CN112287639A - Intelligent customer service work order classification method - Google Patents
Intelligent customer service work order classification method Download PDFInfo
- Publication number
- CN112287639A CN112287639A CN202011194254.3A CN202011194254A CN112287639A CN 112287639 A CN112287639 A CN 112287639A CN 202011194254 A CN202011194254 A CN 202011194254A CN 112287639 A CN112287639 A CN 112287639A
- Authority
- CN
- China
- Prior art keywords
- text
- processed
- words
- word
- processing
- 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
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000012545 processing Methods 0.000 claims abstract description 43
- 238000012549 training Methods 0.000 claims abstract description 40
- 238000004140 cleaning Methods 0.000 claims abstract description 23
- 238000012360 testing method Methods 0.000 claims abstract description 20
- 238000012795 verification Methods 0.000 claims abstract description 17
- 238000002372 labelling Methods 0.000 claims abstract description 14
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000010200 validation analysis Methods 0.000 claims 2
- 238000010586 diagram Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000006872 improvement Effects 0.000 description 4
- 239000000463 material Substances 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000001149 cognitive effect Effects 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 241000238558 Eucarida Species 0.000 description 1
- 244000062793 Sorghum vulgare Species 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 235000019713 millet Nutrition 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000036651 mood Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000007711 solidification Methods 0.000 description 1
- 230000008023 solidification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/103—Formatting, i.e. changing of presentation of documents
- G06F40/117—Tagging; Marking up; Designating a block; Setting of attributes
-
- 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
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- 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
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Human Computer Interaction (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Machine Translation (AREA)
Abstract
The invention relates to the technical field of logistics, in particular to an intelligent customer service work order classification method, which comprises the following steps: acquiring voice conversations between a customer service and a customer, and converting the voice conversations into texts; performing data cleaning, word processing and word processing on the text to obtain a text to be processed; labeling the text type of the text to be processed; dividing the marked text into a training set, a verification set and a test set, and respectively inputting the training set, the verification set and the test set into a preset training model for training to obtain a work order classification result. The work order type is established for the unsolved problem of the intelligent robot, so that the corresponding work order flows to professional personnel for solving, the step of manual work order classification is omitted, the administrative efficiency of an enterprise is greatly improved, the established work order can be rapidly solved, and the timeliness of the problem solving of the enterprise is obviously improved.
Description
Technical Field
The invention belongs to the technical field of logistics, and particularly relates to an intelligent customer service work order classification method.
Background
Early text classification mainly represents texts by establishing a word vector dictionary, and then classifies the texts by adopting a traditional machine learning method, such as SVM and decision tree, and the model trained by the algorithm has poor flexibility and undesirable effect; the word vector dictionary is fixed and does not dynamically change with the increase of the text, and in addition, the context semantic relation of the text is not strong, so that the model effect is poor. The problem that the context semantics of the text are not strongly connected is gradually developed, but the model training and the online service time are too long to be suitable for large-scale commercial application. Aiming at the problem of overlong model response time, because the one-dimensional convolution network has higher convolution speed, models such as TextCNN are gradually formed to solve the NLP classification problem. However, the problem of word dropping vector solidification is not really solved on the basis of the recurrent neural network model and the convolutional neural network model, so that the upper limit of the performance of the trained model is limited, and the performance of the model is greatly reduced in practical application because various trigks only approach the upper limit infinitely.
With the development of NLP technology, the intelligent customer service has more and more prominent important role in customer service, wherein the famous 'Xiaodu' robot with hundreds degrees and 'Xiaoaitong classmate' of millet are in China; the siri function of the apple products is counted abroad; at present, NLP technology is more and more widely served in the field of voice customer service, the intelligent level is gradually improved, normal voice communication with natural people can be realized, and questions proposed by customers can be answered. But the problem beyond the cognitive scope of the intelligent robot needs to establish a work order to wait for later manual solution, which wastes manpower and material resources and reduces the working efficiency of enterprises.
Disclosure of Invention
In view of the above, the present invention is to overcome the defects in the prior art, and provide an intelligent customer service work order classification method to solve the problems in the prior art.
In order to achieve the purpose, the invention adopts the following technical scheme:
an intelligent customer service work order classification method comprises the following steps:
acquiring voice conversations between a customer service and a customer, and converting the voice conversations into texts;
performing data cleaning, word processing and word processing on the text to obtain a text to be processed;
labeling the text type of the text to be processed;
dividing the marked text into a training set, a verification set and a test set, and respectively inputting the training set, the verification set and the test set into a preset training model for training to obtain a work order classification result.
Further, the labeling of the text category of the text to be processed includes,
and matching a preset text type with the text to be processed by adopting a regular expression, and labeling the text type according to a matching result.
Further, the preset text categories include: the courier leaves the job, the network node is closed, the network node is exploded, the network node has poor service attitude and the network node is signed for non-receipt.
Further, performing data cleaning, word segmentation processing and word vector processing on the text to obtain a text to be processed, including:
cleaning data of the text, and removing stop words, tone words and useless contents in the text, wherein the useless contents comprise greetings;
performing word processing and word processing on the text subjected to the data cleaning processing to obtain word vector representation and word vector representation;
and embedding the word vector and the expression of the word vector to obtain the text to be processed.
Further, the method also comprises the step of carrying out secondary cleaning on the text to be processed, and the secondary cleaning comprises the following steps:
judging whether the word vector in the text to be processed is the same as the stop table according to the regular expression to obtain a judgment result;
if the judgment results are the same, removing the same word vector from the text;
and if the judgment results are the same, reserving the word vector.
Further, before dividing the labeled text into a training set, a verification set and a test set, and inputting the training set, the verification set and the test set into a preset training model for training, the method further includes:
processing the text length of the text to be processed to obtain a processed text with the text length of 512 words;
and carrying out word frequency statistics on the processed text and carrying out position coding on each word in the processed text to obtain the relation between semantic contexts of the processed text.
Further, processing the text length of the text to be processed includes: and/or processing the text with the length smaller than 512 words.
Further, the processing the text with the length exceeding 512 words comprises:
the reserved head 512 words are adopted for 15 percent of texts in the form of random numbers;
tail 512 words are reserved for 15% of texts;
the middle 512 words are reserved for 30% of the text;
the head 218 words and the tail 294 words are reserved for the remaining text, respectively.
Further, the processing the text with the length smaller than 512 words includes:
and zero padding is carried out on the tail part of the text line with the length less than 512 words of the text to be processed until 512 words are reached.
Further, performing word frequency statistics on the processed text comprises: word group statistics and single word statistics.
The technical scheme of the invention has the following beneficial effects:
in the technical scheme provided by the invention, the voice conversation is converted into a text by acquiring the voice conversation between the customer service and the customer; performing data cleaning, word processing and word processing on the text to obtain a text to be processed; labeling text categories of the text to be processed; the method comprises the steps of dividing a marked text into a training set, a verification set and a test set, inputting the training set, the verification set and the test set into a preset training model respectively for training, obtaining work order classification results, intelligently establishing work order types for unsolved problems of the intelligent robot, enabling corresponding work orders to flow into professional personnel for solving, saving the step of manual work order classification, greatly improving the administrative efficiency of enterprises, enabling the established work orders to be rapidly solved, and enabling the timeliness of the problem solving of the enterprises to be obviously improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of steps of an intelligent customer service work order classification method according to an embodiment of the present invention.
Fig. 2 is a flowchart of steps of an intelligent customer service work order classification method according to another embodiment of the present invention.
Fig. 3 is a flowchart of steps of an intelligent customer service work order classification method according to another embodiment of the present invention.
Fig. 4 is a flowchart of steps of an intelligent customer service work order classification method according to another embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the examples given herein without any inventive step, are within the scope of the present invention.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
The main defects of the prior art are as follows: the problem exceeding the cognitive range of the intelligent robot needs to be solved by establishing a work order and waiting for later manual work, which wastes manpower and material resources and reduces the working efficiency of enterprises
As shown in fig. 1, a method step diagram provided for one embodiment of the present invention includes:
s101, acquiring a voice conversation between a customer service and a client, and converting the voice conversation into a text;
a voice-to-character interface is adopted to convert voice of the customer service voice conversation from voice to text, so that later text analysis is facilitated; it should be noted that the speech-text interface is a common technique for those skilled in the art, and will not be described in detail herein.
Step S102, performing data cleaning, word processing and word processing on the text to obtain a text to be processed;
the text is cleaned, the text is mainly compared with the stop word list, and words in the stop word list are removed from the text, so that the purpose of cleaning the text is achieved.
Step S103, labeling text types of the text to be processed;
and S104, dividing the marked text into a training set, a verification set and a test set, and respectively inputting the training set, the verification set and the test set into a preset training model for training to obtain a work order classification result.
As an optional implementation test, dividing the labeled text into a training set, a verification set and a test set, wherein the ratio is 7: 2: 1; and then, carrying out Fine adjustment by adopting a BERT pre-training model, wherein when a new deep learning model is trained on a data set of the model by a Fine-tune principle, the Fine adjustment is generally carried out on the pre-trained model, so that a model based on customer service work order classification is trained, and a work order classification result is obtained.
It can be understood that in the technical scheme provided by the invention, the voice conversation is converted into the text by acquiring the voice conversation between the customer service and the customer; performing data cleaning, word processing and word processing on the text to obtain a text to be processed; labeling text categories of the text to be processed; the method comprises the steps of dividing a marked text into a training set, a verification set and a test set, inputting the training set, the verification set and the test set into a preset training model respectively for training, obtaining work order classification results, intelligently establishing work order types for unsolved problems of the intelligent robot, enabling corresponding work orders to flow into professional personnel for solving, saving the step of manual work order classification, greatly improving the administrative efficiency of enterprises, enabling the established work orders to be rapidly solved, and enabling the timeliness of the problem solving of the enterprises to be obviously improved.
As a further improvement of the above method, labeling the text type of the text to be processed includes,
and matching the preset text type with the text to be processed by adopting a regular expression, and labeling the text type according to the matching result.
Specifically, in one embodiment, for a specific scene of express delivery, the preset text categories mainly include five categories: the courier leaves the job, the network node is closed, the network node is exploded, the network node has poor service attitude and the network node is signed for non-receipt. And matching the texts in the five categories by adopting a regular expression to realize the labeling of the text categories.
As a further improvement of the above method, as shown in fig. 2, it is a step diagram of an intelligent customer service work order classification method provided by another embodiment of the present invention in fig. 2.
The method for processing the text by the word segmentation comprises the following steps of carrying out data cleaning, word segmentation processing and word vector processing on the text to obtain the text to be processed:
step S201, cleaning data of the text, and removing stop words, tone words and useless contents in the text, wherein the useless contents comprise greetings;
step S202, performing word processing and word processing on the text subjected to the data cleaning processing to obtain word vector representation and word vector representation;
and step S203, embedding the word vector representation and the word vector representation to obtain a text to be processed.
It can be understood that, in the technical solution provided in this embodiment, the text is subjected to data cleaning, stop words, mood words, and useless contents in the text are removed, the text subjected to data cleaning is subjected to word processing and word processing, word vector representation and word vector representation are obtained, and then the word vector and the word vector representation are embedded, so that the text to be processed is obtained. The coding mode of the mixed coding of the text words is facilitated, and the relation between text contexts can be effectively improved.
As a further improvement of the above method, in an embodiment, as shown in fig. 3, a step diagram of an intelligent customer service work order classification method is provided for another embodiment of the present invention.
The method also comprises the step of carrying out secondary cleaning on the text to be processed, and comprises the following steps:
step S301, judging whether a word vector in a text to be processed is the same as a stop table according to regular expression to obtain a judgment result;
step S302a, if the judgment results are the same, removing the same word vector from the text;
in step S302b, if the determination result is the same, the word vector is retained.
In one embodiment, in consideration of the expression mode of Chinese, the contrary language appearing in the text, such as "you are the service attitude", is the service attitude difference at the complaint site according to the real context, but the regular expression cannot be accurately labeled, and the secondary sampling correction is required manually.
It can be understood that, in the technical solution provided in this embodiment, the secondary cleaning of the text to be processed includes: judging whether the word vector in the text to be processed is the same as the stop table according to the regular expression to obtain a judgment result; if the judgment results are the same, removing the same word vector from the text; if the judgment result is the same, the word vector is reserved. When the regular expression is utilized according to the actual context and cannot be accurately labeled, secondary sampling correction can be carried out manually. The method is beneficial to quickly and accurately realizing the classification of the work order built by the customer service, greatly improves the work order classification and circulation time, effectively improves the time efficiency of enterprise problem handling, and effectively improves the satisfaction degree of the customer service.
As a further improvement of the above method, in an embodiment, as shown in fig. 4, a step diagram of an intelligent customer service work order classification method is provided for another embodiment of the present invention.
Before dividing the marked text into a training set, a verification set and a test set, and respectively inputting the training set, the verification set and the test set into a preset training model for training, the method further comprises the following steps:
step S401, processing the text length of the text to be processed to obtain a processed text with the text length of 512 words;
the method comprises the steps of processing a text with the length exceeding 512 words and/or processing a text with the length smaller than 512 words. The method comprises the following steps:
the reserved head 512 words are adopted for 15 percent of texts in the form of random numbers;
tail 512 words are reserved for 15% of texts;
the middle 512 words are reserved for 30% of the text;
the head 218 words and the tail 294 words are reserved for the remaining text, respectively.
Zero padding is carried out on the tail part of the text line with the length of the text to be processed being less than 512 words until 512 words are reached.
Specifically, because the word vector length input by the BERT model is 512, the text line with the text length exceeding 512 words is processed; the embodiment adopts a random number form and adopts 512 reserved head words for 15 percent of texts; tail 512 words are reserved for 15% of texts; the middle 512 words are reserved for 30% of the text; respectively reserving 218 head words and 294 tail words of the rest text; zero padding is performed on the tail of a text line with the text length less than 512 words until 512 words are reached.
When the text line to be processed is represented by a single word vector, the word vector representation and the word vector representation are ensured to be embedded in order to ensure that the length of the word vector representation is 128.
Step S402, carrying out word frequency statistics on the processed text and carrying out position coding on each word in the processed text to obtain the relation between semantic contexts of the processed text. Specifically, performing word frequency statistics on the processed text comprises: word group statistics and single word statistics.
Each word in the processed text is subjected to position coding, and the embodiment adopts a transform model to carry out position coding on each word in the processed text. The association between semantic contexts can be achieved by encoding the word positions.
It should be noted that the transformer model is a common technique for those skilled in the art, and will not be described in detail herein.
In one embodiment, word frequency statistics is carried out on the processed text, word vector training is carried out by adopting a word2vec tool carried by a scimit-learn library, and the length of a word vector is 128; and loading a trained word vector model and carrying out word vector representation on the processed text, wherein the words can be single words, double words or words exceeding double words, so that for each word of the text with the length of 512, if a word group exists, the word group is used for representing the word vector of the word, and if the word group exists, the word vector can be directly represented.
It should be noted that word2vec tools available from scinit-spare libraries are common techniques for those skilled in the art and will not be described in detail herein.
It can be understood that, in the technical solution provided in this embodiment, the text length of the text to be processed is processed, so as to obtain a processed text with a text length of 512 words; the word frequency statistics is carried out on the processed text, and the position coding is carried out on each word in the processed text, so that the relation between semantic contexts of the processed text is obtained, the content of the text is processed more flexibly, the representation method for converting the indefinite length into the definite length word is realized, the coding mode of text word mixing coding is realized, and the relation between the text contexts is effectively improved; the method is beneficial to quickly and accurately realizing the classification of the work order built by the customer service, greatly improves the work order classification and circulation time, effectively improves the time efficiency of enterprise problem handling, and effectively improves the satisfaction degree of the customer service.
It is understood that the same or similar parts in the above embodiments may be mutually referred to, and the same or similar parts in other embodiments may be referred to for the content which is not described in detail in some embodiments.
It should be noted that, in the description of the present application, the terms "first", "second", etc. are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Further, in the description of the present application, the meaning of "a plurality" means at least two unless otherwise specified.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and the scope of the preferred embodiments of the present application includes other implementations in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present application.
It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present application may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
In the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Although embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application, and that variations, modifications, substitutions and alterations may be made to the above embodiments by those of ordinary skill in the art within the scope of the present application.
Claims (10)
1. An intelligent customer service work order classification method is characterized by comprising the following steps:
acquiring voice conversations between a customer service and a customer, and converting the voice conversations into texts;
performing data cleaning, word processing and word processing on the text to obtain a text to be processed;
labeling the text type of the text to be processed;
dividing the marked text into a training set, a verification set and a test set, and respectively inputting the training set, the verification set and the test set into a preset training model for training to obtain a work order classification result.
2. The method according to claim 1, wherein labeling the text to be processed for text category comprises labeling,
and matching a preset text type with the text to be processed by adopting a regular expression, and labeling the text type according to a matching result.
3. The method of claim 2, wherein the preset text categories comprise: the courier leaves the job, the network node is closed, the network node is exploded, the network node has poor service attitude and the network node is signed for non-receipt.
4. The method of claim 1, wherein performing data cleaning, word segmentation and word vector processing on the text to obtain a text to be processed comprises:
cleaning data of the text, and removing stop words, tone words and useless contents in the text, wherein the useless contents comprise greetings;
performing word processing and word processing on the text subjected to the data cleaning processing to obtain word vector representation and word vector representation;
and embedding the word vector and the expression of the word vector to obtain the text to be processed.
5. The method of claim 4, further comprising performing a secondary cleaning of the text to be processed, comprising:
judging whether the word vector in the text to be processed is the same as the stop table according to the regular expression to obtain a judgment result;
if the judgment results are the same, removing the same word vector from the text;
and if the judgment results are the same, reserving the word vector.
6. The method of claim 1, wherein before dividing the labeled text into a training set, a validation set, and a test set, and inputting the training set, the validation set, and the test set into a preset training model for training, the method further comprises:
processing the text length of the text to be processed to obtain a processed text with the text length of 512 words;
and carrying out word frequency statistics on the processed text and carrying out position coding on each word in the processed text to obtain the relation between semantic contexts of the processed text.
7. The method of claim 5, wherein processing the text length of the text to be processed comprises: and/or processing the text with the length smaller than 512 words.
8. The method of claim 6, wherein processing the text with the length exceeding 512 words comprises:
the reserved head 512 words are adopted for 15 percent of texts in the form of random numbers;
tail 512 words are reserved for 15% of texts;
the middle 512 words are reserved for 30% of the text;
the head 218 words and the tail 294 words are reserved for the remaining text, respectively.
9. The method of claim 6, wherein processing the text with the length less than 512 words comprises:
and zero padding is carried out on the tail part of the text line with the length less than 512 words of the text to be processed until 512 words are reached.
10. The method of claim 6, wherein performing word frequency statistics on the processed text comprises: word group statistics and single word statistics.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011194254.3A CN112287639A (en) | 2020-10-30 | 2020-10-30 | Intelligent customer service work order classification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011194254.3A CN112287639A (en) | 2020-10-30 | 2020-10-30 | Intelligent customer service work order classification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112287639A true CN112287639A (en) | 2021-01-29 |
Family
ID=74353796
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011194254.3A Pending CN112287639A (en) | 2020-10-30 | 2020-10-30 | Intelligent customer service work order classification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112287639A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109492091A (en) * | 2018-09-28 | 2019-03-19 | 科大国创软件股份有限公司 | A kind of complaint work order intelligent method for classifying based on convolutional neural networks |
| WO2019200806A1 (en) * | 2018-04-20 | 2019-10-24 | 平安科技（深圳）有限公司 | Device for generating text classification model, method, and computer readable storage medium |
| CN110413746A (en) * | 2019-06-25 | 2019-11-05 | 阿里巴巴集团控股有限公司 | The method and device of intention assessment is carried out to customer problem |
| CN110472246A (en) * | 2019-08-16 | 2019-11-19 | 上海掌学教育科技有限公司 | Work order classification method, device and storage medium |
-
2020
- 2020-10-30 CN CN202011194254.3A patent/CN112287639A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019200806A1 (en) * | 2018-04-20 | 2019-10-24 | 平安科技（深圳）有限公司 | Device for generating text classification model, method, and computer readable storage medium |
| CN109492091A (en) * | 2018-09-28 | 2019-03-19 | 科大国创软件股份有限公司 | A kind of complaint work order intelligent method for classifying based on convolutional neural networks |
| CN110413746A (en) * | 2019-06-25 | 2019-11-05 | 阿里巴巴集团控股有限公司 | The method and device of intention assessment is carried out to customer problem |
| CN110472246A (en) * | 2019-08-16 | 2019-11-19 | 上海掌学教育科技有限公司 | Work order classification method, device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110555095B (en) | Man-machine conversation method and device | |
| US11004013B2 (en) | Training of chatbots from corpus of human-to-human chats | |
| CN111212190B (en) | Conversation management method, device and system based on conversation strategy management | |
| CN107329967A (en) | Question answering system and method based on deep learning | |
| CN108073600A (en) | A kind of intelligent answer exchange method, device and electronic equipment | |
| CN108416375B (en) | Work order classification method and device | |
| CN109740155A (en) | A kind of customer service system artificial intelligence quality inspection rule self concludes the method and system of model | |
| CN104281615A (en) | Complaint handling method and system | |
| CN111861596B (en) | Text classification method and device | |
| CN113780007A (en) | Corpus screening method, intention recognition model optimization method, equipment and storage medium | |
| CN113536081B (en) | Data center data management method and system based on artificial intelligence | |
| CN108984279A (en) | A kind of streaming computing method of internet of things oriented tradition SQL developer | |
| CN110321564A (en) | A kind of more wheel dialogue intension recognizing methods | |
| CN112966082A (en) | Audio quality inspection method, device, equipment and storage medium | |
| WO2021159655A1 (en) | Data attribute filling method, apparatus and device, and computer-readable storage medium | |
| CN109618067A (en) | Outgoing call dialog process method and system | |
| CN112287639A (en) | Intelligent customer service work order classification method | |
| CN115688868B (en) | Model training method and computing equipment | |
| WO2020151017A1 (en) | Scalable field human-machine dialogue system state tracking method and device | |
| CN115952461A (en) | Pre-training corpus cleaning method, system and storage medium | |
| CN107066450B (en) | Instant messaging session segmentation method based on learning | |
| CN112860873B (en) | Intelligent response method, device and storage medium | |
| US20190213554A1 (en) | User support system with automatic message categorization | |
| CN113705199A (en) | Work order priority confirmation method, work order priority confirmation device, work order priority confirmation electronic equipment, work order priority confirmation medium and work order priority confirmation product | |
| CN113868415A (en) | Knowledge base generation method and device, storage medium and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |