CN111652001B - Data processing method and device - Google Patents
Data processing method and device Download PDFInfo
- Publication number
- CN111652001B CN111652001B CN202010499962.1A CN202010499962A CN111652001B CN 111652001 B CN111652001 B CN 111652001B CN 202010499962 A CN202010499962 A CN 202010499962A CN 111652001 B CN111652001 B CN 111652001B
- Authority
- CN
- China
- Prior art keywords
- user
- behavior
- target
- historical
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3438—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment monitoring of user actions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/01—Customer relationship services
Abstract
The application discloses a data processing method and a device, wherein the method comprises the following steps: receiving current input content of a current user; obtaining at least one target user behavior corresponding to the current input content in the conversation track set; the dialog track set comprises at least one historical user behavior in a target historical dialog, the historical user behavior is a behavior tag corresponding to the input content of a historical user in the interaction between the historical user and the customer service system, and the characteristic value of at least one log characteristic of a target dialog log corresponding to the target historical dialog meets a preset high satisfaction condition; at least one target user behavior is output. According to the method and the device, the target user behavior corresponding to the current input content of the current user is obtained from the historical user behaviors in the historical conversation meeting the high satisfaction condition and is output to the current user, so that the response content can be fed back according to the user behavior selected by the user in a short term, and the accuracy of the response content is improved.
Description
Technical Field
The application relates to the technical field of intelligent customer service, in particular to a data processing method and device.
Background
In an intelligent customer service system, for each consultation problem of a user, the intelligent customer service usually gives one or more statements to feed back to the user.
However, because the question description of the user may imply some information, the intelligent customer service system may not be able to accurately identify the user question, and the response content fed back to the user is not consistent with the user question, so that the response content fed back by the customer service system is inaccurate.
Disclosure of Invention
In view of the above, the present application provides a data processing method and apparatus, as follows:
a method of data processing, comprising:
receiving current input content of a current user;
obtaining at least one target user behavior corresponding to the current input content of the current user in a conversation track set;
the dialog track set comprises at least one historical user behavior in a target historical dialog, the historical user behavior is a behavior tag corresponding to the input content of a historical user in the interaction between the historical user and a customer service system, and the characteristic value of at least one log feature of a target dialog log corresponding to the target historical dialog meets a preset high satisfaction condition;
outputting at least the at least one target user behavior.
The above method, preferably, obtaining at least one target user behavior corresponding to the current input content of the current user in the dialog track set includes:
obtaining a first action set in a preset conversation track set, wherein the first action set comprises at least one first user action, and the semantic similarity between a conversation track corresponding to the first user action and a conversation track corresponding to the current input content is higher than a preset similarity threshold;
ordering first user behaviors in the first set of behaviors;
and selecting the user behaviors ranked at the top M in the first behavior set as target user behaviors, wherein M is a positive integer greater than or equal to 1.
In the method, preferably, the dialog track set further includes a historical system behavior in the target historical dialog, where the historical system behavior is a behavior tag corresponding to feedback content of the customer service system when the historical user interacts with the customer service system;
wherein ranking the first user actions in the first set of actions comprises:
obtaining a corresponding previous target system behavior of the first user behavior in a target history conversation where the first user behavior is located;
obtaining a first frequency value of the target system behavior in the first behavior set, and obtaining a second frequency value of the target system behavior in the dialog track set;
ordering the first user behavior in the first set of behaviors according to the first frequency value and the second frequency value.
The above method, preferably, the dialog trajectory set is obtained by:
obtaining a history log set for interaction between the customer service system and a plurality of history users, wherein the history log set comprises a plurality of history conversation logs;
obtaining at least one log feature of the historical dialog log, wherein the log feature has a feature value which represents the interactive satisfaction degree of the corresponding historical user to the customer service system on the corresponding dimension of the log feature;
obtaining at least one target dialogue log in the history log set according to the characteristic value of the log characteristic, wherein the characteristic value of the log characteristic of any one target dialogue log meets a preset high satisfaction condition;
and obtaining a historical dialogue track in the target historical dialogue corresponding to the target dialogue log, wherein the historical dialogue track comprises at least one node corresponding to historical user behavior and at least one node corresponding to historical system behavior, and the historical user behavior and the historical system behavior form a dialogue track set.
The above method, preferably, the high satisfaction conditions include: and respectively weighting the characteristic values of the log characteristics of the target dialog logs, and then summing the weighted characteristic values to obtain a value larger than a preset satisfaction threshold.
In the above method, preferably, before obtaining at least one target user behavior corresponding to the current input content of the current user in the dialog track set, the method further includes:
judging whether the current input content meets a preset behavior recommendation condition; the behavior recommendation condition comprises: the current input content represents that the previous feedback content of the customer service system does not accord with the previous input content of the current user;
if the current input content meets the behavior recommendation condition, executing the following steps: obtaining at least one target user behavior corresponding to the current input content of the current user in a conversation track set;
and if the current input content does not meet the behavior recommendation condition, outputting the current feedback content generated aiming at the current input content.
The method preferably outputs at least the at least one target user behavior, and includes:
outputting the at least one target user behavior and current feedback content generated for the current input content.
The above method, preferably, outputting the at least one target user behavior and the current feedback content generated for the current input content, includes:
outputting current feedback content generated aiming at the current input content at a target position on an interactive interface corresponding to the current user;
and outputting the at least one target user behavior in a list form in a target area associated with the target position.
The above method, preferably, further comprises:
receiving operation data of the current user for a second user behavior in the at least one target user behavior;
and outputting feedback content corresponding to the second user behavior.
A data processing apparatus comprising:
the input receiving unit is used for receiving the current input content of the current user;
a behavior obtaining unit, configured to obtain at least one target user behavior corresponding to current input content of the current user in a dialog trajectory set;
the dialog track set comprises at least one historical user behavior in a target historical dialog, the historical user behavior is a behavior tag corresponding to the input content of a historical user in the interaction between the historical user and a customer service system, and the characteristic value of at least one log feature of a target dialog log corresponding to the target historical dialog meets a preset high satisfaction condition;
a behavior output unit for outputting at least the at least one target user behavior.
An electronic device, comprising:
the memory is used for storing the application program and the data generated by the operation of the application program;
input/output means for receiving an input and outputting a content;
a processor for executing an application program in memory to implement: receiving current input content of a current user through an input and output device; obtaining at least one target user behavior corresponding to the current input content of the current user in a conversation track set; the dialog track set comprises at least one historical user behavior in a target historical dialog, the historical user behavior is a behavior tag corresponding to the input content of a historical user in the interaction between the historical user and a customer service system, and the characteristic value of at least one log feature of a target dialog log corresponding to the target historical dialog meets a preset high satisfaction condition; outputting at least the at least one target user behavior through an input-output device.
According to the technical scheme, after the current input content of the current user is received, at least one target user behavior corresponding to the current input content is obtained and output from historical user behaviors in at least one target historical conversation contained in a conversation track set, the historical user behaviors are behavior tags corresponding to the input content of the historical user in the interaction between the historical user and a customer service system, and the historical user behaviors are user behaviors in the target historical conversation corresponding to a conversation log with log feature values meeting a high satisfaction condition. Therefore, in the application, the target user behaviors corresponding to the current input content of the current user are obtained from the historical user behaviors in the historical dialogue meeting the high satisfaction condition and are output to the current user, and based on the target user behaviors, the current user can select the user behaviors meeting requirements in the output target user behaviors, so that the response content can be fed back to the current user according to the selected target user behaviors, and the accuracy of the response content is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a flowchart of a data processing method according to an embodiment of the present application;
FIGS. 2-4 are exemplary diagrams of embodiments of the present application, respectively;
fig. 5 and fig. 6 are partial flowcharts of a data processing method according to an embodiment of the present disclosure;
fig. 7 is another flowchart of a data processing method according to an embodiment of the present application;
FIGS. 8-10 are diagrams of another example of an embodiment of the present application, respectively;
fig. 11 is a schematic structural diagram of a data processing apparatus according to a second embodiment of the present application;
fig. 12 is another schematic structural diagram of a data processing apparatus according to a second embodiment of the present application;
fig. 13 is a schematic structural diagram of an electronic device according to a third embodiment of the present application;
fig. 14 is an exemplary flowchart applied to an intelligent customer service system according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Referring to fig. 1, a flowchart of an implementation of a data processing method provided in an embodiment of the present application is shown, where the method may be applied to an electronic device capable of performing data processing, such as a computer or a server configured with a customer service system. The technical scheme in the embodiment is mainly used for recommending one or more user behaviors which possibly meet the requirements of the user to the user according to the current input content of the user so as to avoid the situation that the accuracy of the response content fed back by the wrong user behavior is poor.
Specifically, the method in this embodiment may include the following steps:
step 101: current input content of a current user is received.
The current user can be understood as a user currently interacting with the customer service system, and the current input content refers to the latest input content in the interaction between the current user and the customer service system. As shown in fig. 2, when the user a interacts with the customer service system, the user a inputs "this is favorable" in the interactive interface, and of course, the interactive interface also contains the historical conversation contents of the user a and the customer service system, such as "how much money" and "single machine price 3500, mouse 99" and the like, and the "bar in the year" input by the user a last time is the current input content of the user a.
Step 102: and obtaining at least one target user behavior corresponding to the current input content of the current user in the conversation track set.
Wherein, the conversation track set comprises: historical user behavior in at least one target historical dialog. The target historical dialogue refers to a plurality of historical dialogs formed by interaction between one or more historical users and a customer service system, and the corresponding target historical dialogue has log characteristics with characteristic values meeting preset high satisfaction conditions.
That is, the target history dialogue is a history dialogue filtered out of all history dialogues, and the filtering principle is as follows: the characteristic value of at least one log characteristic of the target dialogue log corresponding to the target historical dialogue meets a high satisfaction condition.
The log feature herein may include: the customer service system does not understand any one or any combination of more than one of the characteristics of times, the characteristics of conversation continuous rounds, the characteristics of conversation turning times, the characteristics of conversation interruption times, the characteristics of negative emotion degrees, the characteristics of user active manual times, the characteristics of user active evaluation and the like, and the characteristic value of the log characteristic meets the high satisfaction condition: the characteristic value of the log characteristic can represent that the satisfaction degree of the historical user in the target historical dialogue to the customer service system is higher than a threshold value.
It should be noted that the historical user behavior in the dialog track set can be understood as: and behavior labels corresponding to the input contents of the historical users in the interaction between the historical users and the customer service system. For example, an input behavior tag for a user's call, an input behavior tag for a user's consultation price, an input behavior tag for a user's inquiry for an offer, an input behavior tag for a user's expression of emotion, and the like.
Of course, the dialog trajectory set may also include at least one historical system behavior in the target historical dialog, and the historical system behavior may be understood as: and behavior labels corresponding to the feedback content of the customer service system during interaction between the historical user and the customer service system. For example, an incoming behavior tag for a call placed by the system, an incoming behavior tag for a price reply by the system, an incoming behavior tag for an offer replied to by the system, an incoming behavior tag for a thank or sorry by the system, and so forth.
It should be noted that the historical user behaviors and the historical system behaviors in the dialog track set exist in the form of dialog tracks, that is, there is a front-back order between the historical user behaviors and between the historical system behaviors and between the historical user behaviors and the historical system behaviors in the dialog track set, and these behaviors form a dialog or behavior track according to the interaction order. Therefore, the dialog track set may include a plurality of historical dialog tracks, the historical dialog tracks are dialog tracks of the target historical dialog, and the feature values of the log features of the corresponding target dialog logs meet the preset high satisfaction condition. And each historical dialogue track comprises one or more historical user behaviors and one or more historical system behaviors, and the behaviors in each historical dialogue track form a track according to the interaction sequence.
Specifically, in this embodiment, semantic analysis may be performed on the current input content of the current user, and a target user behavior corresponding to the current input content may be found in the dialog track set according to an analysis result of the semantic analysis.
Step 103: and at least outputting the at least one target user behavior.
In this embodiment, the target user behaviors may be output on an interactive interface between the current user and the customer service system, as shown in fig. 3, so that the current user can select one or more target user behaviors. Further, in this embodiment, corresponding feedback content (reply content) may be output for the target user behavior selected by the current user to provide for the current user reference.
It can be seen from the foregoing technical solutions that, in a data processing method provided in an embodiment of the present application, after receiving current input content of a current user, at least one target user behavior corresponding to the current input content is obtained and output from historical user behaviors in at least one target historical dialog included in a dialog track set, where the historical user behaviors are behavior tags corresponding to input contents of historical users in interaction between the historical users and a customer service system, and the historical user behaviors are user behaviors in the target historical dialog corresponding to a dialog log in which a feature value of a log feature satisfies a high satisfaction condition. Therefore, in the embodiment, the target user behavior corresponding to the currently input content of the current user is obtained from the historical user behaviors in the historical dialog satisfying the high satisfaction condition and is output to the current user, based on which, the current user can select the user behavior more satisfying the requirement from the output target user behaviors, so that the response content can be fed back to the current user according to the selected target user behavior, thereby improving the accuracy of the response content.
Specifically, after step 103, in this embodiment, operation data of the current user for a second user behavior in the at least one target user behavior may be received, for example, operation data of a position or an area where the second user behavior is located is clicked, or operation data of sliding on the position or the area where the second user behavior is located is received; then, in this embodiment, the feedback content corresponding to the second user behavior is output on the interactive interface of the current user in the customer service system. For example, the customer service system outputs the feedback contents of the corresponding "stand-alone 4000, protective film 299" for the second user action of "consulting new product price" selected by the current user, as shown in fig. 4.
In one implementation, when at least one target user behavior corresponding to the current input content of the current user is obtained in the dialog track set in step 102, the following may be specifically implemented, as shown in fig. 5:
step 501: and obtaining a first behavior set in a preset dialog track set.
The first action set comprises at least one first user action, and the semantic similarity between the conversation track corresponding to the first user action and the conversation track corresponding to the current input content is higher than a preset similarity threshold.
Specifically, in this embodiment, semantic comparison is first performed on a current dialog track in which current input content of a current user is located and a dialog track in a dialog track set, so as to obtain semantic similarity between the current dialog track and each dialog track in the dialog track set, then, dialog tracks with semantic similarity higher than a similarity threshold are extracted from the dialog track set, and user behaviors in the extracted dialog tracks are obtained, that is, first user behaviors, so as to form a first behavior set.
The similarity threshold may be set according to a requirement or an empirical value, such as 80% or 60%, and the semantic similarity is higher than the similarity threshold for characterization: the similarity degree of the user behaviors in the conversation track is higher than a certain threshold value, which can be understood as follows: the similarity between the user behavior corresponding to the current input content of the current user and the first user behavior is higher.
It should be noted that the first behavior set includes, in addition to the first user behavior, a system behavior corresponding to the first user behavior in the dialog track. That is, the first action set includes all user actions and system actions in one or more dialog tracks in the dialog track set, and the semantic similarity of the dialog track corresponding to the current input content is higher than the preset similarity threshold.
Step 502: the first user actions in the first set of actions are ordered.
In this embodiment, the first user behaviors may be sorted according to the magnitude of the frequency value of each first user behavior in the dialog track set and/or in the first behavior set, for example, the higher the frequency of the first user behavior in the first behavior set or the dialog track set, the earlier the sorting is;
alternatively, in this embodiment, the first user behavior may be sorted according to other criteria, for example, sorted according to a related attribute of a previous target system behavior corresponding to the target history dialog in which the first user behavior is located, for example, sorted according to a frequency value of the target system behavior in the first behavior set, a frequency value of the target system behavior in the dialog track set, or a tf-idf characteristic value of the target system behavior.
Step 503: and selecting the user behaviors ranked at the top M in the first behavior set as target user behaviors, wherein M is a positive integer greater than or equal to 1.
For example, M is 3, at this time, the first user behaviors ranked at the top 3 in the first behavior set are target user behaviors, and further, the target user behaviors are output to the current user to represent that the target user behaviors are behaviors that may better meet the input intention of the current user, such as behaviors of consulting prices or consulting offers, and the current user may click on the target user behaviors, and based on this, the customer service system may feed back corresponding reply content for the selected target user behaviors, so that the reply content more meets the requirements of the current user.
Specifically, when the first user behaviors in the first behavior set are sorted in step 502, the following details are provided:
first, a corresponding previous target system behavior of the first user behavior in the target history dialog where the first user behavior is located is obtained. It should be particularly noted here that the dialog track set includes, in addition to the historical user behavior in at least one target historical dialog, the historical system behavior in the corresponding target historical dialog, and therefore, in this embodiment, in the dialog track set, for each first user behavior, the corresponding previous target system behavior in the target historical dialog where the first user behavior is located is obtained. For example, the first user behavior is an input behavior tag for inquiring the benefit of the user, and correspondingly, a behavior tag corresponding to the feedback content of the previous customer service system corresponding to the input behavior for inquiring the benefit of the user, that is, an input behavior tag for the system to reply the price is obtained in the conversation track set; for another example, the first user behavior is an input behavior tag for expressing the happy emotion of the user, and correspondingly, a behavior tag corresponding to the feedback content of the previous customer service system corresponding to the input behavior for expressing the happy emotion of the user is obtained in the conversation track set, and an input behavior tag for responding to the preferential content by the system is obtained; and so on.
Then, a first frequency value of the target system behavior in the first behavior set is obtained, and a second frequency value of the target system behavior in the dialogue track set is obtained.
In this embodiment, the number of times that a target system behavior appears in the first behavior set and the number of times that a primary target system behavior appears in the dialog track set may be counted, and then, based on the total number of historical system behaviors in the first behavior set and the total number of historical system behaviors in the dialog track set, a first frequency value of the target system behavior in the first behavior set and a second frequency value of the target system behavior in the dialog track set are obtained;
alternatively, in this embodiment, after counting the number of times that the target system behavior occurs in the first behavior set and the number of times that the target system behavior occurs in the dialog track set, based on the total number of all behaviors in the first behavior set (including the historical user behavior and the historical system behavior) and the total number of all behaviors in the dialog track set, the first frequency value of the target system behavior in the first behavior set and the second frequency value of the target system behavior in the dialog track set may be obtained.
Finally, the first user behaviors in the first behavior set are ranked according to the first frequency value and the second frequency value.
Specifically, in this embodiment, the first user behaviors in the first behavior set may be sorted separately according to the size of the first frequency value, and at this time, the larger the first frequency value is, the earlier the corresponding first user behavior is sorted;
or, in this embodiment, the first user behaviors in the first behavior set may be sorted separately according to the size of the second frequency value, and at this time, the larger the second frequency value is, the earlier the corresponding first user behavior is sorted;
or, in this embodiment, the first frequency value and the second frequency value may be combined to obtain a tf-idf characteristic value of a previous target system behavior corresponding to the first user behavior in the target history dialog where the first user behavior is located, for example, a value obtained by dividing the first frequency value by the second frequency value is sorted according to the characteristic value, and at this time, the larger the characteristic value is, the earlier the sorting of the corresponding first user behavior is.
In one implementation, the dialog track set in this embodiment may be obtained in advance by the following method, as shown in fig. 6:
step 601: a history log set of interactions between the customer service system and a plurality of history users is obtained.
The history log set comprises a plurality of history dialogue logs, and each history dialogue log corresponds to each history dialogue formed by interaction between each history user and the customer service system. It should be noted that the historical dialog log is generated when the corresponding historical dialog is formed, and is recorded in a log set corresponding to the customer service system, where the log set may be understood as a set formed by the historical dialog log in the storage area of the electronic device where the customer service system is located.
Step 602: at least one log feature of a historical conversation log is obtained.
The log features have characteristic values, and the characteristic values represent the interactive satisfaction degree of corresponding historical users on the customer service system in the corresponding dimension of the log features.
For example, the historical dialog log includes one or more of the following log features: the customer service system does not understand the characteristics of times, the characteristics of conversation continuous turns, the characteristics of conversation turning times, the characteristics of conversation interruption times, the characteristics of negative emotion degree, the characteristics of user active manual times and the characteristics of user active evaluation, and the size of the characteristic value of the log characteristics represents the interactive satisfaction degree of the historical user on the customer service system. For example, the larger the feature value of the times of the customer service system failure understanding, the lower the degree of interaction satisfaction of the representative historical user to the customer service system; the larger the characteristic value of the conversation continuous round number is, the higher the interactive satisfaction degree of the representation historical user to the customer service system is; the larger the characteristic value of the conversation turn-back times is, the higher the interactive satisfaction degree of the representation history user to the customer service system is; the larger the characteristic value of the number of the conversation interruption times is, the lower the interactive satisfaction degree of the representation history user to the customer service system is; the larger the characteristic value of the negative emotion degree is, the lower the interactive satisfaction degree of the representation historical user to the customer service system is; the larger the characteristic value of the number of times of the user actively turning to manual work is, the lower the interactive satisfaction degree of the representation historical user to the customer service system is; the larger the characteristic value actively evaluated by the user is, the higher the interactive satisfaction degree of the characterization history user to the customer service system is, and the like.
Specifically, in the present embodiment, one or more of the above log features may be extracted from the history dialog log, and a feature value of each log feature may be obtained.
Step 603: and obtaining at least one target dialog log in the historical log set according to the characteristic value of the log characteristic.
In this embodiment, the feature value of the log feature of any one of the target dialog logs obtained in this embodiment satisfies the preset high satisfaction condition. That is to say, in the embodiment, the dialog logs whose feature values of the log features satisfy the preset high satisfaction condition are screened out as the target dialog logs, and other dialog logs which do not satisfy the high satisfaction condition are eliminated.
Specifically, the characteristic value of the log characteristic of the target dialog log meets a preset high satisfaction condition, which may be: and performing weighted summation on the characteristic values of the log characteristics according to respective corresponding preset weights to obtain a value which is greater than a satisfaction threshold, such as greater than 0.5 or greater than 0.7. The preset weight corresponding to each log feature may be set according to a requirement, and the preset weights corresponding to different log features may be the same or different, for example, the feature weight of the number of times that the customer service system cannot understand is 0.2, the feature weight of the number of continuous turns of the dialog is 0.3, the feature weight of the number of times that the dialog is turned back is 0.3, and the like.
Step 604: and obtaining a historical dialogue track in the target historical dialogue corresponding to the target dialogue log.
After the target dialog logs are obtained, behavior extraction may be performed on the target history dialog corresponding to the target dialog logs, for example, behavior tags are extracted from the target history dialog by using user behaviors and system behaviors as trace nodes, so as to obtain a history dialog trace corresponding to each target history dialog, where the history dialog trace includes at least one node corresponding to the history user behaviors and at least one node corresponding to the history system behaviors, and the history user behaviors and the history system behaviors form a dialog trace set.
In one implementation, after step 101 and before step 102, the method in this embodiment may further include the following steps, as shown in fig. 7:
step 104: and judging whether the current input content meets a preset behavior recommendation condition, if so, executing step 102, otherwise, executing step 105.
The behavior recommendation condition may be: the current input content represents that the previous feedback content of the customer service system does not accord with the previous input content of the current user, namely the current input content of the current user represents that the satisfaction degree of the current user on the previous feedback content of the customer service system is lower than a certain threshold value.
Specifically, in this embodiment, semantic analysis and other processing may be performed on the current input content, and then whether the current input content meets the behavior recommendation condition is determined according to a semantic analysis result.
For example, in this embodiment, the currently input content is identified as a content expressed by an angry emotion or identified as a content subjected to manual processing, at this time, the satisfaction degree of the current user for the last content fed back by the customer service system is low, and at this time, the behavior recommendation condition is met, then step 102 is executed, so as to output one or more target user behaviors to the current user, where the target user behaviors belong to behaviors that may meet the intention of the current user behavior, and therefore, the current user may select the target user behaviors, so that the customer service system feeds back corresponding reply content according to the target user behavior selected by the current user, and further the reply content more meets the requirement of the current user, so as to improve the accuracy of the reply content.
And if the current input content is the content expressed by the happy emotion, representing that the satisfaction degree of the current user on the previous feedback content of the customer service system is higher, and the behavior recommendation condition is not met, executing step 105.
Step 105: and outputting the current feedback content generated aiming at the current input content.
That is to say, in this embodiment, under the condition that it is found that the previous feedback content of the customer service system matches the previous input content of the current user, that is, the current input content of the current user represents that the previous feedback content of the current user on the customer service system is high in satisfaction, the customer service system may not provide the target user behavior for the current user any longer, but may directly perform feedback on the current input content, as shown in fig. 8, and output the current feedback content on the current input content.
In one implementation, the current feedback content generated for the current input content may be output at the same time when the at least one target user behavior is output in step 103, as shown in fig. 9. Based on this, if the current user thinks that the current feedback content can meet the own requirement, then the information in the current feedback content can be adopted to perform the next operation, and if the current user thinks that the current feedback content cannot meet the own requirement, then a target user behavior meeting the own requirement can be selected from the output target user behaviors.
Specifically, when the at least one target user behavior and the current feedback content generated for the current input content are output in step 103, the current feedback content generated for the current input content may be output at a target position on the interactive interface corresponding to the current user, and at the same time, the at least one target user behavior is output in a form of a list in a target area associated with the target position.
The target area refers to a specific area on the interactive interface corresponding to the current user, for example, the target area may be an area above the target position, and accordingly, the target user behavior is output on the target area in a list form, as shown in fig. 10.
Fig. 11 is a schematic structural diagram of a data processing apparatus according to a second embodiment of the present application, where the apparatus may be configured in an electronic device capable of performing data processing, such as a computer or a server configured with a customer service system. The technical scheme in the embodiment is mainly used for recommending one or more user behaviors which possibly meet the requirements of the user to the user according to the current input content of the user so as to avoid the situation that the accuracy of the response content fed back by the wrong user behavior is poor.
Specifically, the apparatus in this embodiment may include the following units:
an input receiving unit 1101 for receiving current input content of a current user;
a behavior obtaining unit 1102, configured to obtain at least one target user behavior corresponding to current input content of the current user in a dialog track set;
the dialog track set comprises at least one historical user behavior in a target historical dialog, the historical user behavior is a behavior tag corresponding to the input content of a historical user in the interaction between the historical user and a customer service system, and the characteristic value of at least one log feature of a target dialog log corresponding to the target historical dialog meets a preset high satisfaction condition;
a behavior output unit 1103 configured to output at least the at least one target user behavior.
As can be seen from the foregoing technical solutions, in the data processing apparatus provided in the second embodiment of the present application, after receiving the current input content of the current user, at least one target user behavior corresponding to the current input content is obtained and output from the historical user behaviors in at least one target historical dialog included in the dialog track set, where the historical user behaviors are behavior tags corresponding to the input content of the historical user during interaction between the historical user and the customer service system, and the historical user behaviors are user behaviors in the target historical dialog corresponding to the dialog log in which the feature value of the log feature satisfies the high satisfaction condition. Therefore, in the embodiment, the target user behavior corresponding to the currently input content of the current user is obtained from the historical user behaviors in the historical dialog satisfying the high satisfaction condition and is output to the current user, based on which, the current user can select the user behavior more satisfying the requirement from the output target user behaviors, so that the response content can be fed back to the current user according to the selected target user behavior, thereby improving the accuracy of the response content.
In an implementation manner, the behavior obtaining unit 1102 is specifically configured to:
obtaining a first action set in a preset conversation track set, wherein the first action set comprises at least one first user action, and the semantic similarity between a conversation track corresponding to the first user action and a conversation track corresponding to the current input content is higher than a preset similarity threshold; ordering first user behaviors in the first set of behaviors; and selecting the top M user behaviors in the first behavior set as target user behaviors, wherein M is a positive integer greater than or equal to 1.
Optionally, the dialog track set further includes a historical system behavior in the target historical dialog, where the historical system behavior is a behavior tag corresponding to feedback content of the customer service system in the interaction between the historical user and the customer service system;
the behavior obtaining unit 1102 ranks the first user behaviors in the first behavior set, specifically: obtaining a corresponding previous target system behavior of the first user behavior in a target history conversation where the first user behavior is located; obtaining a first frequency value of the target system behavior in the first behavior set, and obtaining a second frequency value of the target system behavior in the dialog track set; ordering the first user behavior in the first set of behaviors according to the first frequency value and the second frequency value.
In one implementation manner, the behavior obtaining unit 1102 in this embodiment is further configured to obtain a dialog track set, where the dialog track set is obtained by:
obtaining a history log set for interaction between the customer service system and a plurality of history users, wherein the history log set comprises a plurality of history conversation logs; obtaining at least one log feature of the historical dialog log, wherein the log feature has a feature value which represents the interactive satisfaction degree of the corresponding historical user to the customer service system on the corresponding dimension of the log feature; obtaining at least one target dialogue log in the history log set according to the characteristic value of the log characteristic, wherein the characteristic value of the log characteristic of any one target dialogue log meets a preset high satisfaction condition; and obtaining a historical dialogue track in the target historical dialogue corresponding to the target dialogue log, wherein the historical dialogue track comprises at least one node corresponding to historical user behavior and at least one node corresponding to historical system behavior, and the historical user behavior and the historical system behavior form a dialogue track set.
Wherein the high satisfaction condition comprises: and respectively weighting the characteristic values of the log characteristics of the target dialog logs, and then summing the weighted characteristic values to obtain a value larger than a preset satisfaction threshold.
In one implementation, the apparatus in this embodiment may further include the following structure, as shown in fig. 12:
a recommendation determining unit 1104, configured to determine whether the current input content meets a preset behavior recommendation condition before the behavior obtaining unit 1102 obtains at least one target user behavior corresponding to the current input content of the current user in the dialog track set; the behavior recommendation condition comprises: the current input content represents that the previous feedback content of the customer service system does not accord with the previous input content of the current user; if the current input content meets the behavior recommendation condition, executing the behavior obtaining unit 1102; and if the current input content does not meet the behavior recommendation condition, triggering the behavior output unit 1103 to output current feedback content generated for the current input content.
In one implementation, the behavior output unit 1103 is specifically configured to: outputting the at least one target user behavior and current feedback content generated for the current input content. For example, outputting the current feedback content generated aiming at the current input content at the target position on the interactive interface corresponding to the current user; and outputting the at least one target user behavior in a list form in a target area associated with the target position.
In one implementation, the input interface unit 1101 is further configured to: receiving operation data of the current user for a second user behavior in the at least one target user behavior;
based on this, the behavior output unit 1103 is further configured to: and outputting feedback content corresponding to the second user behavior.
Referring to fig. 13, a schematic structural diagram of an electronic device according to a third embodiment of the present disclosure is provided, where the electronic device may be an electronic device capable of performing data processing, such as a computer or a server configured with a customer service system. The technical scheme in the embodiment is mainly used for recommending one or more user behaviors which possibly meet the requirements of the user to the user according to the current input content of the user so as to avoid the situation that the accuracy of the response content fed back by the wrong user behavior is poor.
Specifically, the electronic device in this embodiment may include the following structure:
a memory 1301 for storing applications and data generated by the application operations;
an input/output device 1302 for receiving input and outputting content, such as a display with a touch function or the like;
a processor 1303 configured to execute the application program in the memory to implement: receiving current input content of a current user through the input-output device 1302; obtaining at least one target user behavior corresponding to the current input content of the current user in a conversation track set; the dialog track set comprises at least one historical user behavior in a target historical dialog, the historical user behavior is a behavior tag corresponding to the input content of a historical user in the interaction between the historical user and a customer service system, and the characteristic value of at least one log feature of a target dialog log corresponding to the target historical dialog meets a preset high satisfaction condition; at least the at least one target user behavior is output via input-output device 1302.
It can be seen from the foregoing technical solutions that, in the electronic device provided in the third embodiment of the present application, after receiving the current input content of the current user, in the historical user behaviors in at least one target historical dialog included in the dialog track set, at least one target user behavior corresponding to the current input content is obtained and output, where the historical user behaviors are behavior tags corresponding to the input content of the historical user in the interaction between the historical user and the customer service system, and the historical user behaviors are user behaviors in the target historical dialog corresponding to the dialog log in which the feature value of the log feature satisfies the high satisfaction condition. Therefore, in the embodiment, the target user behavior corresponding to the currently input content of the current user is obtained from the historical user behaviors in the historical dialog satisfying the high satisfaction condition and is output to the current user, and based on this, the current user can select the user behavior more meeting the requirement from the output target user behaviors, so that the response content can be fed back to the current user according to the selected target user behavior, and the accuracy of the response content is improved.
Taking an intelligent customer service system as an example, the technical scheme in the application is explained in detail as follows:
firstly, the technical scheme of the application is mainly an implementation scheme which can recommend the user behavior in combination with the conversation track satisfaction degree of the user in the intelligent customer service system, and the user can continue to carry out conversation by referring to the recommended conversation behavior with higher historical satisfaction degree in a user behavior recommendation mode, so that more selections and guidance are provided for the user.
Referring to the flow shown in fig. 14, the technical solution in the present application is mainly divided into the following parts:
1. clustering the existing (historical) offline conversation logs according to log features, wherein the clustering features are log features related to conversation satisfaction, and the specific features are X in the following table 1 1 -X 7 Shown, and the number of clustering categories is 2, namely a category with high satisfaction and a category with low satisfaction;
TABLE 1
2. In the history conversations corresponding to the clustered logs with high satisfaction, all conversation tracks are extracted by taking user behaviors and system behaviors as nodes, wherein the user behaviors are semantic labels obtained according to a semantic understanding module, the system behaviors are a defined conversation strategy set during system design, all the conversation tracks are mapped to corresponding satisfaction categories, and examples of the user behaviors and the system behaviors are shown in the following tables 2 and 3:
TABLE 2 user behavior
| User behavior |
| Asking relevant questions |
| Providing slot position information |
| Chatting machine |
| Manual customer service |
| Open field white |
| The answers are not understood |
| Fail after trial |
| More solutions are needed |
| Seeking help |
| Active exit |
| …… |
TABLE 3 System behavior
3. After the real-time conversation starts, whether negative user feedback occurs in the conversation process is continuously monitored, if negative user feedback occurs, a user behavior recommendation flow is triggered, as shown in a left branch in fig. 13, and the recommendation method can refer to a collaborative filtering method based on item, and the specific steps are as follows:
(1) In a conversation track set with high satisfaction degree, calculating the similarity between the conversation track of the current user and the conversation tracks in the set according to the semantic distance to obtain a next-step user behavior candidate set, wherein the conversation tracks in the candidate set are the conversation tracks which are screened out from the conversation track set and have similarity with the conversation track of the current user higher than a similarity threshold;
(2) Sequencing each user behavior in the candidate set according to the tf-idf characteristic value of the system behavior in the previous round of the conversation where the user behavior is located, so as to obtain a most possible next-step user behavior recommendation list, wherein the user behavior recommendation list comprises a plurality of target user behaviors, and the target user behaviors are sequenced according to the tf-idf characteristic values of the system behaviors in the previous round of the conversation where the target user behaviors are located;
(3) And outputting the user behavior recommendation list and the reply content of the current intelligent customer service system to the user.
Based on this, the user can click the next step on the current recommendation list, or can input again without selecting.
Therefore, in the technical scheme of the application, when the user has bad experience in the conversation process, the intelligent customer service system can actively recommend the next possible behaviors of the user, the recommended behaviors come from the conversation processes with similar conversation tracks in the log, and further guidance is provided for the user who is taken or wants to leave. In the specific implementation of the application, user behaviors and system behaviors are used as nodes of conversation tracks, all conversation tracks in an existing conversation log are mined, the satisfaction degrees of the conversation tracks are clustered, unsupervised conversation track satisfaction degree prediction is realized, further, in the real-time conversation process, when negative feedback user behaviors occur, the intelligent customer service system can actively recommend possible next-step user behaviors to be selected for users, the next-step user behaviors recommended by the intelligent customer service system are based on a collaborative filtering method, the selection range is narrowed during clustering, only the conversation tracks with high satisfaction degrees are searched for according to the prediction results, and the recommended user behaviors can refer to similar conversations with high user satisfaction degrees.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (9)
1. A method for processing data includes the steps of, the method comprises the following steps:
receiving current input content of a current user, wherein the current input content refers to the latest input content in the interaction between the current user and a customer service system;
obtaining at least one target user behavior corresponding to the current input content of the current user from the historical user behaviors in at least one target historical dialogue contained in the dialogue track set;
the dialog track set comprises at least one historical user behavior in a target historical dialog, the historical user behavior is a behavior tag corresponding to the input content of a historical user in the interaction between the historical user and a customer service system, the characteristic value of at least one log feature of a target dialog log corresponding to the target historical dialog meets a preset high satisfaction condition, and the semantic similarity between the dialog track corresponding to the target user behavior and the dialog track corresponding to the current input content is higher than a preset similarity threshold;
outputting at least the at least one target user behavior so that a current user can select one or more target user behaviors therein;
before obtaining at least one target user behavior corresponding to current input content of the current user in the dialog track set, the method further includes:
judging whether the current input content meets a preset behavior recommendation condition; the behavior recommendation condition comprises: the current input content represents that the previous feedback content of the customer service system does not accord with the previous input content of the current user;
if the current input content meets the behavior recommendation condition, executing the following steps: obtaining at least one target user behavior corresponding to the current input content of the current user in a conversation track set;
and if the current input content does not meet the behavior recommendation condition, outputting the current feedback content generated aiming at the current input content.
2. The method of claim 1, obtaining at least one target user behavior in a set of conversation tracks corresponding to current input content of the current user, comprising:
obtaining a first action set in a preset conversation track set, wherein the first action set comprises at least one first user action, and the semantic similarity between a conversation track corresponding to the first user action and a conversation track corresponding to the current input content is higher than a preset similarity threshold;
ordering first user behaviors in the first set of behaviors;
and selecting the user behaviors ranked at the top M in the first behavior set as target user behaviors, wherein M is a positive integer greater than or equal to 1.
3. The method of claim 2, wherein the dialog track set further comprises historical system behaviors in the target historical dialog, and the historical system behaviors are behavior tags corresponding to feedback contents of the customer service system in interaction between the historical user and the customer service system;
wherein ranking the first user actions in the first set of actions comprises:
obtaining a corresponding previous target system behavior of the first user behavior in a target history conversation where the first user behavior is located;
obtaining a first frequency value of the target system behavior in the first behavior set, and obtaining a second frequency value of the target system behavior in the dialog track set;
ranking the first user behavior in the first set of behaviors according to the first frequency value and the second frequency value.
4. The method of claim 1, the set of conversation tracks obtained by:
obtaining a history log set for interaction between the customer service system and a plurality of history users, wherein the history log set comprises a plurality of history conversation logs;
obtaining at least one log feature of the historical dialog log, wherein the log feature has a feature value which represents the interactive satisfaction degree of the corresponding historical user to the customer service system on the corresponding dimension of the log feature;
obtaining at least one target dialogue log in the historical log set according to the characteristic value of the log characteristic, wherein the characteristic value of the log characteristic of any one target dialogue log meets a preset high satisfaction degree condition;
and obtaining a historical dialogue track in the target historical dialogue corresponding to the target dialogue log, wherein the historical dialogue track comprises at least one node corresponding to historical user behavior and at least one node corresponding to historical system behavior, and the historical user behavior and the historical system behavior form a dialogue track set.
5. The method of claim 4, the high satisfaction condition comprising: and respectively weighting the characteristic values of the log characteristics of the target dialog logs, and then summing the weighted characteristic values to obtain a value larger than a preset satisfaction threshold.
6. The method of claim 1, outputting at least the at least one target user behavior, comprising:
outputting the at least one target user behavior and current feedback content generated for the current input content.
7. The method of claim 6, outputting the at least one target user behavior and current feedback content generated for the current input content, comprising:
outputting current feedback content generated aiming at the current input content at a target position on an interactive interface corresponding to the current user;
and outputting the at least one target user behavior in a list form in a target area associated with the target position.
8. The method of claim 1, further comprising:
receiving operation data of the current user for a second user behavior in the at least one target user behavior;
and outputting feedback content corresponding to the second user behavior.
9. A data processing apparatus comprising:
the system comprises an input receiving unit, a processing unit and a display unit, wherein the input receiving unit is used for receiving current input content of a current user, and the current input content refers to the latest input content in the interaction between the current user and a customer service system;
the behavior obtaining unit is used for obtaining at least one target user behavior corresponding to the current input content of the current user from the historical user behaviors in at least one target historical dialogue contained in the dialogue track set;
the dialog track set comprises at least one historical user behavior in a target historical dialog, the historical user behavior is a behavior tag corresponding to the input content of a historical user in the interaction between the historical user and a customer service system, the characteristic value of at least one log feature of a target dialog log corresponding to the target historical dialog meets a preset high satisfaction condition, and the semantic similarity between the dialog track corresponding to the target user behavior and the dialog track corresponding to the current input content is higher than a preset similarity threshold;
the behavior output unit is used for at least outputting the at least one target user behavior so that a current user can select one or more target user behaviors;
the recommendation judging unit is used for judging whether the current input content meets a preset behavior recommendation condition or not before the behavior obtaining unit obtains at least one target user behavior corresponding to the current input content of the current user in the conversation track set; the behavior recommendation condition comprises: the current input content represents that the previous feedback content of the customer service system does not accord with the previous input content of the current user; and if the current input content meets the behavior recommending condition, executing the behavior obtaining unit, and if the current input content does not meet the behavior recommending condition, triggering the behavior output unit to output current feedback content generated aiming at the current input content.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010499962.1A CN111652001B (en) | 2020-06-04 | 2020-06-04 | Data processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010499962.1A CN111652001B (en) | 2020-06-04 | 2020-06-04 | Data processing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111652001A CN111652001A (en) | 2020-09-11 |
| CN111652001B true CN111652001B (en) | 2023-01-17 |
Family
ID=72345048
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010499962.1A Active CN111652001B (en) | 2020-06-04 | 2020-06-04 | Data processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111652001B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112862327A (en) * | 2021-02-19 | 2021-05-28 | 杭州拼便宜网络科技有限公司 | Service label generation method, device and medium |
| CN112949305B (en) * | 2021-05-13 | 2021-08-13 | 平安科技（深圳）有限公司 | Negative feedback information acquisition method, device, equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103389988A (en) * | 2012-05-10 | 2013-11-13 | 腾讯科技（深圳）有限公司 | Method and device for guiding user to carry out information search |
| CN104951428A (en) * | 2014-03-26 | 2015-09-30 | 阿里巴巴集团控股有限公司 | User intention recognition method and device |
| CN106202301A (en) * | 2016-07-01 | 2016-12-07 | 武汉泰迪智慧科技有限公司 | A kind of intelligent response system based on degree of depth study |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106599278B (en) * | 2016-12-23 | 2020-06-12 | 北京奇虎科技有限公司 | Application search intention identification method and device |
| US10387463B2 (en) * | 2017-07-06 | 2019-08-20 | International Business Machines Corporation | Dialog agent for conducting task-oriented computer-based communications |
| CN109670025B (en) * | 2018-12-19 | 2023-06-16 | 北京小米移动软件有限公司 | Dialogue management method and device |
| CN110609890A (en) * | 2019-09-11 | 2019-12-24 | 浙江大搜车软件技术有限公司 | Self-service customer service interaction method and device, computer equipment and storage medium |
| CN111143524A (en) * | 2019-12-06 | 2020-05-12 | 联想(北京)有限公司 | User intention determining method and electronic equipment |
-
2020
- 2020-06-04 CN CN202010499962.1A patent/CN111652001B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103389988A (en) * | 2012-05-10 | 2013-11-13 | 腾讯科技（深圳）有限公司 | Method and device for guiding user to carry out information search |
| CN104951428A (en) * | 2014-03-26 | 2015-09-30 | 阿里巴巴集团控股有限公司 | User intention recognition method and device |
| CN106202301A (en) * | 2016-07-01 | 2016-12-07 | 武汉泰迪智慧科技有限公司 | A kind of intelligent response system based on degree of depth study |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111652001A (en) | 2020-09-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104951428B (en) | User's intension recognizing method and device | |
| CN107391493B (en) | Public opinion information extraction method and device, terminal equipment and storage medium | |
| CN111414462B (en) | Dialogue statement determination method, device, computer equipment and medium | |
| CN113360622B (en) | User dialogue information processing method and device and computer equipment | |
| CN111652001B (en) | Data processing method and device | |
| CN110059182A (en) | Art recommended method and device towards customer service | |
| EP2666101A1 (en) | Method and system of selecting responders | |
| KR102218468B1 (en) | Answer recommendation system and method based on text content and emotion analysis | |
| CN113256397B (en) | Commodity recommendation method and system based on big data and computer-readable storage medium | |
| CN108198017B (en) | Personalized customer service processing method and device | |
| EP3685245B1 (en) | Method, apparatus, and computer-readable media for customer interaction semantic annotation and analytics | |
| CN109933782B (en) | User emotion prediction method and device | |
| CN111626863A (en) | Intelligent recommendation method for financial products | |
| Daniati et al. | Decision Making Framework Based On Sentiment Analysis in Twitter Using SAW and Machine Learning Approach | |
| CN111368131B (en) | User relationship identification method and device, electronic equipment and storage medium | |
| CN112435059A (en) | Method and device for evaluating value of article in real time, electronic equipment and storage medium | |
| CN112182189A (en) | Conversation processing method and device, electronic equipment and storage medium | |
| CN114722266A (en) | Questionnaire pushing method and device, electronic equipment and readable storage medium | |
| CN112365302B (en) | Product recommendation network training method, device, equipment and medium | |
| CN114331572A (en) | Potential customer determination method and device, electronic equipment and storage medium | |
| KR102078541B1 (en) | Issue interest based news value evaluation apparatus and method, storage media storing the same | |
| CN113901328A (en) | Information recommendation method and device, electronic equipment and storage medium | |
| CN112396498A (en) | Commodity sales promotion method, device, equipment and storage medium | |
| CN110929154A (en) | Search result sorting method and system | |
| CN111723189A (en) | Interactive question and answer prompting method and device, storage medium and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |