CN116542676A - Intelligent customer service system based on big data analysis and method thereof - Google Patents
Intelligent customer service system based on big data analysis and method thereof Download PDFInfo
- Publication number
- CN116542676A CN116542676A CN202310823321.0A CN202310823321A CN116542676A CN 116542676 A CN116542676 A CN 116542676A CN 202310823321 A CN202310823321 A CN 202310823321A CN 116542676 A CN116542676 A CN 116542676A
- Authority
- CN
- China
- Prior art keywords
- module
- customer service
- client
- information
- question
- 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 20
- 238000007405 data analysis Methods 0.000 title claims abstract description 18
- 230000008451 emotion Effects 0.000 claims abstract description 51
- 230000008909 emotion recognition Effects 0.000 claims abstract description 16
- 238000000605 extraction Methods 0.000 claims description 11
- 239000008186 active pharmaceutical agent Substances 0.000 description 16
- 238000012549 training Methods 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 239000003086 colorant Substances 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 5
- 230000011218 segmentation Effects 0.000 description 5
- 238000012706 support-vector machine Methods 0.000 description 4
- 238000013145 classification model Methods 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000012217 deletion Methods 0.000 description 2
- 230000037430 deletion Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000003780 insertion Methods 0.000 description 2
- 230000037431 insertion Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000006854 communication Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/01—Customer relationship services
- G06Q30/015—Providing customer assistance, e.g. assisting a customer within a business location or via helpdesk
-
- 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
- G06F16/33—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
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06311—Scheduling, planning or task assignment for a person or group
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Economics (AREA)
- Artificial Intelligence (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- Marketing (AREA)
- Data Mining & Analysis (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Finance (AREA)
- Human Computer Interaction (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Accounting & Taxation (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of intelligent customer service, in particular to an intelligent customer service system and method based on big data analysis. The invention searches the questions of the clients by the question searching module preferentially, if the matched answers exist in the question-answer library, the clients are connected into the electronic customer service module, so that the electronic customer service can give satisfactory answers to the customer service, the step that the clients are connected into the manual module after the electronic customer service can not answer is avoided, the satisfaction degree of the clients is improved, the emotion recognition module is used for recognizing the current emotion state of the clients, thereby better giving the clients service, increasing the trust of the clients to enterprises, and carrying out priority allocation on the electronic customer service module and the manual customer service module by the priority allocation strategy, so that the clients with negative emotion can be connected into the manual customer service module preferentially.
Description
Technical Field
The invention relates to the technical field of intelligent customer service, in particular to an intelligent customer service system and method based on big data analysis.
Background
With the popularization of the internet, more and more users select services such as online shopping, online consultation and the like, and for enterprises, providing efficient and convenient customer service is an important means for maintaining user loyalty, so many intelligent customer service systems are already presented in the market, but there are many defects in these systems, wherein:
the intelligent online customer service comprises an artificial customer service and an electronic customer service, wherein the electronic customer service is to answer a customer question by a machine, the artificial customer service is to answer the customer question by a real person, in the prior art, the problem is directly transferred into the artificial customer service, if the problem is transferred into the electronic customer service, and when the electronic customer service can not be solved, the artificial customer service is accessed, and the unordered customer service question answering mode is adopted to solve the problem with lower efficiency, so that the online experience feeling of the customer is easily reduced;
especially, when some customer emotion is unstable, such as urgent emotion, gas emotion and the like exist, if the traditional intelligent customer service can not timely and satisfactorily answer the problem of the customer according to the current situation of the customer, the customer can transfer to the manual customer service again after answering after passing through the electronic customer service, so that the problem presented by the customer can be solved in one step according to the actual requirement of the customer, the unstable situation of the customer can be fermented, and the customer can lose trust to an enterprise.
In view of this, we propose an intelligent customer service system based on big data analysis and a method thereof to solve the above-mentioned problems.
Disclosure of Invention
The invention aims to provide an intelligent customer service system based on big data analysis and a method thereof, which are used for solving the problems in the background technology.
In order to solve the technical problems, one of the purposes of the invention is to provide an intelligent customer service system based on big data analysis, which comprises an intelligent customer service unit, wherein the intelligent customer service unit comprises a customer problem acquisition module, a manual customer service module, an electronic customer service module and a question and answer library updating module;
the client problem acquisition module is used for acquiring information of client consultation, extracting key words of the client consultation information and matching the information in the question-answering library, and analyzing the current emotion of the client to obtain a question-answering form suitable for the client;
the manual customer service module is used for receiving the consultation information acquired by the problem acquisition module by the customer, and accessing the customer to be processed into the manual customer service capable of replying at the fastest speed by identifying the idleness of the manual customer service;
the electronic customer service module is used for receiving the consultation information transmitted by the customer question acquisition module and outputting a reply text matched with the information of the customer consultation;
the question-answering library updating module is used for updating the question-answering template and updating and optimizing the content in the question-answering library according to the feedback of the client consultation information and the historical service data.
Preferably, the client problem acquisition module comprises a semantic understanding module, a problem retrieval module and an emotion recognition module;
the semantic understanding module is used for identifying wrongly written characters appearing in the customer consultation information, replacing the wrongly written characters according to semantics, and transmitting the keywords in the extracted sentences to the problem retrieval module;
the question retrieval module is used for retrieving a reply text matched with the client consultation information in the question-answer library according to the keywords transmitted by the semantic understanding module;
the emotion recognition module is used for recognizing the current emotion state of the client according to the voice intonation of the client during consultation and negative emotion contained in the consultation information.
Preferably, the problem acquisition module further comprises a priority distribution module;
the priority distribution module is used for receiving the current emotion of the client output by the emotion recognition module, and matching a reply form conforming to the client according to the current emotion of the client by using a priority distribution strategy.
Preferably, the semantic understanding module comprises a mispronounced word replacement module and a keyword extraction module;
the wrongly written word replacement module is used for replacing wrongly written words appearing in the consultation information according to the semantics of the customer consultation information;
the keyword extraction module is used for extracting keywords in the client consultation information.
Preferably, the manual customer service module comprises a first text receiving module and a text distributing module;
the first text receiving module is used for receiving information of client consultation;
the text distribution module is used for searching the currently idle manual customer service and transmitting the customer consultation information received by the first text receiving module to the matched manual customer service.
Preferably, the text distribution module comprises an idle retrieval module and a minimum client identification module;
the idle retrieval module is used for retrieving the manual customer service of the current customers which are not to be processed;
the minimum customer identification module is used for retrieving the manual customer service with the minimum number of currently pending customers.
Preferably, the electronic customer service module comprises a second text receiving module and a text output module;
the second text receiving module is used for receiving information of client consultation;
the text output module is used for outputting matched reply text according to the consultation information received by the second text receiving module.
Preferably, the question-answering library updating module comprises a client opinion feedback module;
the client opinion feedback module is used for collecting opinion information proposed by clients and feeding back the client opinion information to the question-answering library.
The second object of the present invention is to provide a method for intelligent customer service system based on big data analysis, comprising the following steps:
s1, acquiring information of client consultation, and matching the information in a question-answer library by extracting key words of the client consultation information, and analyzing the current emotion of the client to obtain a question-answer form suitable for the client;
s2, identifying the idleness of the artificial customer service, and accessing the customer to be processed into the artificial customer service which can reply the fastest;
s3, transmitting the information of the client consultation to the electronic customer service, and outputting a reply text matched with the information of the client consultation by the electronic customer service;
and S4, updating and optimizing the content in the question-answer library according to the feedback of the client consultation information and the historical service data.
Compared with the prior art, the intelligent customer service system and the intelligent customer service method based on big data analysis have the beneficial effects that:
1. the problem retrieval module is used for retrieving the problems of the clients preferentially, if the matched answers exist in the question and answer library, the clients are connected to the electronic customer service module, so that the electronic customer service can give satisfactory answers to the customer service, the step that the clients are connected to the manual module after the electronic customer service cannot answer is avoided, the satisfaction degree of the clients is improved, the emotion recognition module is used for recognizing the current emotion state of the clients, the clients are better served, and the trust of the clients to enterprises is increased.
2. Compared with the prior art, the intelligent customer service system and the intelligent customer service method based on big data analysis have the beneficial effects that: according to the emotion condition of the customer identified by the emotion identification module, the priority allocation strategy is used for allocating the priority of the electronic customer service module and the priority of the artificial customer service module, so that the customer with negative emotion can access the artificial customer service module preferentially, and the customer is pacified by the artificial customer service, so that the customer is provided with satisfied service.
Drawings
FIG. 1 is a schematic diagram of the overall structure of an embodiment;
FIG. 2 is a schematic diagram of the client problem collection module, the manual customer service module, and the electronic customer service module according to an embodiment;
FIG. 3 is a schematic diagram of a semantic understanding module architecture of an embodiment;
FIG. 4 is a schematic diagram of a text distribution module according to an embodiment;
fig. 5 is an overall method flow diagram of an embodiment.
The meaning of each reference sign in the figure is:
10. an intelligent customer service unit;
11. a customer problem acquisition module;
111. a semantic understanding module; 1111. a misprinted word replacement module; 1112. a keyword extraction module;
112. a question retrieval module;
113. an emotion recognition module;
12. a manual customer service module; 121. a first text receiving module; 122. a text distribution module; 1221. an idle retrieval module; 1222. a minimum customer identification module;
13. an electronic customer service module; 131. a second text receiving module; 132. a text output module;
14. a question-answer library updating module; 141. and a customer opinion feedback module.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1: as shown in fig. 1 to 5, one of the purposes of the present invention is to provide an intelligent customer service system based on big data analysis, which comprises an intelligent customer service unit 10, wherein the intelligent customer service unit 10 comprises a customer problem acquisition module 11, a manual customer service module 12, an electronic customer service module 13 and a question and answer library updating module 14;
the client problem acquisition module 11 is used for acquiring information of client consultation, and matching the information in the question-answering library by extracting key words of the client consultation information, and analyzing the current emotion of the client to obtain a question-answering form suitable for the client;
the client problem acquisition module 11 comprises a semantic understanding module 111, a problem retrieval module 112 and an emotion recognition module 113;
the client problem collecting module 11 collects the problem of client consultation through an API interface, wherein a mode of sending voice exists during client consultation, so that voice needs to be converted into text to be collected, at this time, hundred-degree AI voice recognition in an API service can be used for converting voice into text, and when information collection is carried out through the API interface, the API interface needs to be registered, a development key is applied to obtain the authority of using the API, a programming language supporting the API interface, such as Python, java, C, is used for compiling program codes for accessing the API, so as to realize collection of API interface data, the compiled program codes are integrated with the API interface, and client consultation information needing to be collected is obtained by calling the API interface.
The semantic understanding module 111 is used for identifying wrongly written characters appearing in the customer consultation information, replacing the wrongly written characters according to semantics, and transmitting keywords in the extracted sentences to the question retrieval module 112;
the semantic understanding module 111 includes a misprint word replacement module 1111 and a keyword extraction module 1112;
the wrongly written word replacement module 1111 is configured to replace wrongly written words occurring in the consultation information according to semantics of the customer consultation information;
specifically, firstly, text content is required to be extracted from consultation information submitted by a customer, namely, consultation information in the form of voice or handwriting characters and the like is converted into plain text, then semantic inference is carried out on the extracted text, so that specific meaning in the consultation information of the customer is understood, error detection is carried out by using an error detection algorithm, error words possibly existing in the text are found out, and the detected error words are corrected to correct errors in the text, so that statement smoothness in the information of customer service consultation is ensured, and answers about customer questions are accurately matched in a question-answering library;
the specific calculation steps of the error detection algorithm are as follows:
dividing text data into individual characters or words according to the characters or words by using a Chinese word segmentation technology;
for each word with good word segmentation, calculating the editing distance between the word and the word in the predefined dictionary, wherein the editing distance is calculated by using a dynamic programming algorithm, and the formula is as follows:
；
wherein,,representing calculation to the +.>The individual characters and->The distance between the individual characters, first +.>Is for the score when adding a character or deleting a character, second +.>Is for the score when replacing a character, < >>Judging whether the current two characters are different or not, and finally selecting the word with the minimum editing distance;
when the edit distance is less than a predefined threshold, the wrong word may be marked or replaced, and the wrong text identified.
For example: one customer input: the method comprises the following steps that I want to purchase a latest iphon mobile phone, the word "iphon" in a message is misspelled, the word "iphon" can be corrected to be "iPhone" by using an edit distance algorithm, firstly, a sentence is segmented to obtain the word "I want to purchase the latest iphon mobile phone", and for the word iphon, the edit distance between the iphon and the iPhone can be calculated by using the edit distance algorithm according to the following steps:
initializing a matrix d, wherein the size of the matrix d is m rows and n columns, and m and n are the lengths of two character strings respectively;
initializing a first row and a first column of the matrix from 0 to m and 0 to n;
traversing each element of the matrix from left to right and from top to bottom, and for the element d (i, j) traversed currently, performing the following calculation according to different conditions of the character string:
if word1[ i ] is equal to word2[ j ], then no modification operation is needed, and only copy operation, i.e., calculation, is performed: d (i, j) =d (i-1, j-1);
otherwise, one of the operations of insertion, deletion and replacement is needed, the following three values are respectively calculated, and the minimum value is taken as the value of the current element:
insertion operation: word1[ i ] is inserted before word2[ j ], i.e.: d (i, j) =d (i, j-1) +1;
deletion operation: word1[ i ] is deleted, namely: d (i, j) =d (i-1, j) +1;
replacement operation: word1[ i ] is replaced with word2[ j ], namely: d (i, j) =d (i-1, j-1) +1;
the last element d (m, n) of the final matrix is the editing distance between iphon and iPhone;
according to the above steps, if word 1= "iphon", word 2= "iPhone", the process of calculating the edit distance is as follows:
i p h o n
0 1 2 3 4 5
I 1 1 2 3 4
P 2 2 2 3 4
H 3 3 2 3 4
O 4 4 3 2 3
N 5 5 4 3 2
thus, the edit distance between iphon and iPhone is 2;
from the calculation of the edit distance, we can find the possible correct spelling closest to the target word, i.e. "iPhone";
the misspelled word "iphon" is replaced with the correct word "iPhone" to correct the error.
Therefore, through the steps, the error word 'iphon' in the message can be corrected to 'iPhone' by using an edit distance algorithm, so that the error correction effect is achieved.
The keyword extraction module 1112 is configured to extract keywords in the client consultation information.
The method comprises the following specific steps of:
the client consultation information is segmented, nonsensical contents such as stop words and punctuations are removed, the occurrence times of each word in all documents are counted, the TF-IDF value of each word in each document is calculated, and the formula is as follows:
wherein the method comprises the steps ofRepresents "word frequency-inverse document frequency,">Representing the number of occurrences of the word in the document divided by the total number of documents, +.>A logarithm of the reciprocal of the number of documents representing the word appearing in all documents;
and sorting according to the number of keywords in each document or TF-IDF values, and selecting the words with the top ranking as keywords.
For example, a customer may leave a message: "please ask up to date which colors of iphones are optional? The keyword extraction of the message comprises the following steps:
chinese word segmentation is carried out on the message to obtain' which colors are selectable in the latest IPhone to ask for;
we can calculate TF-IDF values for each word, assuming that our document set contains 100 documents, with 50 documents containing the most recent word, the word "most recent" IDF is log (100/50) =0.301.
For the word "latest" in the above words, the number of occurrences is 1, the total word number is 7, and tf=1/7=0.143, then the TF-IDF value of the word is tfidf= 0.1430.301 =0.043, and similarly, we can calculate TF-IDF values of other words in the above words;
through calculation, counting the occurrence times of each word in all documents, we can find that the words such as the latest words and the iPhone have higher TF-IDF values and can also be used as keywords for extraction.
The question searching module 112 is configured to search a question-answer library for a answer text matching with the client consultation information according to the keyword transmitted by the semantic understanding module 111;
specifically, using a hundred-degree search engine to search in a question-answer library by taking the extracted keywords as search conditions, screening and sorting the searched texts, and finding out the best matched answer text;
for example, consulting information for the following clients: please ask up to date which colors of iphones are optional?
Assume that there are the following text in the question and answer library:
"the latest iPhone has 5 colors in total, black, white, red, yellow and blue, respectively; "
"the selling price of the latest model number of the iPhone can be inquired on the official website, and you can search on the official website; "
The method can perform word segmentation and keyword extraction on the client consultation information to obtain keywords such as 'latest', 'iPhone' and 'color';
then, using the keywords as search conditions, searching in a question-answer library, and searching to find text information containing keywords such as 'latest', 'iPhone' and 'color' mentioned in a first text;
finally, we can answer the matching text and send it to the customer, for example:
answering: the latest iPhone has a total of 5 colors, black, white, red, yellow and blue, respectively.
The emotion recognition module 113 is used for recognizing the current emotion state of the client according to the voice intonation of the client during consultation and the negative emotion contained in the consultation information.
The emotion recognition module 113 adopts a support vector machine model to judge the emotion of the client, and the following steps are adopted:
preparing a text data set marked with emotion polarities, classifying the emotion polarities commonly used into positive, negative and neutral classifications, and marking the texts in the text data set as corresponding polarity categories;
text preprocessing operations such as word segmentation, stop word removal, word stem extraction and the like are carried out on texts in a text data set so as to obtain a better emotion classification result;
through the word frequency-inverse document frequency TF-IDF method, text is expressed as a vector, so that the subsequent emotion classification model establishment and training are facilitated;
dividing the processed text data set into a training set and a testing set, wherein the training set is used for training an emotion classification model, the testing set is used for verifying the accuracy and generalization capability of the model, and generally, the training set accounts for 80% of the data set, and the testing set accounts for 20% of the data set;
using text information of a training set and marked emotion polarities as training data, and establishing an emotion classification model of a support vector machine;
and carrying out emotion classification prediction on the new text information by using the trained support vector machine model so as to determine emotion polarity and emotion strength of the text.
For example, suppose that the customer consultation information is: "I have recently purchased a piece of your company's clothing, but are somewhat disappointed, not of good quality", we need to determine from this sentence whether the customer's emotion is positive or negative;
first we can collect some text data that has been marked with emotional polarity, such as "a very nice piece of half-value skirt-! "buying shoes are praise", "the clothes have good color, but have poor quality, really disappointed", and the like, the stop words are needed to be removed, the text data with clean punctuation marks is obtained, new unknown data are converted into vector representation, emotion judgment is carried out through a trained model, so that client consultation information "I have recently purchased clothes of your company, but have little disappointed, the quality is not good", emotion judgment is carried out through a trained support vector machine classifier, and the consultation information is obtained as negative emotion.
According to the judgment result, the current emotion state of the client is known, so that the client can be answered in consideration of the emotion of the client when replying to the client, the emotion of the client is calmed, and the satisfaction degree of the client is improved.
Further, after the client problem collection module 11 collects the problem of the client consultation, the semantic understanding module 111 firstly carries out sentence understanding on the information of the client consultation, the problem retrieval module 112 retrieves in the question and answer library according to the problem of the client consultation, when the matched answer text is retrieved, the problem of the client consultation is accessed to the electronic customer service module 13, and is communicated with the client through the electronic customer service module 13, if the corresponding answer text is not retrieved in the question and answer library, the client is accessed to the manual customer service module 12, so that the manual customer service is communicated with the client, and the prediction of the current emotion of the client is combined with the emotion recognition module 113 in the communication process, thereby improving the service quality.
The artificial customer service module 12 is used for receiving the consultation information acquired by the customer problem acquisition module 11, and accessing the customer to be processed into the artificial customer service which can reply the fastest by identifying the idleness of the artificial customer service;
the manual customer service module 12 comprises a first text receiving module 121 and a text distributing module 122;
the first text receiving module 121 is used for receiving information of client consultation;
the text distribution module 122 is configured to retrieve the currently idle artificial customer service, and transmit the customer consultation information received by the first text receiving module 121 to the matched artificial customer service.
Wherein text assignment module 122 includes a free retrieval module 1221 and a minimum customer identification module 1222;
the idle searching module 1221 is used for searching the manual customer service of the customers which are not currently to be processed;
specifically, the zendsk data monitoring software may be used to monitor the current session state of the customer service personnel, for example, it may detect whether the customer service personnel is online, and whether there is an available state, for example, "online", "away", "busy", etc., the system may also detect the "knocked-in" state of the customer service personnel, if the customer service personnel uses a Zendesk Support Suite application, when a new customer request has arrived, their application may make a sound or display notification, if the customer service personnel switches to other applications, the system may detect whether these notifications are ignored, or if a positive response has been made, and when it is detected that the customer service of the current customer without treatment is currently available, the customer to be treated may be preferentially allocated to the customer service of the customer without treatment for treatment;
and the zendsk data monitoring software will detect the session window status of the customer service personnel to determine if they are actively talking to the customer, if the customer service personnel has selected to "end session" or window is inactive for a long period of time, the system will consider the session to have ended.
The minimum customer identification module 1222 is used to retrieve the manual customer service that currently has the least number of customers to be processed.
Considering that if each of the artificial customer service has a customer to be processed, the minimum customer identification module 1222 can identify the artificial customer service having the least customer to be processed in the queue to be processed of the artificial customer service, and the customer to be processed can be preferentially allocated to the artificial customer service capable of processing the customer to be processed as soon as possible;
specifically, the state of the message queue is obtained through API call, for example, the message queue tool RabbitMQ provides APIs to obtain the state and statistical information of the message queue, and the APIs can be used to easily obtain the number of the clients to be processed in each manual customer service waiting queue, so that the clients to be processed currently can be allocated to the manual customer service with the least number of the clients to be processed in the waiting queue, and the clients can be ensured to obtain the fastest response.
The electronic customer service module 13 is used for receiving the consultation information transmitted by the client question acquisition module 11 and outputting a reply text matched with the information of the client consultation;
the electronic customer service module 13 includes a second text receiving module 131 and a text output module 132;
the second text receiving module 131 is used for receiving information of client consultation;
the text output module 132 is used for outputting matched reply text according to the consultation information received by the second text receiving module 131.
Specifically, when the information of the client consultation searches the answer text matched with the information in the question and answer library through the hundred-degree engine adopted by the question and answer retrieval module 112, the client is accessed to the electronic customer service module 13 under the condition of no special condition, the hundred-degree search API can enable the chat robot to obtain the search result through sending the HTTP request, and the chat robot displays the obtained answer text in a chat dialog box.
The question and answer library updating module 14 is used for updating a question and answer template, and updating and optimizing the content in the question and answer library according to the feedback of the client consultation information and the historical service data.
The question and answer library updating module 14 includes a customer opinion feedback module 141;
the client opinion feedback module 141 is configured to collect opinion information presented by a client and feed back the client opinion information to the question-answering library.
After processing the information of the client consultation, an email or a short message can be sent to the client, wherein the email or the short message comprises satisfaction survey, client opinion, opinion feedback link and the like of the client when the client replies to the client aiming at customer service, the client can access a feedback interface through the link, submit opinion feedback, and acquire client opinion data by using the acquired API key and the ID of a questionnaire or a problem through the API interface;
when the customer opinion information acquired by the customer is fed back into the question-answering library, the collected customer opinion data is required to be subjected to operations such as screening, classifying and classifying, classifying and the like, the customer opinion data is required to be classified and organized according to relevant conditions such as audiences, topics and the like, the ideas or suggestions of similar topics are classified into the same class, classifying, integrating, summarizing, de-duplication and the like are performed, the aim is to generate an organized data set, then the customer opinion data is processed and analyzed, key information such as questions, doubts, demands, suggestions and the like which are presented by the customer is extracted, the contents can be used as sources of questions and answers in the question-answering library, the collected customer opinion information is organized into a format which can be used for the question-answering library, corresponding questions and answers are created or updated in the question-answering library, the keywords, descriptions, relevant background information and the like which are required to be treated and abstracted into general question-answering types, the comments are ensured to be easily searched and understood, and the contents of the knowledge library are updated according to a certain time interval or each time when the data is added, so that the content of the knowledge library is kept to be always the latest and the latest.
The second object of the present invention is to provide a method for intelligent customer service system based on big data analysis, comprising the following steps:
s1, acquiring information of client consultation, and matching the information in a question-answer library by extracting key words of the client consultation information, and analyzing the current emotion of the client to obtain a question-answer form suitable for the client;
s2, identifying the idleness of the artificial customer service, and accessing the customer to be processed into the artificial customer service which can reply the fastest;
s3, transmitting the information of the client consultation to the electronic customer service, and outputting a reply text matched with the information of the client consultation by the electronic customer service;
and S4, updating and optimizing the content in the question-answer library according to the feedback of the client consultation information and the historical service data.
Example 2: according to the above embodiment 1, after the client problem collection module 11 collects the problem of the client consultation, the problem search module 112 is utilized to search in the question and answer library according to the problem of the client consultation, when the matching answer text is searched, the problem of the client consultation can be accessed to the e-customer service module 13, if the corresponding answer text is not searched in the question and answer library, the client is accessed to the artificial customer service module 12, but when the emotion recognition module 113 recognizes that the current emotion state of the client is angry and discontent, the e-customer service module 13 can process the problem of the client, but is inconvenient to provide humanized service to the client, so that the emotion of the client is inconvenient to pacify, and the client is not satisfied, therefore, the improvement is performed on the basis of the embodiment 1, as shown in fig. 2:
the client problem acquisition module 11 also comprises a priority distribution module;
the priority allocation module is configured to receive the current emotion of the client output by the emotion recognition module 113, and match a reply form conforming to the client according to the current emotion of the client using a priority allocation policy.
According to the above description, when the emotion recognition module 113 recognizes that the current emotion of the client is a negative emotion, the client can be preferentially accessed to the artificial customer service module 12, and the emotion of the client is pacified through the artificial customer service, so that the satisfaction degree of the client is improved.
The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the above-described embodiments, and that the above-described embodiments and descriptions are only preferred embodiments of the present invention, and are not intended to limit the invention, and that various changes and modifications may be made therein without departing from the spirit and scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (6)
1. The utility model provides an intelligent customer service system based on big data analysis, includes intelligent customer service unit (10), its characterized in that: the intelligent customer service unit (10) comprises a customer problem acquisition module (11), a manual customer service module (12), an electronic customer service module (13) and a question and answer library updating module (14);
the client problem acquisition module (11) is used for acquiring information of client consultation, extracting keywords of the client consultation information and matching the information in the question-answering library, and analyzing the current emotion of the client to obtain a question-answering form suitable for the client; the client problem acquisition module (11) comprises a semantic understanding module (111), a problem retrieval module (112) and an emotion recognition module (113); the semantic understanding module (111) is used for identifying wrongly written characters appearing in the customer consultation information, replacing the wrongly written characters according to semantics, and transmitting keywords in the extracted sentences to the problem retrieval module (112); the question retrieval module (112) is used for retrieving a reply text matched with the client consultation information in a question and answer library according to the keywords transmitted by the semantic understanding module (111); the emotion recognition module (113) is used for recognizing the current emotion state of the client according to the voice intonation of the client during consultation and the negative emotion contained in the consultation information;
the manual customer service module (12) is used for receiving the consultation information acquired by the problem acquisition module (11) by the customer, and accessing the customer to be processed into the manual customer service capable of replying at the fastest speed by identifying the idleness of the manual customer service; the manual customer service module (12) comprises a first text receiving module (121) and a text distributing module (122); the first text receiving module (121) is used for receiving information of client consultation; the text distribution module (122) is used for retrieving currently idle manual customer service and transmitting the customer consultation information received by the first text receiving module (121) to the matched manual customer service;
the electronic customer service module (13) is used for receiving the consultation information transmitted by the customer question acquisition module (11) and outputting a reply text matched with the information of the customer consultation; the electronic customer service module (13) comprises a second text receiving module (131) and a text output module (132); the second text receiving module (131) is used for receiving information of client consultation; the text output module (132) is used for outputting matched reply text according to the consultation information received by the second text receiving module (131);
the question and answer library updating module (14) is used for updating a question and answer template and updating and optimizing the content in the question and answer library according to the feedback of the client consultation information and the historical service data.
2. The intelligent customer service system based on big data analysis of claim 1, wherein: the problem acquisition module (11) further comprises a priority distribution module;
the priority allocation module is used for receiving the current emotion of the client output by the emotion recognition module (13), and matching a reply form conforming to the client according to the current emotion of the client by using a priority allocation strategy.
3. The intelligent customer service system based on big data analysis of claim 1, wherein: the semantic understanding module (111) comprises a mispronounced word replacement module (111) and a keyword extraction module (112);
the wrongly written word replacement module (111) is used for replacing wrongly written words appearing in the consultation information according to the semantics of the customer consultation information;
the keyword extraction module (112) is used for extracting keywords in the client consultation information.
4. The intelligent customer service system based on big data analysis of claim 1, wherein: the text distribution module (122) includes a free retrieval module (1221) and a minimum customer identification module (1222);
the idle retrieval module (1221) is used for retrieving the manual customer service of the customers which are not currently to be processed;
the minimum customer identification module (1222) is configured to retrieve a human customer service having a minimum number of currently pending customers.
5. The intelligent customer service system based on big data analysis of claim 4, wherein: the question and answer library updating module (14) comprises a client opinion feedback module (141);
the client opinion feedback module (141) is used for collecting opinion information proposed by clients and feeding back the client opinion information to the question-answering library.
6. A method applied to an intelligent customer service system based on big data analysis as claimed in any of claims 1-5, characterized in that: the method comprises the following steps:
s1, acquiring information of client consultation, and matching the information in a question-answer library by extracting key words of the client consultation information, and analyzing the current emotion of the client to obtain a question-answer form suitable for the client;
s2, identifying the idleness of the artificial customer service, and accessing the customer to be processed into the artificial customer service which can reply the fastest;
s3, transmitting the information of the client consultation to the electronic customer service, and outputting a reply text matched with the information of the client consultation by the electronic customer service;
and S4, updating and optimizing the content in the question-answer library according to the feedback of the client consultation information and the historical service data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310823321.0A CN116542676A (en) | 2023-07-06 | 2023-07-06 | Intelligent customer service system based on big data analysis and method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310823321.0A CN116542676A (en) | 2023-07-06 | 2023-07-06 | Intelligent customer service system based on big data analysis and method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116542676A true CN116542676A (en) | 2023-08-04 |
Family
ID=87454592
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310823321.0A Pending CN116542676A (en) | 2023-07-06 | 2023-07-06 | Intelligent customer service system based on big data analysis and method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116542676A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117408708A (en) * | 2023-11-09 | 2024-01-16 | 南方电网储能股份有限公司信息通信分公司 | Customer service center dispatching system based on big data |
| CN118333636A (en) * | 2024-06-17 | 2024-07-12 | 杭州干城新物联科技有限公司 | Automatic customer service processing system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105808694A (en) * | 2016-03-04 | 2016-07-27 | 上海携程商务有限公司 | Online customer service response system and method |
| US20160247068A1 (en) * | 2013-11-01 | 2016-08-25 | Tencent Technology (Shenzhen) Company Limited | System and method for automatic question answering |
| CN109040484A (en) * | 2018-07-16 | 2018-12-18 | 安徽信尔联信息科技有限公司 | A kind of Auto-matching contact staff method |
| CN110149450A (en) * | 2019-05-22 | 2019-08-20 | 欧冶云商股份有限公司 | Intelligent customer service answer method and system |
| CN113377943A (en) * | 2021-08-16 | 2021-09-10 | 中航信移动科技有限公司 | Multi-round intelligent question-answering data processing system |
| CN115643341A (en) * | 2022-10-14 | 2023-01-24 | 杭州半云科技有限公司 | Artificial intelligence customer service response system |
-
2023
- 2023-07-06 CN CN202310823321.0A patent/CN116542676A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160247068A1 (en) * | 2013-11-01 | 2016-08-25 | Tencent Technology (Shenzhen) Company Limited | System and method for automatic question answering |
| CN105808694A (en) * | 2016-03-04 | 2016-07-27 | 上海携程商务有限公司 | Online customer service response system and method |
| CN109040484A (en) * | 2018-07-16 | 2018-12-18 | 安徽信尔联信息科技有限公司 | A kind of Auto-matching contact staff method |
| CN110149450A (en) * | 2019-05-22 | 2019-08-20 | 欧冶云商股份有限公司 | Intelligent customer service answer method and system |
| CN113377943A (en) * | 2021-08-16 | 2021-09-10 | 中航信移动科技有限公司 | Multi-round intelligent question-answering data processing system |
| CN115643341A (en) * | 2022-10-14 | 2023-01-24 | 杭州半云科技有限公司 | Artificial intelligence customer service response system |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117408708A (en) * | 2023-11-09 | 2024-01-16 | 南方电网储能股份有限公司信息通信分公司 | Customer service center dispatching system based on big data |
| CN118333636A (en) * | 2024-06-17 | 2024-07-12 | 杭州干城新物联科技有限公司 | Automatic customer service processing system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110765257B (en) | Intelligent consulting system of law of knowledge map driving type | |
| CN116756295B (en) | Knowledge base retrieval method, device and storage medium | |
| CN110096581B (en) | System and method for establishing question-answer system recommendation questions based on user behaviors | |
| US9710829B1 (en) | Methods, systems, and articles of manufacture for analyzing social media with trained intelligent systems to enhance direct marketing opportunities | |
| CN116542676A (en) | Intelligent customer service system based on big data analysis and method thereof | |
| US20100079464A1 (en) | Information processing apparatus capable of easily generating graph for comparing of a plurality of commercial products | |
| KR20180022762A (en) | Method, system and computer-readable recording medium for providing customer counseling service using real-time response message generation | |
| KR20170137419A (en) | Method, system and computer-readable recording medium for providing customer counseling service using real-time response message generation | |
| CN109858626B (en) | Knowledge base construction method and device | |
| CN111062220B (en) | End-to-end intention recognition system and method based on memory forgetting device | |
| TWI734085B (en) | Dialogue system using intention detection ensemble learning and method thereof | |
| CN111782793A (en) | Intelligent customer service processing method, system and equipment | |
| CN114969297A (en) | Method for analyzing power customer appeal relevancy | |
| CN113051380A (en) | Information generation method and device, electronic equipment and storage medium | |
| CN113591487A (en) | Scenic spot comment emotion analysis method based on deep learning | |
| CN117909466A (en) | Domain question-answering system, construction method, electronic device and storage medium | |
| CN117235213A (en) | Interactive customer service method and system | |
| CN111708870A (en) | Deep neural network-based question answering method and device and storage medium | |
| CN114186041A (en) | Answer output method | |
| CN114186040A (en) | Operation method of intelligent robot customer service | |
| CN113553419A (en) | Civil aviation knowledge map question-answering system | |
| CN110532374B (en) | Insurance information processing method and device | |
| CN117113973A (en) | Information processing method and related device | |
| CN111104422A (en) | Training method, device, equipment and storage medium of data recommendation model | |
| JP2009053743A (en) | Document similarity derivation apparatus, document similarity derivation method and document similarity derivation 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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20230804 |