CN111984764A - Stream bill reason analysis method and device, computer equipment and readable storage medium - Google Patents
Stream bill reason analysis method and device, computer equipment and readable storage medium Download PDFInfo
- Publication number
- CN111984764A CN111984764A CN202010906344.4A CN202010906344A CN111984764A CN 111984764 A CN111984764 A CN 111984764A CN 202010906344 A CN202010906344 A CN 202010906344A CN 111984764 A CN111984764 A CN 111984764A
- Authority
- CN
- China
- Prior art keywords
- reason
- flow
- intention
- sample
- dialog text
- 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.)
- Withdrawn
Links
Images
Classifications
-
- 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
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- 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
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
Abstract
The application provides a method and a device for analyzing reason of a flow list, computer equipment and a readable storage medium, comprising the following steps: when determining that a target user in a target project is lost, acquiring a dialog text between the target user and a service user; inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text; and determining the reason of the flow order of the target user based on the flow order reason label. According to the method and the device, the conversation text between the target user and the service user can be analyzed in real time based on the flow list recognition model, and the analysis efficiency of the conversation text is improved, so that the flow list direction and reason of the target user can be rapidly responded, and the efficiency of the service user in processing the flow list problem is improved.
Description
Technical Field
The present application relates to the field of machine learning technologies, and in particular, to a method and an apparatus for analyzing a reason for a flow order, a computer device, and a readable storage medium.
Background
In the service industry, especially in some online service industries, in order to better serve users, the content of a conversation between a user and a customer service staff is generally analyzed to analyze the advantages and disadvantages of the customer service staff in communicating with the user, so as to better communicate with the user in the following.
At present, the analysis work of the conversation content is often performed manually, that is, the information such as different questions and answers, the willingness and attitude shown by the user in the conversation content is summarized and summarized by a manual analysis method so as to optimize the subsequent communication process, but because the processing is performed manually, many subjective factors are often mixed, and for a large amount of conversation content, many human resources are required to be consumed, and the processing efficiency is low.
Disclosure of Invention
In view of the above, an object of the present application is to provide a method, an apparatus, a computer device, and a readable storage medium for analyzing a reason of a flow sheet, which can analyze a dialog text between a target user and a service user in real time based on a flow sheet recognition model, thereby improving an analysis efficiency of the dialog text, so as to quickly cope with a flow single direction and a reason of the target user, and improve an efficiency of the service user in handling a flow sheet problem.
The embodiment of the application provides a flow list reason analysis method, which comprises the following steps:
when determining that a target user in a target project is lost, acquiring a dialog text between the target user and a service user;
inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text;
and determining the reason of the flow order of the target user based on the flow order reason label.
In an optional embodiment, determining the target user churn includes:
acquiring a dialog text between the target user and a service user;
inputting the dialog text into a pre-trained intention recognition model, and outputting an intention recognition result corresponding to the dialog text;
determining that the target user is lost when the intent recognition result is a negative intent.
In an alternative embodiment, the intent recognition model is trained by:
constructing a first sample dialog text; wherein the first sample dialog text comprises: sample intention keywords for representing the intention of the user and intention labels corresponding to the sample intention keywords; wherein the intent tags include an active intent tag and a passive intent tag;
and training model parameters of the intention recognition model according to an output result of the intention recognition model based on the input sample intention keywords and intention labels corresponding to the sample intention keywords until the loss of the intention recognition model meets a preset condition, so as to obtain the intention recognition model comprising the trained model parameters.
In an alternative embodiment, the constructing the first sample dialog text includes:
acquiring an original dialog text of a user and a service person;
performing word segmentation on the original dialog text, extracting sample intention keywords used for representing the intention of a user in the original dialog text, and setting an intention label of each sample intention keyword;
and constructing the first sample dialog text according to the original dialog text, the sample intention keywords and the intention label of each sample intention keyword.
In an alternative embodiment, the training of the flow sheet recognition model is performed by a method comprising:
constructing a second sample dialog text; wherein the second sample dialog text comprises: the system comprises a receipt reason keyword and an intention label, wherein the receipt reason keyword is used for representing a receipt reason;
and taking the second sample dialog text as the input of a flow sheet recognition model, training model parameters of the flow sheet recognition model according to the output result of the flow sheet recognition model based on the input second sample dialog text and a flow sheet reason label corresponding to the flow sheet reason keyword in the second sample dialog text until the loss of the flow sheet recognition model meets a preset condition, and obtaining the flow sheet recognition model comprising the trained model parameters.
In an optional embodiment, the analysis method further comprises:
acquiring conversation voice between a target user and a service user;
and carrying out voice recognition processing on the conversation voice to obtain a conversation text corresponding to the conversation voice.
In an optional embodiment, the determining the reason for the target user based on the reason for the flow order tag includes:
counting the total number of each flow list reason label based on the flow list reason label corresponding to each dialog text;
selecting a target flow list reason label with the maximum corresponding total number from the flow list reason labels based on the total number of each flow list reason label;
and determining the reason of the flow list indicated by the target flow list reason label as the reason of the flow list of the target user.
In an optional embodiment, the analysis method further comprises:
determining a solution corresponding to the reason of the flow order based on the reason of the flow order of the target user;
operating the target item based on the solution.
The embodiment of the present application further provides a flow list reason analysis device, where the analysis device includes:
the first acquisition module is used for acquiring a dialog text between a target user and a service user when the target user is determined to be lost in a target project;
the first input module is used for inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text;
and the first determining module is used for determining the reason of the flow list of the target user based on the flow list reason label.
In an alternative embodiment, the analysis device further comprises:
the second acquisition module is used for acquiring a dialog text between the target user and the service user;
the second input module is used for inputting the dialog text into a pre-trained intention recognition model and outputting an intention recognition result corresponding to the dialog text;
a second determination module to determine that the target user is lost when the intent recognition result is a negative intent.
In an alternative embodiment, the analysis device further comprises: a first training module.
The first training module is configured to train the intent recognition model, and the first training module includes:
a first construction unit for constructing a first sample dialog text; wherein the first sample dialog text comprises: sample intention keywords for representing the intention of the user and intention labels corresponding to the sample intention keywords; wherein the intent tags include an active intent tag and a passive intent tag;
the first training unit is used for taking a sample intention keyword as input of an intention recognition model, training model parameters of the intention recognition model according to an output result of the intention recognition model based on the input sample intention keyword and an intention label corresponding to the sample intention keyword until loss of the intention recognition model meets a preset condition, and obtaining the intention recognition model comprising the trained model parameters.
In an alternative embodiment, the first construction element is specifically configured to:
acquiring an original dialog text of a user and a service person;
performing word segmentation on the original dialog text, extracting sample intention keywords used for representing the intention of a user in the original dialog text, and setting an intention label of each sample intention keyword;
and constructing the first sample dialog text according to the original dialog text, the sample intention keywords and the intention label of each sample intention keyword.
In an alternative embodiment, the analysis device further comprises: a second training module.
The second training module is configured to train the slip recognition model, and the second training module includes:
a second construction unit for constructing a second sample dialogue text; wherein the second sample dialog text comprises: the system comprises a receipt reason keyword and an intention label, wherein the receipt reason keyword is used for representing a receipt reason;
and the second training unit is used for taking the second sample dialog text as the input of the flow sheet recognition model, and training the model parameters of the flow sheet recognition model according to the output result of the flow sheet recognition model based on the input second sample dialog text and the flow sheet reason label corresponding to the flow sheet reason keyword in the second sample dialog text until the loss of the flow sheet recognition model meets the preset condition, so as to obtain the flow sheet recognition model comprising the trained model parameters.
In an alternative embodiment, the analysis device further comprises:
the third acquisition module is used for acquiring the conversation voice between the target user and the service user;
and the voice recognition module is used for carrying out voice recognition processing on the conversation voice to obtain a conversation text corresponding to the conversation voice.
In an optional implementation manner, the first determining module is specifically configured to:
counting the total number of each flow list reason label based on the flow list reason label corresponding to each dialog text;
selecting a target flow list reason label with the maximum corresponding total number from the flow list reason labels based on the total number of each flow list reason label;
and determining the reason of the flow list indicated by the target flow list reason label as the reason of the flow list of the target user.
In an alternative embodiment, the analysis device further comprises:
a third determining module, configured to determine, based on a reason for the flow order of the target user, a solution corresponding to the reason for the flow order;
an operation module for operating the target project based on the solution.
There is also provided a computer device, a processor, and a memory, where the memory stores machine-readable instructions executable by the processor, and the processor is configured to execute the machine-readable instructions stored in the memory, and when the machine-readable instructions are executed by the processor, the machine-readable instructions are executed by the processor to perform the steps of any one of the above-mentioned possible embodiments.
An embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and the computer program is executed to perform the steps in any one of the above-mentioned possible implementation manners.
The embodiment of the application provides a method and a device for analyzing a reason of a flow list, computer equipment and a readable storage medium, and the method comprises the following steps: when determining that a target user in a target project is lost, acquiring a dialog text between the target user and a service user; inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text; and determining the reason of the flow order of the target user based on the flow order reason label. According to the method and the device, the conversation text between the target user and the service user can be analyzed in real time based on the flow list recognition model, and the analysis efficiency of the conversation text is improved, so that the flow list direction and reason of the target user can be rapidly responded, and the efficiency of the service user in processing the flow list problem is improved.
In order to make the aforementioned objects, features and advantages of the present application more comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.
Fig. 1 is a flowchart illustrating a method for analyzing a reason for a flow list according to an embodiment of the present disclosure;
fig. 2 shows one of schematic diagrams of a flow single reason analysis apparatus provided in an embodiment of the present application;
fig. 3 shows a second schematic diagram of a single-reason flow analysis apparatus provided in the embodiment of the present application;
fig. 4 shows a third schematic diagram of a single-reason flow analysis apparatus provided in an embodiment of the present application;
fig. 5 shows a schematic structural diagram of a computer device provided in an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, 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 the embodiments. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without making any creative effort, shall fall within the protection scope of the present application.
Research shows that at present, the analysis work of the conversation content is often performed manually, that is, the information such as different questions and answers, the willingness and attitude shown by the user in the conversation content is summarized and summarized by a manual analysis method so as to optimize the subsequent communication process.
Based on the above research, an embodiment of the present application provides a method, an apparatus, a computer device, and a readable storage medium for analyzing a reason for a flow order, including: when determining that a target user in a target project is lost, acquiring a dialog text between the target user and a service user; inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text; and determining the reason of the flow order of the target user based on the flow order reason label. According to the method and the device, the conversation text between the target user and the service user can be analyzed in real time based on the flow list recognition model, and the analysis efficiency of the conversation text is improved, so that the flow list direction and reason of the target user can be rapidly responded, and the efficiency of the service user in processing the flow list problem is improved.
The above-mentioned drawbacks are the results of the inventor after practical and careful study, and therefore, the discovery process of the above-mentioned problems and the solution proposed by the present application to the above-mentioned problems in the following should be the contribution of the inventor to the present application in the process of the present application.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.
To facilitate understanding of the present embodiment, first, a detailed description is given to a flow list reason analysis method disclosed in an embodiment of the present application, where an execution subject of the flow list reason analysis method provided in the embodiment of the present application is generally a computer device with certain computing capability, and the computer device includes, for example: a terminal device, which may be a User Equipment (UE), a mobile device, a User terminal, a cellular phone, a cordless phone, a Personal Digital Assistant (PDA), a handheld device, a computing device, a vehicle mounted device, a wearable device, or a server or other processing device. In some possible implementations, the flow single-reason analysis method may be implemented by a processor calling computer-readable instructions stored in a memory.
The following describes a method for analyzing a reason for a flow list provided in the embodiment of the present application, by taking an execution subject as a terminal device as an example.
Referring to fig. 1, a flowchart of a flow list reason analysis method provided in an embodiment of the present application is shown, where the analysis method includes steps S101 to S103, where:
s101: when determining that a target user in a target project is lost, acquiring a dialog text between the target user and a service user;
s102: inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text;
s103: and determining the reason of the flow order of the target user based on the flow order reason label.
According to the method and the device, when the loss of the target user in the target project is determined, the dialog text between the target user and the service user is obtained; inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text; and determining the reason of the flow order of the target user based on the flow order reason label. According to the method and the device, the conversation text between the target user and the service user can be analyzed in real time based on the flow list recognition model, and the analysis efficiency of the conversation text is improved, so that the flow list direction and reason of the target user can be rapidly responded, and the efficiency of the service user in processing the flow list problem is improved.
The following describes each of the above-mentioned steps S101 to S103 in detail.
First, in the above S101, after determining that the target user in the target item runs away, a dialog text between the target user and the service user in the corresponding target item is obtained.
The target user may be determined to be lost first, and the method for determining the target user loss includes:
acquiring a dialog text between the target user and a service user;
inputting the dialog text into a pre-trained intention recognition model, and outputting an intention recognition result corresponding to the dialog text;
determining that the target user is lost when the intent recognition result is a negative intent.
In this step, the objective is to determine the intention of the target user, including positive intention and negative intention, wherein, taking the clothing purchase as an example, the positive intention may include "want", "like", "fit", etc., and words representing the purchase intention, and the corresponding negative intention may include "dislike", "do not want", "expensive", "not fit", etc., and words not representing the purchase intention are not wanted, and when the positive intention is determined, no subsequent processing is needed to reduce the data processing amount. However, when the negative intention is determined, in order to promote the target user to purchase the commodity, the service user may take a certain follow-up measure against the negative intention of the target user, that is, only when the negative intention is determined, the follow-up process is performed to reduce the amount of data processing.
Preferably, the intention recognition result corresponding to the dialog text, namely the positive intention and the negative intention, can be output by inputting the dialog text into a pre-trained intention recognition model.
Specifically, for an intention recognition model, the intention recognition model may be trained by:
constructing a first sample dialog text; wherein the first sample dialog text comprises: sample intention keywords for representing the intention of the user and intention labels corresponding to the sample intention keywords; wherein the intent tags include an active intent tag and a passive intent tag;
and training model parameters of the intention recognition model according to an output result of the intention recognition model based on the input sample intention keywords and intention labels corresponding to the sample intention keywords until the loss of the intention recognition model meets a preset condition, so as to obtain the intention recognition model comprising the trained model parameters.
The method for constructing the first sample dialog text comprises the following steps:
acquiring an original dialog text of a user and a service person;
performing word segmentation on the original dialog text, extracting sample intention keywords used for representing the intention of a user in the original dialog text, and setting an intention label of each sample intention keyword;
and constructing the first sample dialog text according to the original dialog text, the sample intention keywords and the intention label of each sample intention keyword.
For example, an intention recognition model can be constructed in advance, and keywords which embody the negative intention of the target user can be set. For example, "i do not want this," "too expensive," "dislike this color," etc., and give a certain intention label to the keyword, and store the keyword into an intention recognition model in advance as a dialog text recognition instruction, so that the dialog text can be quickly recognized when performing dialog text recognition, and the dialog text can be compared and matched. That is, when there is a keyword with negative intention such as "this commodity is too expensive" or "this color is disliked" appearing in the dialog text, the slip reason recognition mechanism is triggered to provide a prerequisite for the subsequent slip reason recognition. That is, the dialog text is compared with the keywords in the intention recognition model, and if the keywords of the negative intention in the intention recognition model appear, the target user is considered to belong to the possible waybill, that is, the target user does not have the commodity purchasing tendency, and then the next waybill reason analysis is carried out.
In addition to text dialog between a target user and a service user during online service, there are some dialog voices corresponding to offline service, or voice messages sent by the target user during online service, and the like, and for the dialog voices, in the embodiment of the present application, the following contents are included:
for conversational speech, the method for analyzing the reason for the flow list may further include:
acquiring conversation voice between a target user and a service user;
and carrying out voice recognition processing on the conversation voice to obtain a conversation text corresponding to the conversation voice.
In the embodiment of the present application, a portable mobile terminal capable of recording and displaying text may be adopted, for example: the intelligent mobile phone, the recording pen and the like are used for recording the conversation voice between the target user and the service user in real time and carrying out voice recognition processing on the conversation voice. Or the mobile terminal can also transmit the conversation voice to the server in real time, and the server performs voice recognition processing on the conversation voice.
Illustratively, the dialog Speech may be converted into corresponding dialog text by Automatic Speech Recognition (ASR). However, due to the diversity and complexity of the speech to be processed, speech recognition systems can only achieve satisfactory performance with certain constraints or can only be used in certain specific situations.
In the embodiment of the application, the acquired conversation voice can be identified, and the identification result is displayed on the display screen of the mobile terminal for the service user to check.
Secondly, in the above S102, after determining that the target user churn occurs and acquiring the dialog text, the corresponding dialog text of the flow sheet may be input into the flow sheet recognition model to analyze the reason for the corresponding flow sheet.
The flow sheet recognition model can be trained by the following method, including:
constructing a second sample dialog text; wherein the second sample dialog text comprises: the system comprises a receipt reason keyword and an intention label, wherein the receipt reason keyword is used for representing a receipt reason;
and taking the second sample dialog text as the input of a flow sheet recognition model, training model parameters of the flow sheet recognition model according to the output result of the flow sheet recognition model based on the input second sample dialog text and a flow sheet reason label corresponding to the flow sheet reason keyword in the second sample dialog text until the loss of the flow sheet recognition model meets a preset condition, and obtaining the flow sheet recognition model comprising the trained model parameters.
For example, a plurality of flow order reason keywords may be preset, for example: the method includes the steps of selecting a plurality of stream single-reason keywords such as 'noble', 'small size', 'dislike color', 'old style', and the like, wherein the corresponding label of 'noble' is 'price', 'small size' is 'size', and the like. And inputting the constructed second sample dialog text into a flow sheet recognition model, training model parameters of the flow sheet recognition model based on an output result and a flow sheet reason label corresponding to the flow sheet reason keyword in the second sample dialog text, and obtaining the flow sheet recognition model comprising the trained model parameters when the model loss of the flow sheet recognition model reaches a preset condition.
In the embodiment of the application, the dialog text is input into the model based on the pre-trained flow sheet recognition model, so that a flow sheet reason label corresponding to the dialog text can be obtained and used for subsequent analysis of the flow sheet reason.
Thirdly, in the above S103, after the reason tag of the flow list is acquired, the flow list reason of the target user can be determined based on the reason tag.
Wherein the determining the reason for the flow order of the target user based on the flow order reason label comprises:
counting the total number of each flow list reason label based on the flow list reason label corresponding to each dialog text;
selecting a target flow list reason label with the maximum corresponding total number from the flow list reason labels based on the total number of each flow list reason label;
and determining the reason of the flow list indicated by the target flow list reason label as the reason of the flow list of the target user.
For example, for one dialog text, there may be a plurality of corresponding reason labels for the flow list, but the reason may be a primary or secondary score, so that the service user may preferentially process based on the primary reason. For example: if the price label appears 8 times and the color label appears 3 times, the price can be judged to be the main reason of the flow list, and the processing can be preferentially carried out based on the price reason.
In the embodiment of the present application, the recognition process for the dialog text may further include the following processes: the method comprises the steps of splitting data of a conversation text to obtain a plurality of conversation short texts, inputting each conversation short text into a single reason recognition model, and determining a specific reason label of each conversation short text; counting the number of all reason labels, taking the reason label with the highest number as the final reason label of the recognized text, and determining the reason of the flow list according to the final reason label, so that the speed of text recognition can be increased, and the processing efficiency can be improved.
After determining the reason for the flow order of the target user, in the embodiment of the present application, a coping method is further provided, that is, the reason for the flow order of the current target user is analyzed through a flow order reason identification model, and a corresponding solution is searched to form notification information and send the notification information to the service user in real time, where the coping method specifically includes:
determining a solution corresponding to the reason of the flow order based on the reason of the flow order of the target user;
operating the target item based on the solution.
Illustratively, the analyze flow sheet reason label is: if the color is not preferred, a solution list corresponding to the reason label is searched, wherein the solution list can be preset based on human or expert research. From the solution list, the corresponding solution is determined to be: and searching all color products of the target commodity and the product allowance of each color.
Specifically, when the service user and the target user discuss the target article a, the reason label is analyzed to be color dislike according to the conversation content, all color products and the allowance of the target article a are quickly inquired and displayed on the mobile terminal of the service user in real time, and the display information is as follows:
a 1-yellow, remainder: some (the balance is over 10);
a 2-green, remainder: 0;
a 3-violet, remainder: 2.
thus, the service user can conduct subsequent discussion with the target user based on the display content to promote positive intentions of the target user.
According to the embodiment of the application, the flow note reason label is added to the conversation text in the transaction scene through the manual and machine learning algorithm, model training is carried out, and the flow note reason identification model is obtained, so that the flow note reason of the conversation text can be determined based on the model, a corresponding solution is adopted, and the transaction rate is promoted.
It will be understood by those skilled in the art that in the method of the present invention, the order of writing the steps does not imply a strict order of execution and any limitations on the implementation, and the specific order of execution of the steps should be determined by their function and possible inherent logic.
Based on the same inventive concept, the embodiment of the present application further provides a flow sheet reason analysis device corresponding to the flow sheet reason analysis method, and because the principle of solving the problem of the device in the embodiment of the present application is similar to that of the flow sheet reason analysis method in the embodiment of the present application, the implementation of the device may refer to the implementation of the method, and repeated details are not repeated.
Referring to fig. 2, 3 and 4, fig. 2 is a schematic diagram of a single flow cause analysis apparatus according to an embodiment of the present disclosure, fig. 3 is a second schematic diagram of the single flow cause analysis apparatus according to the embodiment of the present disclosure, and fig. 4 is a third schematic diagram of the single flow cause analysis apparatus according to the embodiment of the present disclosure. The analysis device includes: a first obtaining module 210, a first input module 220, a first determining module 230;
the first obtaining module 210 is configured to obtain a dialog text between a target user and a service user when it is determined that the target user is lost in a target item;
the first input module 220 is configured to input the dialog text into a pre-trained flow sheet recognition model, so as to obtain a flow sheet reason tag corresponding to the dialog text;
a first determining module 230, configured to determine a reason for the flow order of the target user based on the flow order reason tag.
The embodiment of the application provides a method for acquiring a dialog text between a target user and a service user when the target user is lost in a target project; inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text; and determining the reason of the flow order of the target user based on the flow order reason label. According to the method and the device, the conversation text between the target user and the service user can be analyzed in real time based on the flow list recognition model, and the analysis efficiency of the conversation text is improved, so that the flow list direction and reason of the target user can be rapidly responded, and the efficiency of the service user in processing the flow list problem is improved.
In an alternative embodiment, as shown in fig. 3, the analysis device further comprises:
a second obtaining module 310, configured to obtain a dialog text between the target user and a service user;
the second input module 320 is configured to input the dialog text into a pre-trained intent recognition model, and output an intent recognition result corresponding to the dialog text;
a second determination module 330 for determining the target user churn when the intent recognition result is a negative intent.
In an alternative embodiment, the analysis device further comprises: a first training module 340.
The first training module 340 is configured to train the intention recognition model, and the first training module 340 includes:
a first construction unit for constructing a first sample dialog text; wherein the first sample dialog text comprises: sample intention keywords for representing the intention of the user and intention labels corresponding to the sample intention keywords; wherein the intent tags include an active intent tag and a passive intent tag;
the first training unit is used for taking a sample intention keyword as input of an intention recognition model, training model parameters of the intention recognition model according to an output result of the intention recognition model based on the input sample intention keyword and an intention label corresponding to the sample intention keyword until loss of the intention recognition model meets a preset condition, and obtaining the intention recognition model comprising the trained model parameters.
In an alternative embodiment, the first construction element is specifically configured to:
acquiring an original dialog text of a user and a service person;
performing word segmentation on the original dialog text, extracting sample intention keywords used for representing the intention of a user in the original dialog text, and setting an intention label of each sample intention keyword;
and constructing the first sample dialog text according to the original dialog text, the sample intention keywords and the intention label of each sample intention keyword.
In an alternative embodiment, as shown in fig. 4, the analysis device further comprises: a second training module 240.
The second training module 240 is configured to train the slip recognition model, and the second training module includes:
a second construction unit for constructing a second sample dialogue text; wherein the second sample dialog text comprises: the system comprises a receipt reason keyword and an intention label, wherein the receipt reason keyword is used for representing a receipt reason;
and the second training unit is used for taking the second sample dialog text as the input of the flow sheet recognition model, and training the model parameters of the flow sheet recognition model according to the output result of the flow sheet recognition model based on the input second sample dialog text and the flow sheet reason label corresponding to the flow sheet reason keyword in the second sample dialog text until the loss of the flow sheet recognition model meets the preset condition, so as to obtain the flow sheet recognition model comprising the trained model parameters.
In an alternative embodiment, the analysis device further comprises:
a third obtaining module 250, configured to obtain a dialogue voice between the target user and the service user;
and the voice recognition module 260 is configured to perform voice recognition processing on the conversation voice to obtain a conversation text corresponding to the conversation voice.
In an optional implementation manner, the first determining module 230 is specifically configured to:
counting the total number of each flow list reason label based on the flow list reason label corresponding to each dialog text;
selecting a target flow list reason label with the maximum corresponding total number from the flow list reason labels based on the total number of each flow list reason label;
and determining the reason of the flow list indicated by the target flow list reason label as the reason of the flow list of the target user.
In an alternative embodiment, the analysis device further comprises:
a third determining module 270, configured to determine, based on the reason for the flow order of the target user, a solution corresponding to the reason for the flow order;
an operation module 280 for operating the target item based on the solution.
An embodiment of the present application further provides a computer device, as shown in fig. 5, which is a schematic structural diagram of the computer device provided in the embodiment of the present application, where the computer device 10 includes:
a processor 11 and a memory 12; the memory 12 stores machine-readable instructions executable by the processor 11, which when executed by the computer device 10 are executed by the processor 11 to implement the steps of:
when determining that a target user in a target project is lost, acquiring a dialog text between the target user and a service user;
inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text;
and determining the reason of the flow order of the target user based on the flow order reason label.
In an alternative embodiment, the instructions executed by processor 11 to determine the target user churn include:
acquiring a dialog text between the target user and a service user;
inputting the dialog text into a pre-trained intention recognition model, and outputting an intention recognition result corresponding to the dialog text;
determining that the target user is lost when the intent recognition result is a negative intent.
In an alternative embodiment, the processor 11 executes instructions that train the intent recognition model by:
constructing a first sample dialog text; wherein the first sample dialog text comprises: sample intention keywords for representing the intention of the user and intention labels corresponding to the sample intention keywords; wherein the intent tags include an active intent tag and a passive intent tag;
and training model parameters of the intention recognition model according to an output result of the intention recognition model based on the input sample intention keywords and intention labels corresponding to the sample intention keywords until the loss of the intention recognition model meets a preset condition, so as to obtain the intention recognition model comprising the trained model parameters.
In an alternative embodiment, the instructions executed by processor 11 for constructing the first sample dialog text include:
acquiring an original dialog text of a user and a service person;
performing word segmentation on the original dialog text, extracting sample intention keywords used for representing the intention of a user in the original dialog text, and setting an intention label of each sample intention keyword;
and constructing the first sample dialog text according to the original dialog text, the sample intention keywords and the intention label of each sample intention keyword.
In an alternative embodiment, the processor 11 executes instructions that train the flow sheet recognition model by a method comprising:
constructing a second sample dialog text; wherein the second sample dialog text comprises: the system comprises a receipt reason keyword and an intention label, wherein the receipt reason keyword is used for representing a receipt reason;
and taking the second sample dialog text as the input of a flow sheet recognition model, training model parameters of the flow sheet recognition model according to the output result of the flow sheet recognition model based on the input second sample dialog text and a flow sheet reason label corresponding to the flow sheet reason keyword in the second sample dialog text until the loss of the flow sheet recognition model meets a preset condition, and obtaining the flow sheet recognition model comprising the trained model parameters.
In an alternative embodiment, in the instructions executed by the processor 11, the analysis method further includes:
acquiring conversation voice between a target user and a service user;
and carrying out voice recognition processing on the conversation voice to obtain a conversation text corresponding to the conversation voice.
In an alternative embodiment, the determining, by the processor 11 in the instruction executed based on the flow order reason tag, a flow order reason of the target user includes:
counting the total number of each flow list reason label based on the flow list reason label corresponding to each dialog text;
selecting a target flow list reason label with the maximum corresponding total number from the flow list reason labels based on the total number of each flow list reason label;
and determining the reason of the flow list indicated by the target flow list reason label as the reason of the flow list of the target user.
In an alternative embodiment, in the instructions executed by the processor 11, the analysis method further includes:
determining a solution corresponding to the reason of the flow order based on the reason of the flow order of the target user;
operating the target item based on the solution.
The embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the computer program performs the steps of the method for analyzing a flow order cause in the above method embodiment. The storage medium may be a volatile or non-volatile computer-readable storage medium.
The computer program product of the flow list reason analysis method provided in the embodiment of the present application includes a computer readable storage medium storing a program code, where instructions included in the program code may be used to execute the steps of the flow list reason analysis method described in the above method embodiment, which may be referred to in the above method embodiment specifically, and are not described herein again.
The embodiments of the present application also provide a computer program, which when executed by a processor implements any one of the methods of the foregoing embodiments. The computer program product may be embodied in hardware, software or a combination thereof. In an alternative embodiment, the computer program product is embodied in a computer storage medium, and in another alternative embodiment, the computer program product is embodied in a Software product, such as a Software Development Kit (SDK), or the like.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the system and the apparatus described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again. In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present application, and are used for illustrating the technical solutions of the present application, but not limiting the same, and the scope of the present application is not limited thereto, and although the present application is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope disclosed in the present application; such modifications, changes or substitutions do not depart from the spirit and scope of the exemplary embodiments of the present application, and are intended to be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (11)
1. A flow order cause analysis method is characterized by comprising the following steps:
when determining that a target user in a target project is lost, acquiring a dialog text between the target user and a service user;
inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text;
and determining the reason of the flow order of the target user based on the flow order reason label.
2. The method of claim 1, wherein determining the target user churn comprises:
acquiring a dialog text between the target user and a service user;
inputting the dialog text into a pre-trained intention recognition model, and outputting an intention recognition result corresponding to the dialog text;
determining that the target user is lost when the intent recognition result is a negative intent.
3. The method of claim 2, wherein the intent recognition model is trained by:
constructing a first sample dialog text; wherein the first sample dialog text comprises: sample intention keywords for representing the intention of the user and intention labels corresponding to the sample intention keywords; wherein the intent tags include an active intent tag and a passive intent tag;
and training model parameters of the intention recognition model according to an output result of the intention recognition model based on the input sample intention keywords and intention labels corresponding to the sample intention keywords until the loss of the intention recognition model meets a preset condition, so as to obtain the intention recognition model comprising the trained model parameters.
4. The method of claim 3, wherein the constructing a first sample dialog text comprises:
acquiring an original dialog text of a user and a service person;
performing word segmentation on the original dialog text, extracting sample intention keywords used for representing the intention of a user in the original dialog text, and setting an intention label of each sample intention keyword;
and constructing the first sample dialog text according to the original dialog text, the sample intention keywords and the intention label of each sample intention keyword.
5. The method for analyzing the reason for the flow sheet according to claim 1, wherein the flow sheet recognition model is trained by a method comprising:
constructing a second sample dialog text; wherein the second sample dialog text comprises: the system comprises a receipt reason keyword and an intention label, wherein the receipt reason keyword is used for representing a receipt reason;
and taking the second sample dialog text as the input of a flow sheet recognition model, training model parameters of the flow sheet recognition model according to the output result of the flow sheet recognition model based on the input second sample dialog text and a flow sheet reason label corresponding to the flow sheet reason keyword in the second sample dialog text until the loss of the flow sheet recognition model meets a preset condition, and obtaining the flow sheet recognition model comprising the trained model parameters.
6. The flow sheet cause analysis method according to claim 1, further comprising:
acquiring conversation voice between a target user and a service user;
and carrying out voice recognition processing on the conversation voice to obtain a conversation text corresponding to the conversation voice.
7. The method for analyzing the reason for the flow order according to claim 1, wherein the determining the reason for the flow order of the target user based on the flow order reason tag comprises:
counting the total number of each flow list reason label based on the flow list reason label corresponding to each dialog text;
selecting a target flow list reason label with the maximum corresponding total number from the flow list reason labels based on the total number of each flow list reason label;
and determining the reason of the flow list indicated by the target flow list reason label as the reason of the flow list of the target user.
8. The flow sheet cause analysis method according to claim 1, further comprising:
determining a solution corresponding to the reason of the flow order based on the reason of the flow order of the target user;
operating the target item based on the solution.
9. A flow order cause analysis apparatus, characterized in that the analysis apparatus comprises:
the first acquisition module is used for acquiring a dialog text between a target user and a service user when the target user is determined to be lost in a target project;
the first input module is used for inputting the conversation text into a pre-trained flow sheet recognition model to obtain a flow sheet reason label corresponding to the conversation text;
and the first determining module is used for determining the reason of the flow list of the target user based on the flow list reason label.
10. A computer device, comprising: a processor, a memory storing machine readable instructions executable by the processor, the processor to execute the machine readable instructions stored in the memory, the processor to perform the steps of the flow single cause analysis method of any one of claims 1 to 8 when the machine readable instructions are executed by the processor.
11. A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, which computer program, when executed by a computer device, performs the steps of the flow singles reason analysis method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010906344.4A CN111984764A (en) | 2020-09-01 | 2020-09-01 | Stream bill reason analysis method and device, computer equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010906344.4A CN111984764A (en) | 2020-09-01 | 2020-09-01 | Stream bill reason analysis method and device, computer equipment and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111984764A true CN111984764A (en) | 2020-11-24 |
Family
ID=73447294
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010906344.4A Withdrawn CN111984764A (en) | 2020-09-01 | 2020-09-01 | Stream bill reason analysis method and device, computer equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111984764A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113450153A (en) * | 2021-07-02 | 2021-09-28 | 京东科技控股股份有限公司 | Data processing method and device |
| CN114528849A (en) * | 2022-01-20 | 2022-05-24 | 浙江百应科技有限公司 | Method, system, apparatus, and medium for recognizing valid text based on dialog data |
-
2020
- 2020-09-01 CN CN202010906344.4A patent/CN111984764A/en not_active Withdrawn
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113450153A (en) * | 2021-07-02 | 2021-09-28 | 京东科技控股股份有限公司 | Data processing method and device |
| CN114528849A (en) * | 2022-01-20 | 2022-05-24 | 浙江百应科技有限公司 | Method, system, apparatus, and medium for recognizing valid text based on dialog data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104951428B (en) | User's intension recognizing method and device | |
| CN109446509B (en) | Dialogue corpus intention analysis method and system and electronic equipment | |
| CN113360622B (en) | User dialogue information processing method and device and computer equipment | |
| CN111260437A (en) | Product recommendation method based on commodity aspect level emotion mining and fuzzy decision | |
| US9875486B2 (en) | Extracting product purchase information from electronic messages | |
| CN111597320A (en) | Intention recognition device, method, equipment and storage medium based on hierarchical classification | |
| CN109461039A (en) | A kind of text handling method and intelligent customer service method | |
| CN110019703B (en) | Data marking method and device and intelligent question-answering method and system | |
| CN111324698B (en) | Deep learning method, evaluation viewpoint extraction method, device and system | |
| CN111984764A (en) | Stream bill reason analysis method and device, computer equipment and readable storage medium | |
| CN111143530A (en) | Intelligent answering method and device | |
| CN110955750A (en) | Combined identification method and device for comment area and emotion polarity, and electronic equipment | |
| CN112053703A (en) | Information processing method and device, computer equipment and storage medium | |
| CN112784039A (en) | Method, device and storage medium for distributing online customer service | |
| CN111429157A (en) | Method, device and equipment for evaluating and processing complaint work order and storage medium | |
| CN113051380A (en) | Information generation method and device, electronic equipment and storage medium | |
| US20160267425A1 (en) | Data processing techniques | |
| CN114817507A (en) | Reply recommendation method, device, equipment and storage medium based on intention recognition | |
| CN113822071A (en) | Background information recommendation method and device, electronic equipment and medium | |
| CN113609865A (en) | Text emotion recognition method and device, electronic equipment and readable storage medium | |
| CN108073567A (en) | A kind of Feature Words extraction process method, system and server | |
| CN111787042A (en) | Method and device for pushing information | |
| CN114297380A (en) | Data processing method, device, equipment and storage medium | |
| CN114003811A (en) | Method and device for recommending business knowledge information, electronic equipment and storage medium | |
| JP7216627B2 (en) | INPUT SUPPORT METHOD, INPUT SUPPORT SYSTEM, AND PROGRAM |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20201124 |