独词划分为同一类，得到多个单独词子集。对于得到的每一个单独词子集，服务器可再从中挑选出出现次数最高的单独词作为特征词，来代表相应的单独词子集，从而得到特征词集。对于特征词集中每个特征词，由于每个单独词子集中的单独词属于相同的语义，服务器可将各特征词挑选自的单独词子集中各单独词的出现次数之和，作为各特征词的出现次数；再计算各特征词的出现次数占各特征词的出现次数之和的占比。
上述实施例中，将属于相同语义的单独词划分为同一类，并从中选出出现次数最高的单独词作为代表，然后将作为代表的特征词挑选自的单独词子集中，各单独词的出现次数之和作为该特征词的出现次数，使得对特征词出现次数的统计更合理，从而使得挑选出的特征词更能反映用户近期关注的重点。
在其中一个实施例中，该通话数据处理方法还包括：收集每个坐席标识对应的历史服务数据；确定各坐席标识的历史服务数据与选取的特征词之间的相关度；按照相关度降序对相应的坐席标识进行排序；从排序的坐席标识首位起选取预设比例的坐席标识；将选取的坐席标识对应的坐席终端关联至专用通话通道。
坐席标识用于唯一标识一个坐席人员。坐席标识可以是包括数字、字母和符号中的至少一种字符的字符串。历史服务数据是指坐席人员在历史时间内为用户提供在线业务咨 询和业务办理的在线通话记录。历史服务数据包括所接通的在线通话数量、每通在线通话相应的通话记录和用户反馈等数据。
具体地，服务器可获取每个坐席标识对应的历史服务数据，计算各坐席标识的历史服务数据与选取的特征词之间的相关度，从而根据该相关度来对相应的坐席标识进行排序。排序时，按照相关度降序排序，相关度高的靠前而相关度低的靠后。服务器可再从排序的坐席标识中，从相关度最高的坐席标识开始，按照排序的坐席标识总量的预设百分比来选取坐席标识。比如共有X个坐席标识排序，则取排名靠前的X*10％的坐席标识，此时预设比例可为10％。服务器可再将选取的坐席标识对应的坐席终端，与选取的特征词相应的专用通话通道关联。
在本实施例中，以坐席标识的历史服务数据与选取的特征词之间的相关度为依据，为选取的特征词分配的专用通话通道关联坐席终端，通过熟悉的坐席人员为用户提供服务，从而提高电话交互的效率。
在其中一个实施例中，该通话数据处理方法还包括：与用户终端建立通信连接后向用户终端发送语音引导信息；实时接收用户终端根据引导信息选择的专用通话通道；建立用户终端与选择的专用通话通道对应的坐席终端之间的通信连接。
语音引导信息是引导用户选择业务服务的信息。比如，健康险咨询请执行XX操作，万能险账户查询请执行XX操作等。
具体地，服务器可调用IVR系统向用户终端发送语音引导信息，该语音引导信息包括对业务类型的选择的语音提示信息。用户终端可播放该语音提示信息，将所获取的用户的操作信息发送至服务器。服务器在接收到用户终端发送的操作信息后，根据预设的操作信息和特征词之间的对应关系确定对应的特征词，并建立用户终端与确定的特征词相应的专用通话通道对应的坐席终端之间的通信连接。
在本实施例中，服务器通过语音引导信息获取用户选择的专用通话通道，实时建立用户终端与该专用通话通道的对应的坐席终端之间的通信连接，从而实时连接到为该专用通话通道分配的专业的坐席人员与用户进行电话交互，提高电话交互效率。
在其中一个实施例中，该通话数据处理方法还包括：确定选取的特征词所属的业务类型；从与业务类型对应的业务数据库中获取业务知识信息；根据业务知识信息生成与业务类型对应的应答模板；将应答模板发送至选取的坐席标识对应的坐席终端。
业务类型包括信用卡业务类型、银行业务类型、证券业务类型以及保险业务类型等 其中的多种。而保险业务类型又可包括寿险业务类型、健康险业务类型、投连险业务类型等子业务类型。
具体地，坐席业务数据库中存储了多个业务类型的特征词。并设置了每种业务类型与一个或多个特征词的对应关系，与业务类型具有对应关系的特征词，即为从属于该业务类型的特征词。服务器在确定了用户终端选取的特征词后，可根据该特征词相应业务类型的对应关系，确定其所属的业务类型。
进一步地，服务器还针对不同业务类型设置了与业务数据库对应的调用接口，并为不同的业务类型分配了对应相同或不同数量的接口调用资源。在确定了业务类型后，服务器可根据调用与该业务类型对应的接口，与该业务类型对应的业务数据库建立连接。
更进一步地，业务数据库中存储了该业务数据库相应业务类型的业务知识信息。业务知识信息包括业务相关知识和问题与回答的知识信息。知识信息包括某个专业名词解释、计算公式、业务流程等信息。服务器在与该业务类型对应的业务数据库建立连接后，从该业务数据库中获取业务知识信息。服务器可再根据业务知识信息生成与业务类型对应的应答模板，并将应答模板发送至选取的坐席标识对应的坐席终端。
在本实施例中，通过进一步向坐席终端自动提供查询出的与特征词相关的业务知识信息与应答模板，使坐席人员在需要检索相关的业务知识信息来解答用户咨询的问题时，提高对问题答案的获取效率。
如图4所示，在其中一个具体的实施例中，通话数据处理方法包括以下步骤：
S402，按时间周期获取历史语音会话中的用户语音数据。
S404，对识别用户语音数据得到的会话文本分词得到单独词；按单独词所属语义对单独词分类得到单独词子集；挑选各单独词子集中出现次数最高的单独词作为特征词得到特征词集。
S406，对于特征词集中每个特征词，获取当前特征词挑选自的单独词子集中各单独词的出现次数之和，作为当前特征词的出现次数；确定各特征词的出现次数占各特征词的出现次数之和的占比。
S408，确定各特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值。
S410，选取相应确定的差值大于预设差值、且与预设词不一致的特征词；为选取的特征词分配相应的专用通话通道。
S412，收集每个坐席标识对应的历史服务数据；确定各坐席标识的历史服务数据与选取的特征词之间的相关度；按照相关度降序对相应的坐席标识进行排序；从排序的坐席标识首位起选取预设比例的坐席标识；将选取的坐席标识对应的坐席终端关联至专用通话通道。
S414，确定选取的特征词所属的业务类型；从与业务类型对应的业务数据库中获取业务知识信息；根据业务知识信息生成与业务类型对应的应答模板；将应答模板发送至选取的坐席标识对应的坐席终端。
S416，与用户终端建立通信连接后向用户终端发送语音引导信息；实时接收用户终端根据引导信息选择的专用通话通道；建立用户终端与选择的专用通话通道对应的坐席终端之间的通信连接。
在本实施例中，按时间周期从用户在历史语音会话中的用户语音数据提取代表用户需求的特征词，获取这些特征词各自的出现次数占这些特征词的出现次数之和的占比，并确定各特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值。当特征词相应的差值大于预设差值、且与预设词不一致时，判定近期该特征词相关的业务量明显增多，自动为该特征词分配专用通话通道，这样即可通过分配的专用通话通道主动为用户提供特征词相关业务的信息，从而提高电话交互的效率。
其次，为该特征词分配的专用通话通道关联了熟悉高的坐席人员对应的坐席终端，并提供查询出的与特征词相关的业务知识信息与应答模板，使坐席人员在需要检索相关的业务知识信息来解答用户咨询的问题时，提高对问题答案的获取效率，进一步提高了电话交互的效率。
应该理解的是，上述流程图中的各个步骤按照箭头的指示依次显示，但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明，这些步骤的执行并没有严格的顺序限制，这些步骤可以以其它的顺序执行。而且，上述流程图中的至少一部分步骤可以包括多个子步骤或者多个阶段，这些子步骤或者阶段并不必然是在同一时刻执行完成，而是可以在不同的时刻执行，这些子步骤或者阶段的执行顺序也不必然是依次进行，而是可以与其它步骤或者其它步骤的子步骤或者阶段的至少一部分轮流或者交替地执行。
如图5所示，在其中一个实施例中，提供了一种通话数据处理装置500，该通话数据处理装置500包括：第一获取模块501、分词模块502、第二获取模块503、确定模块504、 选取模块505以及分配模块506。
第一获取模块501，用于按时间周期获取历史语音会话中的用户语音数据。
分词模块502，用于对识别用户语音数据得到的会话文本分词得到特征词集。
第二获取模块503，用于获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比。
确定模块504，用于确定各特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值。
选取模块505，用于选取相应确定的差值大于预设差值、且与预设词不一致的特征词。
分配模块506，用于为选取的特征词分配相应的专用通话通道。
上述通话数据处理装置500，按时间周期从用户在历史语音会话中的用户语音数据提取代表用户需求的特征词，获取这些特征词各自的出现次数占这些特征词的出现次数之和的占比，并确定各特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值。当特征词相应的差值大于预设差值、且与预设词不一致时，判定近期该特征词相关的业务量明显增多，自动为该特征词分配专用通话通道，这样即可通过分配的专用通话通道主动为用户提供特征词相关业务的信息，从而提高电话交互的效率。
在其中一个实施例中，分词模块502还用于对识别用户语音数据得到的会话文本分词得到单独词；按单独词所属语义对单独词分类得到单独词子集；挑选各单独词子集中出现次数最高的单独词作为特征词得到特征词集。第二获取模块503还用于对于特征词集中每个特征词，获取当前特征词挑选自的单独词子集中各单独词的出现次数之和，作为当前特征词的出现次数；确定各特征词的出现次数占各特征词的出现次数之和的占比。
如图6所示，在其中一个实施例中，通话数据处理装置500还包括：关联模块507。
关联模块507，用于收集每个坐席标识对应的历史服务数据；确定各坐席标识的历史服务数据与选取的特征词之间的相关度；按照相关度降序对相应的坐席标识进行排序；从排序的坐席标识首位起选取预设比例的坐席标识；将选取的坐席标识对应的坐席终端关联至专用通话通道。
在其中一个实施例中，关联模块507还用于与用户终端建立通信连接后向用户终端发送语音引导信息；实时接收用户终端根据引导信息选择的专用通话通道；建立用户终端与选择的专用通话通道对应的坐席终端之间的通信连接。
在其中一个实施例中，关联模块507还用于确定选取的特征词所属的业务类型；从与业务类型对应的业务数据库中获取业务知识信息；根据业务知识信息生成与业务类型对应的应答模板；将应答模板发送至选取的坐席标识对应的坐席终端。
关于数据处理装置的具体限定可以参见上文中对于数据处理方法的限定，在此不再赘述。上述数据处理装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中，也可以以软件形式存储于计算机设备中的存储器中，以便于处理器调用执行以上各个模块对应的操作。
一个或多个存储有计算机可读指令的非易失性存储介质，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行以下步骤：按时间周期获取历史语音会话中的用户语音数据；对识别用户语音数据得到的会话文本分词得到特征词集；获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；确定各特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；选取相应确定的差值大于预设差值、且与预设词不一致的特征词；为选取的特征词分配相应的专用通话通道。
在一个实施例中，对识别用户语音数据得到的会话文本分词得到特征词集，包括：对识别用户语音数据得到的会话文本分词得到单独词；按单独词所属语义对单独词分类得到单独词子集；挑选各单独词子集中出现次数最高的单独词作为特征词得到特征词集。获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比，包括：对于特征词集中每个特征词，获取当前特征词挑选自的单独词子集中各单独词的出现次数之和，作为当前特征词的出现次数；确定各特征词的出现次数占各特征词的出现次数之和的占比。
在一个实施例中，计算机可执行指令被处理器执行时，还使得处理器执行以下步骤：收集每个坐席标识对应的历史服务数据；确定各坐席标识的历史服务数据与选取的特征词之间的相关度；按照相关度降序对相应的坐席标识进行排序；从排序的坐席标识首位起选取预设比例的坐席标识；将选取的坐席标识对应的坐席终端关联至专用通话通道。
在一个实施例中，计算机可执行指令被处理器执行时，还使得处理器执行以下步骤：与用户终端建立通信连接后向用户终端发送语音引导信息；实时接收用户终端根据引导信息选择的专用通话通道；建立用户终端与选择的专用通话通道对应的坐席终端之间的通信连接。
在一个实施例中，计算机可执行指令被处理器执行时，还使得处理器执行以下步骤： 确定选取的特征词所属的业务类型；从与业务类型对应的业务数据库中获取业务知识信息；根据业务知识信息生成与业务类型对应的应答模板；将应答模板发送至选取的坐席标识对应的坐席终端。
一种计算机设备，包括存储器和一个或多个处理器，存储器中储存有计算机可读指令，计算机可读指令被处理器执行时，使得一个或多个处理器执行以下步骤：按时间周期获取历史语音会话中的用户语音数据；对识别用户语音数据得到的会话文本分词得到特征词集；获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；确定各特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；选取相应确定的差值大于预设差值、且与预设词不一致的特征词；为选取的特征词分配相应的专用通话通道。
在一个实施例中，对识别用户语音数据得到的会话文本分词得到特征词集，包括：对识别用户语音数据得到的会话文本分词得到单独词；按单独词所属语义对单独词分类得到单独词子集；挑选各单独词子集中出现次数最高的单独词作为特征词得到特征词集。获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比，包括：对于特征词集中每个特征词，获取当前特征词挑选自的单独词子集中各单独词的出现次数之和，作为当前特征词的出现次数；确定各特征词的出现次数占各特征词的出现次数之和的占比。
在一个实施例中，计算机可执行指令被处理器执行时，还使得处理器执行以下步骤：收集每个坐席标识对应的历史服务数据；确定各坐席标识的历史服务数据与选取的特征词之间的相关度；按照相关度降序对相应的坐席标识进行排序；从排序的坐席标识首位起选取预设比例的坐席标识；将选取的坐席标识对应的坐席终端关联至专用通话通道。
在一个实施例中，计算机可执行指令被处理器执行时，还使得处理器执行以下步骤：与用户终端建立通信连接后向用户终端发送语音引导信息；实时接收用户终端根据引导信息选择的专用通话通道；建立用户终端与选择的专用通话通道对应的坐席终端之间的通信连接。
在一个实施例中，计算机可执行指令被处理器执行时，还使得处理器执行以下步骤：确定选取的特征词所属的业务类型；从与业务类型对应的业务数据库中获取业务知识信息；根据业务知识信息生成与业务类型对应的应答模板；将应答模板发送至选取的坐席标识对应的坐席终端。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通 过计算机可读指令来指令相关的硬件来完成，所述的计算机可读指令可存储于一非易失性计算机可读取存储介质中，该计算机可读指令在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用，均可包括非易失性和/或易失性存储器。非易失性存储器可包括只读存储器(ROM)、可编程ROM(PROM)、电可编程ROM(EPROM)、电可擦除可编程ROM(EEPROM)或闪存。易失性存储器可包括随机存取存储器(RAM)或者外部高速缓冲存储器。作为说明而非局限，RAM以多种形式可得，诸如静态RAM(SRAM)、动态RAM(DRAM)、同步DRAM(SDRAM)、双数据率SDRAM(DDRSDRAM)、增强型SDRAM(ESDRAM)、同步链路(Synchlink)DRAM(SLDRAM)、存储器总线(Rambus)直接RAM(RDRAM)、直接存储器总线动态RAM(DRDRAM)、以及存储器总线动态RAM(RDRAM)等。
以上实施例的各技术特征可以进行任意的组合，为使描述简洁，未对上述实施例中的各个技术特征所有可能的组合都进行描述，然而，只要这些技术特征的组合不存在矛盾，都应当认为是本说明书记载的范围。
以上所述实施例仅表达了本申请的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对本申请专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本申请构思的前提下，还可以做出若干变形和改进，这些都属于本申请的保护范围。因此，本申请专利的保护范围应以所附权利要求为准。

Claims (20)

  1. 一种通话数据处理方法，包括：
    按时间周期获取历史语音会话中的用户语音数据；
    对识别所述用户语音数据得到的会话文本分词得到特征词集；
    获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；
    确定各所述特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；
    选取相应确定的差值大于预设差值、且与预设词不一致的特征词；及
    为选取的所述特征词分配相应的专用通话通道。
  2. 根据权利要求1所述的方法，其特征在于，所述对识别所述用户语音数据得到的会话文本分词得到特征词集，包括：
    对识别所述用户语音数据得到的会话文本分词得到单独词；
    按所述单独词所属语义对所述单独词分类得到单独词子集；及
    挑选各单独词子集中出现次数最高的单独词作为特征词得到特征词集；
    所述获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比，包括：
    对于特征词集中每个特征词，获取当前特征词挑选自的单独词子集中各单独词的出现次数之和，作为当前特征词的出现次数；及
    确定各特征词的出现次数占各特征词的出现次数之和的占比。
  3. 根据权利要求1所述的方法，其特征在于，还包括：
    收集每个坐席标识对应的历史服务数据；
    确定各坐席标识的历史服务数据与选取的所述特征词之间的相关度；
    按照相关度降序对相应的坐席标识进行排序；
    从排序的坐席标识首位起选取预设比例的坐席标识；及
    将选取的坐席标识对应的坐席终端关联至所述专用通话通道。
  4. 根据权利要求3所述的方法，其特征在于，还包括：
    与用户终端建立通信连接后向所述用户终端发送语音引导信息；
    实时接收所述用户终端根据所述引导信息选择的专用通话通道；及
    建立所述用户终端与选择的所述专用通话通道对应的坐席终端之间的通信连接。
  5. 根据权利要求3所述的方法，其特征在于，还包括：
    确定选取的所述特征词所属的业务类型；
    从与所述业务类型对应的业务数据库中获取业务知识信息；
    根据所述业务知识信息生成与所述业务类型对应的应答模板；及
    将所述应答模板发送至选取的所述坐席标识对应的坐席终端。
  6. 一种通话数据处理装置，包括：
    第一获取模块，用于按时间周期获取历史语音会话中的用户语音数据；
    分词模块，用于对识别所述用户语音数据得到的会话文本分词得到特征词集；
    第二获取模块，用于获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；
    确定模块，用于确定各所述特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；
    选取模块，用于选取相应确定的差值大于预设差值、且与预设词不一致的特征词；及
    分配模块，用于为选取的所述特征词分配相应的专用通话通道。
  7. 根据权利要求6所述的装置，其特征在于，所述分词模块还用于对识别所述用户语音数据得到的会话文本分词得到单独词；按所述单独词所属语义对所述单独词分类得到单独词子集；挑选各单独词子集中出现次数最高的单独词作为特征词得到特征词集；
    所述第二获取模块还用于对于特征词集中每个特征词，获取当前特征词挑选自的单独词子集中各单独词的出现次数之和，作为当前特征词的出现次数；确定各特征词的出现次数占各特征词的出现次数之和的占比。
  8. 根据权利要求6所述的装置，其特征在于，所述装置还包括：
    关联模块，用于收集每个坐席标识对应的历史服务数据；确定各坐席标识的历史服务数据与选取的所述特征词之间的相关度；按照相关度降序对相应的坐席标识进行排序；从排序的坐席标识首位起选取预设比例的坐席标识；将选取的坐席标识对应的坐席终端关联至所述专用通话通道。
  9. 根据权利要求8所述的装置，其特征在于，所述关联模块还用于与用户终端建立通信连接后向所述用户终端发送语音引导信息；实时接收所述用户终端根据所述引导信息选择的专用通话通道；及建立所述用户终端与选择的所述专用通话通道对应的坐席终端之间的通信连接。
  10. 根据权利要求8所述的装置，其特征在于，所述关联模块还用于确定选取的所述特征词所属的业务类型；从与所述业务类型对应的业务数据库中获取业务知识信息；根据所述业务知识信息生成与所述业务类型对应的应答模板；及将所述应答模板发送至选取的所述坐席标识对应的坐席终端。
  11. 一个或多个存储有计算机可读指令的非易失性计算机可读存储介质，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行以下步骤：
    按时间周期获取历史语音会话中的用户语音数据；
    对识别所述用户语音数据得到的会话文本分词得到特征词集；
    获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；
    确定各所述特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；
    选取相应确定的差值大于预设差值、且与预设词不一致的特征词；及
    为选取的所述特征词分配相应的专用通话通道。
  12. 根据权利要求11所述的存储介质，其特征在于，所述对识别所述用户语音数据得到的会话文本分词得到特征词集，包括：
    对识别所述用户语音数据得到的会话文本分词得到单独词；
    按所述单独词所属语义对所述单独词分类得到单独词子集；及
    挑选各单独词子集中出现次数最高的单独词作为特征词得到特征词集；
    所述获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比，包括：
    对于特征词集中每个特征词，获取当前特征词挑选自的单独词子集中各单独词的出现次数之和，作为当前特征词的出现次数；及
    确定各特征词的出现次数占各特征词的出现次数之和的占比。
  13. 根据权利要求11所述的存储介质，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    收集每个坐席标识对应的历史服务数据；
    确定各坐席标识的历史服务数据与选取的所述特征词之间的相关度；
    按照相关度降序对相应的坐席标识进行排序；
    从排序的坐席标识首位起选取预设比例的坐席标识；及
    将选取的坐席标识对应的坐席终端关联至所述专用通话通道。
  14. 根据权利要求13所述的存储介质，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    与用户终端建立通信连接后向所述用户终端发送语音引导信息；
    实时接收所述用户终端根据所述引导信息选择的专用通话通道；及
    建立所述用户终端与选择的所述专用通话通道对应的坐席终端之间的通信连接。
  15. 根据权利要求13所述的存储介质，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    确定选取的所述特征词所属的业务类型；
    从与所述业务类型对应的业务数据库中获取业务知识信息；
    根据所述业务知识信息生成与所述业务类型对应的应答模板；及
    将所述应答模板发送至选取的所述坐席标识对应的坐席终端。
  16. 一种计算机设备，包括存储器及一个或多个处理器，所述存储器中储存有计算机可读指令，所述计算机可读指令被所述一个或多个处理器执行时，使得所述一个或多个处理器执行以下步骤：
    按时间周期获取历史语音会话中的用户语音数据；
    对识别所述用户语音数据得到的会话文本分词得到特征词集；
    获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比；
    确定各所述特征词在当前时间周期获取的占比，与在当前时间周期之前的历史时间周期获取的占比之间的差值；
    选取相应确定的差值大于预设差值、且与预设词不一致的特征词；及
    为选取的所述特征词分配相应的专用通话通道。
  17. 根据权利要求16所述的计算机设备，其特征在于，所述对识别所述用户语音数据得到的会话文本分词得到特征词集，包括：
    对识别所述用户语音数据得到的会话文本分词得到单独词；
    按所述单独词所属语义对所述单独词分类得到单独词子集；及
    挑选各单独词子集中出现次数最高的单独词作为特征词得到特征词集；
    所述获取特征词集中各特征词的出现次数占各特征词的出现次数之和的占比，包括：
    对于特征词集中每个特征词，获取当前特征词挑选自的单独词子集中各单独词的出现次数之和，作为当前特征词的出现次数；及
    确定各特征词的出现次数占各特征词的出现次数之和的占比。
  18. 根据权利要求16所述的计算机设备，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    收集每个坐席标识对应的历史服务数据；
    确定各坐席标识的历史服务数据与选取的所述特征词之间的相关度；
    按照相关度降序对相应的坐席标识进行排序；
    从排序的坐席标识首位起选取预设比例的坐席标识；及
    将选取的坐席标识对应的坐席终端关联至所述专用通话通道。
  19. 根据权利要求18所述的计算机设备，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    与用户终端建立通信连接后向所述用户终端发送语音引导信息；
    实时接收所述用户终端根据所述引导信息选择的专用通话通道；及
    建立所述用户终端与选择的所述专用通话通道对应的坐席终端之间的通信连接。
  20. 根据权利要求18所述的计算机设备，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    确定选取的所述特征词所属的业务类型；
    从与所述业务类型对应的业务数据库中获取业务知识信息；
    根据所述业务知识信息生成与所述业务类型对应的应答模板；及
    将所述应答模板发送至选取的所述坐席标识对应的坐席终端。
PCT/CN2018/087327 2017-07-25 2018-05-17 通话数据处理方法、装置、存储介质和计算机设备 Ceased WO2019019778A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710612991.2A CN107800900B (zh) 2017-07-25 2017-07-25 通话数据处理方法、装置、存储介质和计算机设备
CN201710612991.2 2017-07-25

Publications (1)

Publication Number Publication Date
WO2019019778A1 true WO2019019778A1 (zh) 2019-01-31

Family

ID=61531124

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/087327 Ceased WO2019019778A1 (zh) 2017-07-25 2018-05-17 通话数据处理方法、装置、存储介质和计算机设备

Country Status (2)

Country Link
CN (1) CN107800900B (zh)
WO (1) WO2019019778A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107800900B (zh) * 2017-07-25 2020-08-28 平安科技（深圳）有限公司 通话数据处理方法、装置、存储介质和计算机设备
CN109543177B (zh) * 2018-10-19 2022-04-12 中国平安人寿保险股份有限公司 消息数据处理方法、装置、计算机设备和存储介质
CN111901554B (zh) * 2020-07-27 2022-11-11 中国平安人寿保险股份有限公司 基于语义聚类的通话通道构建方法、装置和计算机设备
CN113129866B (zh) * 2021-04-13 2022-08-02 重庆度小满优扬科技有限公司 语音处理方法、装置、存储介质及计算机设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101094271A (zh) * 2007-07-20 2007-12-26 华为技术有限公司 自动语音应答系统菜单的生成方法及装置
CN104375847A (zh) * 2013-08-14 2015-02-25 华为技术有限公司 识别业务类型的方法及装置
WO2015047272A1 (en) * 2013-09-26 2015-04-02 Interactive Intelligence, Inc. System and method for context based knowledge retrieval
CN104598632A (zh) * 2015-02-05 2015-05-06 北京航空航天大学 热点事件检测方法和装置
US9036806B1 (en) * 2014-08-27 2015-05-19 Xerox Corporation Predicting the class of future customer calls in a call center
CN107800900A (zh) * 2017-07-25 2018-03-13 平安科技（深圳）有限公司 通话数据处理方法、装置、存储介质和计算机设备

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101094271A (zh) * 2007-07-20 2007-12-26 华为技术有限公司 自动语音应答系统菜单的生成方法及装置
CN104375847A (zh) * 2013-08-14 2015-02-25 华为技术有限公司 识别业务类型的方法及装置
WO2015047272A1 (en) * 2013-09-26 2015-04-02 Interactive Intelligence, Inc. System and method for context based knowledge retrieval
US9036806B1 (en) * 2014-08-27 2015-05-19 Xerox Corporation Predicting the class of future customer calls in a call center
CN104598632A (zh) * 2015-02-05 2015-05-06 北京航空航天大学 热点事件检测方法和装置
CN107800900A (zh) * 2017-07-25 2018-03-13 平安科技（深圳）有限公司 通话数据处理方法、装置、存储介质和计算机设备

Also Published As

Publication number Publication date
CN107800900A (zh) 2018-03-13
CN107800900B (zh) 2020-08-28

Similar Documents

Publication Publication Date Title
CN112434501B (zh) 工单智能生成的方法、装置、电子设备及介质
CN110493019B (zh) 会议纪要的自动生成方法、装置、设备及存储介质
US11503439B2 (en) Message management methods and systems
CN108521525A (zh) 基于用户标签体系的智能机器人客服营销方法和系统
US20130102290A1 (en) Intelligent communication with contact lists
WO2019019778A1 (zh) 通话数据处理方法、装置、存储介质和计算机设备
CN107809550A (zh) 调整业务语音播放顺序的方法及设备
CN109064315A (zh) 逾期账单智能催收方法、装置、计算机设备及存储介质
US10698934B1 (en) Interface between a virtual personal assistant and an interactive voice response system
CN103514165A (zh) 用于识别对话中所提及的人的方法和装置
CN108810296B (zh) 一种智能外呼方法及装置
US20210184876A1 (en) Automatic conference management tool
CN113761127B (zh) 一种任务处理方法、装置、电子设备和存储介质
WO2018188459A9 (zh) 确定推荐信息的方法及装置、信息推荐方法及装置
CN108899015B (zh) 一种向通话中的用户提供语音反馈信息的方法和装置
CN110312046A (zh) 外呼数据优化方法、装置、计算机设备及存储介质
CN112185383A (zh) 一种用于客服回访的处理方法及系统
CN113192505B (zh) 一种会话信息生成方法、装置及服务机器人
CN114222028A (zh) 语音识别方法、装置、计算机设备和存储介质
CN119691117A (zh) 在线咨询方法、系统、设备、存储介质及程序产品
CN113111157B (zh) 问答处理方法、装置、计算机设备和存储介质
CN108632475B (zh) 语音留言方法、装置、计算机设备和存储介质
CN110011905B (zh) 一种通信信息处理方法、装置及设备
CN114500747A (zh) 一种用于应急话务切换的智能话务容灾平台搭建方法
JP2023076003A (ja) プログラム、情報処理システム及び情報処理方法

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: 18839398

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: 18839398

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC , EPO FORM 1205A DATED 04.08.2020.

122 Ep: pct application non-entry in european phase

Ref document number: 18839398

Country of ref document: EP

Kind code of ref document: A1