CN111782793A - Intelligent customer service processing method, system and equipment - Google Patents
Intelligent customer service processing method, system and equipment Download PDFInfo
- Publication number
- CN111782793A CN111782793A CN202010798753.7A CN202010798753A CN111782793A CN 111782793 A CN111782793 A CN 111782793A CN 202010798753 A CN202010798753 A CN 202010798753A CN 111782793 A CN111782793 A CN 111782793A
- Authority
- CN
- China
- Prior art keywords
- customer service
- classification model
- text
- preprocessing
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- 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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/27—Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
-
- 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
- G06F40/35—Discourse or dialogue representation
Abstract
The invention discloses an intelligent customer service processing method, system and equipment. The method comprises the following steps: acquiring multi-turn conversation contents in the intelligent customer service system, performing model training according to the multi-turn conversation contents, and constructing a multi-class label classification model of the context, wherein each label class corresponds to a reply item; and sending the request content input by the user into a multi-class label classification model for classification, and outputting corresponding reply items according to the classified labels. The scheme of the invention has higher flexibility and can process more forms of user requests; compared with a customer service system based on a retrieval mode, the classification model is used for replacing the retrieval mode, so that the robustness of the whole system is improved; the semantic gap problem of the original retrieval system can be solved; the influence of noise on the system selection reply item can be reduced; the accuracy is higher.
Description
Technical Field
The invention relates to the technical field of computers, in particular to an intelligent customer service processing method, system and equipment.
Background
The dialog System (Spoken dialog System) can be regarded as an interactive System simulating human-human communication through Natural Language, and has been widely applied to customer service systems as the technologies of Automatic Speech Recognition (ASR) and Natural Language Processing (NLP) become mature. In 2006, a deep learning method is widely researched and applied, a plurality of machine learning tasks are developed in a breakthrough manner, and in 2013, word2vec, a language model and a sequence-to-sequence model appear in succession, and massive data are collected and applied to training and designing of a system.
In a traffic card service scene, the most important service is oriented to business consultation such as card handling and activation, and the traffic card service scene belongs to an intelligent service system in the vertical field. According to the input request of the user, the reply item is retrieved, and the traditional method is to use a series of rules defined manually in advance to carry out semantic analysis and match the reply item. The conventional method is based on a search-type model, the main idea of the search-type dialogue system is to match a most suitable reply item in a plurality of reply candidate items as an output according to an input request of a user, and the matching modes are two. The first one is a matching mode based on representation, text feature extraction is respectively carried out on input and reply candidate items in an initial stage, then similarity calculation is carried out on the obtained text representation through a similarity function, and the highest matching layer degree is used as output. The second mode is a matching mode based on interaction, which is different from the mode in which the similarity is calculated for the text representation only at the last stage, the matching mode based on interaction is used for interacting the text representation at the front stage of the model, through similarity matrix calculation, the model can obtain the matching relationship of different levels and different granularities, and finally the reply item with the highest matching layer degree is returned.
The prior art has the problems that the method based on the manual definition rule is not flexible enough, and various and changeable user requests cannot be well solved; the core of the method is that the input request information of a user and the reply item are subjected to correlation Matching (Relevance Matching), however, the information-reply sometimes have no direct relation on the word, the problem of semantic gap is overcome, in addition, because the information-reply are matched, the quality of two text messages is very important, if irrelevant contents appear in the text, the noise easily affects a dialogue system based on retrieval when the reply item is matched, and the robustness is low.
Disclosure of Invention
The invention aims to provide an intelligent customer service processing method, an intelligent customer service processing system and intelligent customer service processing equipment, which are used for overcoming the defects in the prior art.
In order to achieve the purpose, the invention adopts the following technical scheme.
In a first aspect, an intelligent customer service processing method is provided, including: acquiring multi-turn conversation contents in the intelligent customer service system, performing model training according to the multi-turn conversation contents, and constructing a multi-class label classification model of the context, wherein each label class corresponds to a reply item; and sending the request content input by the user into a multi-class label classification model for classification, and outputting corresponding reply items according to the classified labels.
In some possible implementations, the model training according to multiple rounds of dialog contents to construct the multi-class label classification model of the above context includes: acquiring multi-turn conversation contents from an intelligent customer service system; associating the acquired multi-turn conversation content with the content, and labeling the associated text data, wherein the labeled categories comprise chatting, business and complaints, the business categories are further labeled in a category manner, and each category of the business corresponds to one reply item; preprocessing the labeled text data, wherein the preprocessing comprises data cleaning and word segmentation processing, and dividing the preprocessed text data into a training set and a test set; extracting text features of the text data of the training set; and training the classification model by adopting the extracted text features to obtain the multi-class label classification model of the context.
In some possible implementation manners, the sending the request content input by the user into the multi-class tag classification model for classification, and outputting the corresponding reply item according to the classified tag includes: receiving request content currently input by a user; associating the request content currently input by the user with the content to generate input data; preprocessing the generated input data, wherein the preprocessing comprises data cleaning and word segmentation; performing text feature extraction on the preprocessed input data; and sending the extracted text features into a multi-class label classification model for classification, and outputting corresponding reply items according to the classified labels.
In a second aspect, an intelligent customer service system is provided, comprising: the classification module is used for acquiring multi-turn conversation contents in the intelligent customer service system, performing model training according to the multi-turn conversation contents, and constructing a multi-class label classification model of the context, wherein each label class corresponds to a reply item; and the reply module is used for sending the request content input by the user into the multi-class label classification model for classification and outputting a corresponding reply item according to the classified label.
In some possible implementations, the classification module includes: the acquisition unit is used for acquiring multi-turn conversation contents from the intelligent customer service system and storing the multi-turn conversation contents into the database; the labeling unit is used for associating the multi-turn conversation content with the above content and labeling the associated text data, wherein the labeled categories comprise chatting, business and complaints, the business categories are further labeled in categories, each category of the business corresponds to one reply item, and the labeled text data is stored in a database; the system comprises a preprocessing unit, a training set and a testing set, wherein the preprocessing unit is used for preprocessing the labeled text data, the preprocessing comprises data cleaning and word segmentation processing, and the preprocessed text data is divided into the training set and the testing set; the text feature extraction unit is used for extracting text features of the text data of the training set; and the model construction unit is used for training the classification model by adopting the extracted text features to obtain the multi-class label classification model of the context.
In some possible implementations, the reply module includes: the receiving unit is used for receiving the request content currently input by the user; the association unit is used for associating the request problem currently input by the user with the content to generate input data; the preprocessing unit is used for preprocessing the generated input data, and the preprocessing comprises data cleaning and word segmentation; the text feature extraction unit is used for extracting text features of the preprocessed input data; and the output unit is used for sending the extracted text features into the multi-class label classification model for classification and outputting corresponding reply items according to the classified labels.
In a third aspect, there is provided a computer device comprising a processor and a memory, the memory having a program stored therein, the program comprising computer executable instructions, the processor executing the computer executable instructions stored in the memory when the computer device is running, to cause the computer device to perform the intelligent customer service processing method according to the first aspect.
In a fourth aspect, there is provided a computer readable storage medium storing one or more programs, the one or more programs comprising computer executable instructions, which when executed by a computer device, cause the computer device to perform the intelligent customer service method of the first aspect.
According to the technical scheme, the embodiment of the invention has the following advantages:
firstly, compared with the traditional manual rule method, the method has higher flexibility and can process more forms of user requests;
compared with a customer service system based on a retrieval mode, the method of replacing the retrieval mode with the classification module is beneficial to improving the robustness of the whole system;
thirdly, because the text content of the reply item is not required to be matched, the method can solve the semantic gap problem faced by the original retrieval system;
thirdly, the invention can reduce the influence of noise on the system selection reply item;
and thirdly, compared with the existing intelligent customer service system, the system and the method fully utilize the above dialogue content information to reply the current user request in a mode of being associated with the above, so that the accuracy is higher.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the embodiments and the drawings used in the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of an intelligent customer service processing method according to an embodiment of the present invention;
FIG. 2 is a block diagram of an intelligent customer service system provided by an embodiment of the present invention;
FIG. 3 is a flow chart of constructing a multi-label classification model in an embodiment of the invention;
FIG. 4 is a flowchart of text annotation according to an embodiment of the present invention;
FIG. 5 is a flow chart of a reply to a user portion in an embodiment of the present invention;
fig. 6 is a block diagram of a computer device according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The terms "first," "second," "third," and the like in the description and in the claims, and in the above-described drawings, are used for distinguishing between different objects and not for describing a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
The following are detailed descriptions of the respective embodiments.
The embodiment of the invention provides an intelligent customer service system and a corresponding processing method. The system is an end-to-end intelligent customer service system based on a classification method, and mainly comprises two parts, wherein one part is a construction part of a multi-label classification model, and the other part is a reply user part. The construction part of the multi-label classification model mainly obtains multi-round conversation contents in the intelligent customer service system, performs text related preprocessing, and then models the processed multi-round conversation contents into the multi-class label classification model of the context, wherein each class corresponds to a specific reply item. The reply user part processes the request content input by the user, classifies the request content of the user and outputs a corresponding reply item.
Referring to fig. 1, an embodiment of the present invention provides an intelligent customer service processing method, which includes the following steps:
s1, obtaining multi-round conversation contents in the intelligent customer service system, carrying out model training according to the multi-round conversation contents, and constructing a multi-class label classification model of the context, wherein each label class corresponds to a reply item;
and S2, sending the request content input by the user into a multi-class label classification model for classification, and outputting corresponding reply items according to the classified labels.
Further, the performing model training according to multiple rounds of dialog contents in step S1 to construct the multi-class label classification model of the above context may include: acquiring multi-turn conversation contents from an intelligent customer service system; associating the acquired multi-turn conversation content with the content, and labeling the associated text data, wherein the labeled categories comprise chatting, business and complaints, the business categories are further labeled in a category manner, and each category of the business corresponds to one reply item; preprocessing the labeled text data, wherein the preprocessing comprises data cleaning and word segmentation processing, and dividing the preprocessed text data into a training set and a test set; extracting text features of the text data of the training set; and training the classification model by adopting the extracted text features, and constructing to obtain the multi-class label classification model of the context.
Further, step S2 may include: receiving request content currently input by a user; associating the request content currently input by the user with the content to generate input data; preprocessing the generated input data, wherein the preprocessing comprises data cleaning and word segmentation; performing text feature extraction on the preprocessed input data; and sending the extracted text features into a multi-class label classification model for classification, and outputting corresponding reply items according to the classified labels.
Referring to fig. 2, another embodiment of the present invention provides an intelligent customer service system, which includes:
the classification module 10 is configured to obtain multiple rounds of conversation contents in the intelligent customer service system, perform model training according to the multiple rounds of conversation contents, and construct a multi-class label classification model of the context, where each label class corresponds to one reply item.
And the reply module 20 is configured to send the request content input by the user into the multi-class label classification model for classification, and output a corresponding reply item according to the classified label.
Further, the classification module 10 may include:
the acquisition unit is used for acquiring multi-turn conversation contents from the intelligent customer service system and storing the multi-turn conversation contents into the database;
the labeling unit is used for associating the multi-turn conversation content with the above content and labeling the associated text data, wherein the labeled categories comprise chatting, business and complaints, the business categories are further labeled in categories, each category of the business corresponds to one reply item, and the labeled text data is stored in a database;
the system comprises a preprocessing unit, a training set and a testing set, wherein the preprocessing unit is used for preprocessing the labeled text data, the preprocessing comprises data cleaning and word segmentation processing, and the preprocessed text data is divided into the training set and the testing set;
the text feature extraction unit is used for extracting text features of the text data of the training set;
and the model construction unit is used for training the classification model by adopting the extracted text features to construct and obtain the multi-class label classification model of the context.
Further, the reply module 20 may include:
the receiving unit is used for receiving the request content currently input by the user;
the association unit is used for associating the request problem currently input by the user with the content to generate input data;
the preprocessing unit is used for preprocessing the generated input data, and the preprocessing comprises data cleaning and word segmentation;
the text feature extraction unit is used for extracting text features of the preprocessed input data;
and the output unit is used for sending the extracted text features into the multi-class label classification model for classification and outputting corresponding reply items according to the classified labels.
Next, the multi-label classification model construction part and the reply user part in the technical solution of the embodiment of the present invention are respectively compared to further explain.
Construction of a multi-label classification model
Referring to fig. 3, the construction of the multi-label classification model includes the following steps:
1.1 obtaining multiple rounds of dialog text data
A) Acquiring online real conversation data from an intelligent customer service system, for example, a mass transit intelligent customer service system facing a card scene of an ETC (Electronic Toll Collection) card scene;
B) storing the acquired data in a database, for example: mongo DB. MongoDB is a database based on distributed file storage.
1.2 text information labeling
Referring to fig. 4, the process of text annotation is as follows.
A) Firstly, labeling text data information of multi-turn conversation content:
a.1) importing multi-turn conversation contents to be labeled into a data labeling system;
a.2) the data annotation personnel associate the current conversation content annotation text with the above content;
a.3) carrying out text annotation according to conversation contents, wherein the main categories comprise chatting, complaints and services;
a.4) class marking can be further carried out in the service classes, and each class corresponds to a reply item;
a.5) submitting the annotation data.
B) And (3) checking the marked data:
b.1) examiners examine the annotated data;
b.2) whether the marked data is rejected, if so, marking personnel mark the data again and review again;
and B.3) warehousing the approved data for model increment training.
C) And (5) all the marked data are put into a warehouse, and the construction of the traffic card field data set is completed.
1.3 text information preprocessing
A) And (4) text data cleaning, namely removing special characters in the text data in a regular mode by using an existing tool, such as a re module in python.
B) Text word segmentation:
b.1) aiming at the preprocessing of Chinese texts, because of the structural relationship of Chinese, the characteristic granularity is that the granularity of words is far better than the granularity of words, so word segmentation processing needs to be carried out on Chinese;
b.2) the word segmentation method mainly comprises a word segmentation method based on character string matching, a word segmentation method based on statistics and a word segmentation method based on understanding, and the word segmentation processing can be carried out on the Chinese text by utilizing the existing tools, such as the ending word segmentation;
C) the preprocessed text data is proportionally divided into a training set Tr _ set and a test set Te _ set.
1.4 construction of Classification Module
A) Text feature extraction:
a.1) text characteristic engineering mainly extracts characteristic representation capable of embodying text characteristics according to a database;
a.2) inputting a training set Tr _ set of multi-turn dialogue contents subjected to word segmentation, and performing text feature extraction by using the existing engineering technology such as TF-IDF;
a.3) TF-IDF (term frequency-inverse document frequency) is a commonly used weighting technique for information retrieval and data mining. TF is Term Frequency (Term Frequency) and IDF is Inverse text Frequency index (Inverse Document Frequency).
Word frequency (TF) is the number of times a word occurs divided by the total number of words in the document. If the total number of words in a document is 100 and the word "activate" occurs 3 times, then the word frequency for "activate" in the document is 0.03 (3/100). One way to calculate the Document Frequency (DF) is to determine how many documents have been "activated" and then divide by the total number of documents contained in the document set. Therefore, if the term "activate" occurs over 1,000 documents and the total number of documents is 10,000,000, the document frequency is 0.0001(1000/10,000,000). Finally, the TF-IDF score can be derived by calculating the word frequency x (the logarithm of the reciprocal of the document frequency). In the above example, the term "activate" would have a TF-IDF score of 0.12 ═ 0.03xlog (1/0.0001) in the document set.
B) Classification model
B.1) the classification model adopted in the text belongs to a supervised learning model and is input into a multi-turn conversation feature representation q and a corresponding service label l which are associated with the text; here, the multi-turn dialog feature representation q includes feature representations of the current turn and the previous turns of the dialog;
b.2) the classification algorithm adopts the existing engineering technology with supervised learning, such as a support vector machine, and represents a training model according to the input text characteristics;
b.3) calculating confidence of each service label;
b.4) the classification model outputs a service label;
C) and exporting and storing the multi-class label classification model, and recording the multi-class label classification model as CM.
(II) replying to the user
Referring to fig. 5, the replying to the user part includes the following steps:
2.1 user input request
Request content q input by useriUsually, the question contains a series of keywords, such as "abnormal card activation", "bluetooth connection impossible", etc.
2.2 associating the request entered by the user with the content above
Because the conversation content in the intelligent customer service often belongs to a multi-turn question-answering context, the current turn of user request content q is addediAnd the contents of { q } abovei-n,ri-n,…,qi-2,ri-2,qi-1,ri-1Get associated, where q is the request r in reply, resulting in the input data q associated with abovei ’The method can effectively help reply to the user request.
Requesting content q for the ith round of usersiNot only need to input q when classifyingiAlso input n turns of dialogue data before the ith turn, i.e. { q }i-n,ri-n,…,qi-2,ri-2,qi-1,ri-1}. That is, the input data q herei ’Not only include qiAnd also includes { qi-n,ri-n,…,qi-2,ri-2,qi-1,ri-1}. Where n is an empirical value, and may be determined according to specific needs, for example, may be 3 or 4, and is not limited.
2.3 request information q associated with the abovei ’Pretreatment, specifically refer to step 1.3 above;
2.4 feature extraction qi ’Specifically, please refer to the text feature extraction in step 1.4 above;
2.5 into the classification model CM, output according to the modelThe service label l corresponds to the reply item ri；
2.6 output reply term riTo the user.
Referring to fig. 6, an embodiment of the present invention further provides a computer device 60, which includes a processor 61 and a memory 62, where the memory 62 stores a program, and the program includes computer execution instructions, and when the computer device 60 runs, the processor 61 executes the computer execution instructions stored in the memory, so as to make the computer device 60 execute the intelligent customer service processing method as described above.
An embodiment of the present invention also provides a computer readable storage medium storing one or more programs, the one or more programs comprising computer executable instructions, which when executed by a computer device, cause the computer device to perform the intelligent customer service processing method as described above.
To sum up, the embodiment of the invention discloses an intelligent customer service processing method, an intelligent customer service processing system and related equipment, and the embodiment of the invention has the following advantages that:
firstly, compared with the traditional manual rule method, the method has higher flexibility and can process more forms of user requests;
compared with a customer service system based on a retrieval mode, the method of replacing the retrieval mode with the classification module is beneficial to improving the robustness of the whole system;
thirdly, because the text content of the reply item is not required to be matched, the method can solve the semantic gap problem faced by the original retrieval system;
thirdly, the invention can reduce the influence of noise on the system selection reply item;
and thirdly, compared with the existing intelligent customer service system, the system and the method fully utilize the above dialogue content information to reply the current user request in a mode of being associated with the above, so that the accuracy is higher.
In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to the related descriptions of other embodiments.
The above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; those of ordinary skill in the art will understand that: the technical solutions described in the above embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (8)
1. An intelligent customer service processing method is characterized by comprising the following steps:
acquiring multi-turn conversation contents in the intelligent customer service system, performing model training according to the multi-turn conversation contents, and constructing a multi-class label classification model of the context, wherein each label class corresponds to a reply item;
and sending the request content input by the user into a multi-class label classification model for classification, and outputting corresponding reply items according to the classified labels.
2. The method of claim 1, wherein the model training from multiple rounds of dialog content to construct a multi-class label classification model of the context above comprises:
acquiring multi-turn conversation contents from an intelligent customer service system;
associating the acquired multi-turn conversation content with the content, and labeling the associated text data, wherein the labeled categories comprise chatting, business and complaints, the business categories are further labeled in a category manner, and each category of the business corresponds to one reply item;
preprocessing the labeled text data, wherein the preprocessing comprises data cleaning and word segmentation processing, and dividing the preprocessed text data into a training set and a test set;
extracting text features of the text data of the training set;
and training the classification model by adopting the extracted text features to obtain the multi-class label classification model of the context.
3. The method of claim 1, wherein the step of inputting the request content input by the user into a multi-category tag classification model for classification and outputting the corresponding reply item according to the classified tag comprises:
receiving request content currently input by a user;
associating the request content currently input by the user with the content to generate input data;
preprocessing the generated input data, wherein the preprocessing comprises data cleaning and word segmentation;
performing text feature extraction on the preprocessed input data;
and sending the extracted text features into a multi-class label classification model for classification, and outputting corresponding reply items according to the classified labels.
4. An intelligent customer service system, comprising:
the classification module is used for acquiring multi-turn conversation contents in the intelligent customer service system, performing model training according to the multi-turn conversation contents, and constructing a multi-class label classification model of the context, wherein each label class corresponds to a reply item;
and the reply module is used for sending the request content input by the user into the multi-class label classification model for classification and outputting a corresponding reply item according to the classified label.
5. The system of claim 4, wherein the classification module comprises:
the acquisition unit is used for acquiring multi-turn conversation contents from the intelligent customer service system and storing the multi-turn conversation contents into the database;
the labeling unit is used for associating the multi-turn conversation content with the above content and labeling the associated text data, wherein the labeled categories comprise chatting, business and complaints, the business categories are further labeled in categories, each category of the business corresponds to one reply item, and the labeled text data is stored in a database;
the system comprises a preprocessing unit, a training set and a testing set, wherein the preprocessing unit is used for preprocessing the labeled text data, the preprocessing comprises data cleaning and word segmentation processing, and the preprocessed text data is divided into the training set and the testing set;
the text feature extraction unit is used for extracting text features of the text data of the training set;
and the model construction unit is used for training the classification model by adopting the extracted text features to obtain the multi-class label classification model of the context.
6. The system of claim 4, wherein the reply module comprises:
the receiving unit is used for receiving the request content currently input by the user;
the association unit is used for associating the request problem currently input by the user with the content to generate input data;
the preprocessing unit is used for preprocessing the generated input data, and the preprocessing comprises data cleaning and word segmentation;
the text feature extraction unit is used for extracting text features of the preprocessed input data;
and the output unit is used for sending the extracted text features into the multi-class label classification model for classification and outputting corresponding reply items according to the classified labels.
7. A computer device comprising a processor and a memory, the memory having stored therein a program comprising computer-executable instructions that, when executed by the computer device, the processor executes the computer-executable instructions stored by the memory to cause the computer device to perform the intelligent customer service processing method of any one of claims 1-3.
8. A computer readable storage medium storing one or more programs, the one or more programs comprising computer executable instructions, which when executed by a computer device, cause the computer device to perform the intelligent customer service method of any of claims 1-3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010798753.7A CN111782793A (en) | 2020-08-11 | 2020-08-11 | Intelligent customer service processing method, system and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010798753.7A CN111782793A (en) | 2020-08-11 | 2020-08-11 | Intelligent customer service processing method, system and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111782793A true CN111782793A (en) | 2020-10-16 |
Family
ID=72761949
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010798753.7A Pending CN111782793A (en) | 2020-08-11 | 2020-08-11 | Intelligent customer service processing method, system and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111782793A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112328871A (en) * | 2020-10-27 | 2021-02-05 | 深圳集智数字科技有限公司 | Reply generation method, device, equipment and storage medium based on RPA module |
| CN112883183A (en) * | 2021-03-22 | 2021-06-01 | 北京大学深圳研究院 | Method for constructing multi-classification model, intelligent customer service method, and related device and system |
| CN113282755A (en) * | 2021-06-11 | 2021-08-20 | 上海寻梦信息技术有限公司 | Dialogue type text classification method, system, equipment and storage medium |
| CN117059082A (en) * | 2023-10-13 | 2023-11-14 | 北京水滴科技集团有限公司 | Outbound call conversation method, device, medium and computer equipment based on large model |
| CN112328871B (en) * | 2020-10-27 | 2024-04-26 | 深圳集智数字科技有限公司 | Reply generation method, device, equipment and storage medium based on RPA module |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109543030A (en) * | 2018-10-12 | 2019-03-29 | 平安科技（深圳）有限公司 | Customer service machine conference file classification method and device, equipment, storage medium |
| CN110059182A (en) * | 2019-03-21 | 2019-07-26 | 阿里巴巴集团控股有限公司 | Art recommended method and device towards customer service |
| CN111177359A (en) * | 2020-04-10 | 2020-05-19 | 支付宝(杭州)信息技术有限公司 | Multi-turn dialogue method and device |
-
2020
- 2020-08-11 CN CN202010798753.7A patent/CN111782793A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109543030A (en) * | 2018-10-12 | 2019-03-29 | 平安科技（深圳）有限公司 | Customer service machine conference file classification method and device, equipment, storage medium |
| CN110059182A (en) * | 2019-03-21 | 2019-07-26 | 阿里巴巴集团控股有限公司 | Art recommended method and device towards customer service |
| CN111177359A (en) * | 2020-04-10 | 2020-05-19 | 支付宝(杭州)信息技术有限公司 | Multi-turn dialogue method and device |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112328871A (en) * | 2020-10-27 | 2021-02-05 | 深圳集智数字科技有限公司 | Reply generation method, device, equipment and storage medium based on RPA module |
| CN112328871B (en) * | 2020-10-27 | 2024-04-26 | 深圳集智数字科技有限公司 | Reply generation method, device, equipment and storage medium based on RPA module |
| CN112883183A (en) * | 2021-03-22 | 2021-06-01 | 北京大学深圳研究院 | Method for constructing multi-classification model, intelligent customer service method, and related device and system |
| CN113282755A (en) * | 2021-06-11 | 2021-08-20 | 上海寻梦信息技术有限公司 | Dialogue type text classification method, system, equipment and storage medium |
| CN117059082A (en) * | 2023-10-13 | 2023-11-14 | 北京水滴科技集团有限公司 | Outbound call conversation method, device, medium and computer equipment based on large model |
| CN117059082B (en) * | 2023-10-13 | 2023-12-29 | 北京水滴科技集团有限公司 | Outbound call conversation method, device, medium and computer equipment based on large model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11334635B2 (en) | Domain specific natural language understanding of customer intent in self-help | |
| CN110727779A (en) | Question-answering method and system based on multi-model fusion | |
| CN109872162B (en) | Wind control classification and identification method and system for processing user complaint information | |
| CN110580308B (en) | Information auditing method and device, electronic equipment and storage medium | |
| CN111782793A (en) | Intelligent customer service processing method, system and equipment | |
| CN111783394A (en) | Training method of event extraction model, event extraction method, system and equipment | |
| CN112395421B (en) | Course label generation method and device, computer equipment and medium | |
| CN109582788A (en) | Comment spam training, recognition methods, device, equipment and readable storage medium storing program for executing | |
| CN112069312A (en) | Text classification method based on entity recognition and electronic device | |
| CN113051380A (en) | Information generation method and device, electronic equipment and storage medium | |
| CN111462752A (en) | Client intention identification method based on attention mechanism, feature embedding and BI-L STM | |
| CN109446522B (en) | Automatic test question classification system and method | |
| CN113722492A (en) | Intention identification method and device | |
| CN113220999A (en) | User feature generation method and device, electronic equipment and storage medium | |
| CN112883183B (en) | Method for constructing multi-classification model, intelligent customer service method, and related device and system | |
| CN115017385A (en) | Article searching method, device, equipment and storage medium | |
| CN114067343A (en) | Data set construction method, model training method and corresponding device | |
| CN114186041A (en) | Answer output method | |
| CN109597879B (en) | Service behavior relation extraction method and device based on 'citation relation' data | |
| CN113095078A (en) | Associated asset determination method and device and electronic equipment | |
| CN112579666A (en) | Intelligent question-answering system and method and related equipment | |
| CN110717029A (en) | Information processing method and system | |
| CN115618968B (en) | New idea discovery method and device, electronic device and storage medium | |
| CN114417001B (en) | Chinese writing intelligent analysis method, system and medium based on multi-mode | |
| CN112308453B (en) | Risk identification model training method, user risk identification method and related devices |
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 |