CN110427620B - Service quality optimization management system based on community system - Google Patents
Service quality optimization management system based on community system Download PDFInfo
- Publication number
- CN110427620B CN110427620B CN201910664648.1A CN201910664648A CN110427620B CN 110427620 B CN110427620 B CN 110427620B CN 201910664648 A CN201910664648 A CN 201910664648A CN 110427620 B CN110427620 B CN 110427620B
- Authority
- CN
- China
- Prior art keywords
- dialogue
- service
- text
- emotion
- marketing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 28
- 230000008451 emotion Effects 0.000 claims abstract description 160
- 238000004891 communication Methods 0.000 claims abstract description 34
- 238000004364 calculation method Methods 0.000 claims abstract description 33
- 238000000034 method Methods 0.000 claims description 27
- 230000011218 segmentation Effects 0.000 claims description 19
- 230000008569 process Effects 0.000 claims description 8
- 230000002123 temporal effect Effects 0.000 claims description 8
- 238000010586 diagram Methods 0.000 description 10
- 230000003993 interaction Effects 0.000 description 6
- 230000009286 beneficial effect Effects 0.000 description 5
- 238000004590 computer program Methods 0.000 description 4
- 239000000470 constituent Substances 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000008909 emotion recognition Effects 0.000 description 4
- 230000008520 organization Effects 0.000 description 4
- 230000000737 periodic effect Effects 0.000 description 4
- 206010023644 Lacrimation increased Diseases 0.000 description 2
- 230000004317 lacrimation Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
-
- 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Tourism & Hospitality (AREA)
- General Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Information Transfer Between Computers (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a service quality optimization management system for analyzing dialogue messages between staff and users based on a community system, and further identifying the popularity of marketing information through emotion analysis optimization, so that the staff can push the popular marketing information to improve the service quality, which is characterized by comprising the following steps: a service management device; at least one employee terminal provided with a community module corresponding to the community system and an employee account for access; the system comprises a plurality of user terminals, a community module and a user account number, wherein the user terminals are provided with a community module and a user account number for access; the service management device logs in the community through the administrator account and periodically acquires dialogue messages in the community, wherein the service management device is provided with a dialogue information acquisition part, a dialogue identification part, an emotion score calculation part, a session emotion calculation part, a marketing content ordering acquisition part and a service side communication part.
Description
Technical Field
The invention belongs to the field of customer relationship management, and particularly relates to a service quality optimization management system based on a community system.
Background
Modern enterprises survive in the market, and are a very important ring for the management of own commodities, and the enterprises can control whether the commodities in the market are welcome by consumers or not through various means, so that the competitiveness of the enterprises in the market is enhanced.
With the development of social media, some enterprises can interact with consumers by establishing online communities, for example, service providers of infant products can enable staff to provide the consumers with problems related to infant care and push products related to infants through communities. The data show that the support of the consumers can be improved by 25% by simply replying to the complaints of the consumers on social media, so that the mode is not only beneficial to improving the satisfaction of the consumers to the enterprises, but also beneficial to popularizing the awareness of the enterprises and the products thereof. Even be favorable to the consumer to know in time that daily is careless or that there is no channel to acquire.
However, enterprises only increase consumer satisfaction through communities, and messages of communities are not well utilized. Many conversation messages in the community messages are mutually doped, and effective utilization of the conversation messages is difficult.
Disclosure of Invention
In order to solve the problems, the invention provides a service quality optimization management system for analyzing dialogue messages between staff and users based on a community system, and further optimizing recognition of popularity of marketing information through emotion analysis, so that the staff can push more popular marketing information to improve service quality, and the invention adopts the following technical scheme:
the invention provides a service quality optimization management system based on a community system, which is characterized by comprising the following steps: a service management device; each employee terminal is provided with an employee account for accessing the community module, each employee account corresponds to one community, and the employee account is used for enabling employees to log into the communities through the employee account and provide inquiry and answer marketing service information for users; the system comprises a plurality of user terminals, a community module and a user terminal, wherein the user terminals are provided with user accounts for accessing the community module and are used for enabling users to log into communities through the user accounts and acquiring inquiry and answer marketing service information provided by staff; the service management device logs in the communities through the administrator account numbers and periodically acquires dialogue messages in the communities, wherein the service management device is provided with a dialogue information acquisition part, a dialogue identification part, an emotion calculation part, a session emotion calculation part, a marketing content ordering acquisition part and a service side communication part, the dialogue information acquisition part periodically accesses the communities through the administrator account numbers and acquires dialogue message flows which correspond to all communities and are composed of a plurality of dialogue messages, the dialogue identification part carries out dialogue identification on all dialogue messages in the dialogue message flows so as to acquire a plurality of dialogue messages which are related to content types and contain a plurality of corresponding dialogue messages, the dialogue is divided into service dialogue and marketing dialogue, the emotion calculation part sequentially carries out emotion calculation on all dialogue messages corresponding to all dialogue messages and obtains the message emotion score of each dialogue message, and the session emotion calculation part calculates the marketing emotion score corresponding to each dialogue according to the dialogue message of a corresponding user in each dialogue and the corresponding message emotion score, and the marketing content ordering acquisition part carries out marketing content ordering according to the marketing message score corresponding to all the dialogue messages corresponding to each dialogue and the corresponding to all emotion message types and sends the marketing content ordering recommendation to the marketing terminal as a plurality of marketing recommendation with low score and no marketing content recommendation.
The community system-based service quality optimization management system provided by the invention can be further characterized by comprising a staff terminal and a staff side input display part, wherein the staff terminal is provided with a staff side picture storage part and a staff side input display part, the staff side picture storage part is used for storing a marketing content viewing picture, and the staff side input display part is used for displaying the marketing content viewing picture and enabling staff to view and push recommended marketing content when the staff terminal receives recommended marketing content and non-recommended marketing content.
The community system-based service quality optimization management system provided by the invention can also have the technical characteristics that the service management device is further provided with a user message statistics part, the user message statistics part sequentially counts the number of user accounts for sending dialogue messages in each marketing session so as to obtain the number of user replies corresponding to each marketing session, and the marketing content ordering acquisition part also eliminates marketing sessions with the number of user replies lower than a certain number from an ordering sequence according to the number of user replies when ordering the marketing sessions.
The service quality optimization management system based on the community system provided by the invention can be further characterized by further comprising an average emotion score calculation part and a service quality sorting part, wherein the session emotion calculation part also calculates service session emotion scores corresponding to each service session according to dialogue messages corresponding to user accounts in each session and corresponding message emotion scores, the average emotion score calculation part calculates the average value of the service session emotion scores corresponding to each employee account according to the service session emotion scores so as to obtain service session average emotion scores, the service quality sorting part sorts the employee accounts according to the service session average emotion scores so as to form a service quality sequence, and the service side communication part sends the service quality sequence to all employee terminals.
The service quality optimization management system based on the community system provided by the invention can also have the technical characteristics that the dialogue message comprises dialogue texts and time stamps, and the dialogue recognition part performs dialogue recognition by the following method: identifying the start of three types of conversations and contextual clustering the conversational text based on a time window to obtain a plurality of conversations, wherein one contextual conversational text m j Time window W (m) j ) Is that;
in the formula (1), t i For dialogue text m i Time stamp, t j For dialogue text m j The value of D is 6 or 12 hours; each dialog text m contains three contexts, namely an author context C A (m) refer to context C C (m) and temporal context C T (m) word vector of dialogue text m is denoted v (m), using dialogue text m i Extends the dialog text m in three contexts i The dialog text m i Is an extended representation m of i ′ The method comprises the following steps:
in the formula (2), the meaning of each parameter is: p (P) A (d ij ) Dialog text m j Belonging to the author context C A (m i ) Time of dayDialog text m i With dialog text m j Probability of belonging to the same session; p (P) C (d ij ) Dialog text m j Belonging to reference context C C (m i ) Time dialogue text m i With dialog text m j Probability of belonging to the same session; p (P) T (d ij ) Dialog text m j Belonging to time context C T (m i ) Time dialogue text m i With dialog text m j Probability of belonging to the same session; alpha-the weight of the content of the dialog text m in the expanded representation; lambda (lambda) A -relative weight of author context; lambda (lambda) C -relative weights of the contexts; lambda (lambda) T -relative weights of time contexts, λ A +λ C +λ T =1; further, text clustering is carried out on the dialogue text m, the start of three types of sessions is marked as a clustering prototype T, and the dialogue text m is calculated i And dialog text m j Text approximation sim (m i ，m j )：
For the remaining other dialog text m j Calculating similarity sim (m j ，T)：
In the formula, if
Dialog text m j The belonging cluster is->
If->
Then m is j Not belonging to the dialog to be identified herein, t thresh Is m j Belonging to->
The dialog text m corresponding to each clustering prototype T can be obtained through formulas (4) and (5), and the dialog text m to which each clustering prototype T belongs is sequentially combined according to each clustering prototype T so as to form each session.
The service quality optimization management system based on the community system provided by the invention can also have the technical characteristics that the author context C A (m) is composed of all the dialog texts of which the dialog text m belongs to the same author a as the dialog text m:
wherein N represents a probability density function of normal distribution, μ a Mean value of probability density function representing normal distribution, set to 0, sigma a Representing standard deviation of normal distribution probability density function belonging to author context, referring to context C C (m) is composed of all other dialog texts of author a mentioned by dialog text m and all other dialog texts of author a mentioned by dialog text m:
wherein M is a All dialog text representing author a, at dialog text m j Belonging to reference context C C (m) at the time of dialogue text m i With dialog text m j Probability P belonging to the same session C (d ij ) The method comprises the following steps:
wherein mu is c Mean value of probability density function representing normal distribution, set to 0, sigma c Representing standard deviation of probability density functions belonging to normal distribution referring to context, time context C T (m) is made up of all other sentences except the dialog text m:
C T (m)＝M\m (10)
where M represents all dialog text, where M is the dialog text j Belonging to time context C T (m) dialog text m i With dialog text m j Probability P belonging to the same session T (d ij ) The method comprises the following steps:
wherein mu is T Mean value of probability density function representing normal distribution, set to 0, sigma T Representing the standard deviation of the normal distribution probability density function belonging to the temporal context.
The service quality optimization management system based on the community system provided by the invention can also have the technical characteristics that the dialogue message at least contains dialogue text, and the emotion score calculation part comprises: the text word segmentation unit is used for respectively segmenting the dialogue text of each dialogue message; the word segmentation emotion segmentation calculation unit is used for carrying out emotion analysis on each word segment according to a pre-stored emotion dictionary so as to obtain emotion scores of each word segment; and the message emotion score setting unit is used for setting the message emotion score of each dialogue message according to the sum of emotion scores of all the segmentation words in the dialogue message.
The community system-based service quality optimization management system provided by the invention can also have the technical characteristics that the emotion dictionary comprises a word emotion dictionary and an expression emotion dictionary.
The service quality optimization management system based on the community system provided by the invention can also have the technical characteristics that the session emotion calculating part calculates the session emotion score of each session total session emotion score of test_Senti when calculating the service session emotion score and marketing session emotion time sharing it The method comprises the following steps:
wherein C represents the total number of each session, J represents the number of users, and Cu_Senti ijtc The message emotion score of the user in the conversation process c of the employee i and the user j is once in the t week, wherein i represents one employee, j represents one consumer, t represents one week, and c represents a certain conversation in one week.
The actions and effects of the invention
According to the community system-based service quality optimization management system, the dialogue identification part is provided, so that the dialogue information flow acquired from the communities by the dialogue information acquisition part can be identified, and the service session and the marketing session can be obtained. The conversation emotion calculating part is further provided with an emotion score calculating part for carrying out emotion recognition on each conversation message in the conversation message stream, so that the conversation emotion calculating part can further calculate conversation emotion scores of each conversation according to the message emotion scores of the conversation messages, and analysis of emotion of users and staff in communities is realized. And the marketing content ordering and acquiring part is used for ordering the marketing session according to the session emotion score of the marketing session, so that the marketing content popular with users in a period of time can be obtained more simply and accurately. By extracting the popular marketing contents, the invention enables staff and enterprises to better know the user preference, thereby improving the service quality of the staff and being beneficial to the enterprises to grasp the whole market to a certain extent through communities.
Drawings
FIG. 1 is a block diagram of a quality of service optimization management system in an embodiment of the present invention;
FIG. 2 is a block diagram showing the construction of a service management apparatus in an embodiment of the present invention;
FIG. 3 is a block diagram of an employee terminal in an embodiment of the invention;
fig. 4 is a block diagram of a user terminal in an embodiment of the present invention;
FIG. 5 is a block diagram of an administrator terminal in an embodiment of the present invention; and
fig. 6 is a flow chart of a quality of service optimization management process in an embodiment of the invention.
Detailed Description
In order to make the technical means, creation characteristics, achievement purposes and effects of the present invention easy to understand, the community system-based quality of service optimization management system of the present invention is specifically described below with reference to the embodiments and the accompanying drawings.
< example >
Fig. 1 is a block diagram of a quality of service management system in an embodiment of the present invention.
As shown in fig. 1, the quality of service optimization management system 100 includes a service management apparatus 1, a plurality of employee terminals 2, a plurality of user terminals 3, and an administrator terminal 4.
The service management device 1 is held by an organization that needs to manage services, employee terminals 2 are held by employees having different employee accounts, user terminals 3 are held by users having different user accounts, and an administrator terminal 4 is held by an administrator of the organization.
In this embodiment, the user terminals 3 are all smart mobile devices (e.g., smart phones) held by the corresponding personnel, and the administrator terminal 4 and the service management apparatus 1 are each operation terminals (e.g., computers) connected in communication with each other. In other embodiments, the respective constituent elements of the administrator terminal 4 and the service management apparatus 1 may be part of a server held by an organization.
In this embodiment, the employee terminal 2, the user terminal 3, and the administrator terminal 4 are all provided with a community module, which is a software application corresponding to a community system (for example, a system providing community communication such as WeChat, QQ, etc. and provided with a community communication server 200), and each terminal may communicate with the community communication server 200 through its own community module (for example, by accessing the community communication server 200 through a communication network). The terminals use the community module to log in the community through the account numbers so as to mutually send and receive the question-answer marketing service information (namely, the dialogue information of the staff providing the question-answer marketing service).
In this embodiment, each employee terminal, user terminal, and administrator terminal has an employee account, a user account, and an administrator account, respectively. Each employee account corresponds to a community one by one, and the employee logs into the community through the employee account and sends inquiry and answer marketing service information to users in the community; each community corresponds to a plurality of user accounts, and a user accesses the corresponding community through the user account and sends a question message to staff in the community or obtains question and answer marketing service information; only one administrator account can be used to log into all communities and receive session messages therein. In such a community structure, users (or consumers) typically receive service messages and make questions in the community as members, and employees typically take charge of daily community management as a community master of the community and communicate with the users and provide question-answering services, and an administrator account is typically used for the service management apparatus 1 to obtain rights to access each community.
Fig. 2 is a block diagram of a service management apparatus in an embodiment of the present invention.
As shown in fig. 2, the service management device 1 includes a dialogue information acquisition unit 11, a dialogue identification unit 12, an emotion score calculation unit 13, a dialogue emotion calculation unit 14, a user message statistics unit 15, a marketing content ranking acquisition unit 16, an average emotion score calculation unit 17, a quality of service ranking unit 18, a service-side communication unit 19, and a service-side control unit 110.
The service-side communication unit 19 is for performing data communication between the respective components of the service management device 1 and between the service management device 1 and other terminals, and the service-side control unit 110 includes a computer program for controlling operations of the respective components of the service management device 1.
The dialogue information acquiring unit 11 is configured to access the community module through the administrator account periodically and acquire a dialogue message stream corresponding to the community.
In this embodiment, the session information acquiring unit 11 can send an acquisition request to the administrator terminal 4 through the service-side communication unit 19, so that the administrator terminal 4 automatically logs into each community through the administrator account number and acquires session messages generated by the interaction between the staff members and the users in each community.
In the present embodiment, the periodic setting values periodically acquired by the dialogue information acquisition unit 11 are acquired at one-week intervals in the present embodiment, and may be set at one-day intervals, one-month intervals, or the like in other embodiments, depending on the actual situation. The dialogue information acquiring unit 11 has a time judging unit for judging whether or not the current time has reached a periodic set value from the time acquired last time.
In this embodiment, the dialogue message stream is composed of all dialogue messages in a community within a period of time (in this embodiment, a week), and each dialogue message is composed of a dialogue text, a timestamp when the dialogue text is transmitted, a transmission account number (employee account number, user account number) of the dialogue text, and the like.
The dialogue recognizing section 12 is configured to perform dialogue recognition on each piece of dialogue text in the dialogue message stream to acquire a plurality of sessions related to the content type.
In this embodiment, the session is classified into an active service session, a request service session, and a marketing session according to the content type. The active service session is a session generated by staff user interaction after staff actively sends a service message, the request service session is a session generated by staff user interaction after a user sends a question message, and the marketing session is a session generated by staff user interaction after staff sends a marketing message.
In this embodiment, the dialogue identifying unit 12 first identifies the start of the three types of sessions (i.e., the service message actively sent by the employee, the question message sent by the user, and the marketing message sent by the employee, which may be obtained by classifying and identifying by a conventional classifier, for example, a naive bayes classifier), and performs context text clustering on the dialogue text based on a time window, thereby obtaining a plurality of sessions. Specifically, the parameters and calculation formulas involved in the above process are as follows:
A) Context dialogue text m j Time window W (m) j ) Is that;
in the formula (1), t i For dialogue text m i Time stamp, t j For dialogue text m j The value of D is 6 or 12 hours. Dialog text m for all contexts j Are all within a time window W (m j ) Inner, and a second part of the inner part of the outer part of the inner part of.
B) Each dialog text m contains three contexts, namely an author context C A (m) refer to context C C (m) and temporal context C T (m)。
Author context C A (m) is composed of all the dialog texts of which the dialog text m belongs to the same author a as the dialog text m:
dialog text m j Belonging to the author context C A (m) at the time of dialogue text m i With dialog text m j Probability P belonging to the same session A (d ij ) The method comprises the following steps:
in the formula (3), N represents a probability density function of normal distribution, μ a Representing positiveThe mean value of the state distribution probability density function is set to 0, sigma a The standard deviation representing the normal distribution probability density function belonging to the author context, the value of which can be estimated from the training set.
Refer to context C C (m) is made up of all other dialog texts of author a mentioned by dialog text m (e.g. dialog text containing "@ D user") and of author a mentioned by this dialog text m:
In the formula (4), M a Representing all dialog text of author a.
Dialog text m j Belonging to reference context C C (m) at the time of dialogue text m i With dialog text m j Probability P belonging to the same session C (d ij ) The method comprises the following steps:
mu in the formula (5) c Mean value of probability density function representing normal distribution, set to 0, sigma c The standard deviation of the normal distribution probability density function belonging to the reference context is represented.
Time context C T (m) is made up of all other sentences except the dialog text m:
C T (m)＝M\m (6)
in the formula (6), M represents all the dialog texts.
Dialog text m j Belonging to time context C T (m) dialog text m i With dialog text m j Probability P belonging to the same session T (d ij ) The method comprises the following steps:
mu in the formula (7) T Mean value of probability density function representing normal distribution, set to 0, sigma T Representing the standard deviation of the normal distribution probability density function belonging to the temporal context.
C) The word vector of the word wordbedding of the dialog text m, denoted v (m), is used with the dialog text m i Extends the dialog text m in three contexts i The dialog text m i Is expressed m' i The method comprises the following steps:
in the formula (8), the meaning of each parameter is:
P A (d ij ) Dialog text m j Belonging to the author context C A (m i ) Time dialogue text m i With dialog text m j Probability of belonging to the same session;
P C (d ij ) Dialog text m j Belonging to reference context C C (m i ) Time dialogue text m i With dialog text m j Probability of belonging to the same session;
P T (d ij ) Dialog text m j Belonging to time context C T (m i ) Time dialogue text m i With dialog text m j Probability of belonging to the same session;
alpha-the weight of the content of the dialog text m in the expanded representation;
λ A -relative weight of author context;
λ C -relative weights of the contexts;
λ T -relative weights of time contexts, λ A +λ C +λ T ＝1。
Through the formulas (2) to (8), the representation of one sentence can be expanded by using three contexts, thereby being beneficial to dialog recognition. The author context has the effect that the probability that messages published by the same author at adjacent times belong to the same dialogue is high; the reference context means that if someone is @ in the message, then the likelihood that the message of the person being @ belongs to the same conversation is high; temporal context refers to a high likelihood that temporally adjacent messages belong to the same conversation.
D) Further, text clustering is performed on the dialogue text m, the start of three types of dialogues is marked as a clustering prototype T, and the dialogue text m is calculated i And dialog text m j Text approximation sim (m i ，m j )：
For the remaining other dialog text m j Calculating similarity sim (m j ，T)：
In the dialogue text m j The belonged cluster is
If->
Dialog text m j The belonged cluster is
If->
Then m is j Not belonging to the dialog to be identified herein, t thresh Is m j Belonging to->
Is used for the minimum similarity threshold of (2).
And (3) obtaining dialogue texts m corresponding to each clustering prototype T through formulas (4) and (5), and combining the dialogue texts m according to each clustering prototype T in sequence to form each session.
By the above formulas (1) to (11), the conversation recognition portion 12 can divide conversation messages in each conversation message stream into a plurality of conversations.
An example of the identification of a session is shown in the following table:
table 1 session identification example
In table 1, the tags are session tags, and the session messages of the same tag belong to the same session, and the time, sender and content respectively correspond to the time stamp, sending account number and session text of the session message. Wherein session tag 1 is a marketing session and session tags 2, 3, 4 are all request service sessions.
The emotion score calculation section 13 calculates a message emotion score corresponding to each dialogue message for each dialogue message.
In this embodiment, the emotion score calculation section 13 adopts a text-and-emoji expression-based dual-mode emotion recognition method that performs emotion analysis by comparing a conventional text emotion dictionary (chinese LIWC emotion dictionary) and an emotion dictionary (e.g., emoji emotion dictionary "Novak, petra Kralj, et al," set of emojis "PloS one10.12 (2015): e 0144296)") to a dialogue text.
Specifically, emotion score calculation section 13 completes the dual mode emotion recognition method by text word segmentation section 131, word segmentation emotion score calculation section 132, and message emotion score setting section 133.
When emotion score calculation section 13 performs emotion score calculation for one dialogue message, text segmentation section 131 performs segmentation on the dialogue text of the dialogue message to obtain a plurality of segmented words. The word segmentation is divided into text word segmentation (i.e. conventional text word segmentation) and expression word segmentation (i.e. emoji expression sent by the user).
Further, the word segmentation emotion segmentation calculation unit 132 performs emotion analysis on each word segment according to a pre-stored emotion dictionary to obtain emotion scores of each word segment, that is, performs a one-to-one comparison on each word segment through the emotion dictionary to obtain emotion scores of each word segment. Wherein the emotion of a word (e.g. "laugh") is classified as 1 when the word represents a positive emotion, the emotion of the word (e.g. "lacrimation") is classified as-1 when the word represents a negative emotion, and the emotion of the word (e.g. "lachrymal") is classified as 0 when the word does not have emotion.
After the emotion scores of all the partial words are obtained, the message emotion score setting unit 133 sets the message emotion score of the dialogue message based on the sum of emotion scores of all the partial words in the dialogue message. If the sum of all positive and negative emotion scores (net score) in a dialogue message is positive, the emotion of the dialogue message is classified as 1, if the net score is negative, the emotion of the dialogue message is classified as-1, and if the net score is 0, the emotion of the dialogue message is classified as 0.
An example of the calculation of the emotion score for a message is shown in the following table:
table 2 message emotion score example
In Table 2, "[ lacrimation ]" is an emoji expression, and is identified in the emotion dictionary as representing negative emotion, i.e., the corresponding word segmentation emotion is classified as-1.
The session emotion calculating section 14 is configured to calculate a service session emotion score corresponding to each service session and a marketing session emotion score corresponding to each marketing session, based on the dialogue message corresponding to the user account and the corresponding message emotion score in each session.
In this embodiment, the service session emotion and the marketing session emotion score are calculated by the same method, and the average value of message emotion scores of all dialogue messages uttered by the user in a session, that is, the total session emotion score of each session, namely, test_senti, is calculated it The method comprises the following steps:
wherein C represents the total number of each session, J represents the number of users, and Cu_Senti ijtc In a conversation process c of an employee i and a user j in a t week, the emotion degree of the user is divided into message emotion degrees, wherein i represents an employee, j represents a consumer, t represents a week, and c represents a certain session in the week.
The user message statistics unit 15 is configured to count the number of user accounts for transmitting dialogue messages in each marketing session, so as to obtain the number of user replies corresponding to each marketing session.
In this embodiment, the number of user replies can reflect the number of users participating in each marketing session, thereby reflecting the popularity of the marketing session to some extent.
The marketing content ranking obtaining section 16 is configured to rank all marketing sessions according to the marketing session emotion scores and the user reply numbers corresponding to the respective marketing sessions, and obtain a plurality of marketing sessions with higher ranks as recommended marketing content and a plurality of marketing sessions with lower ranks as non-recommended marketing content.
In this embodiment, the marketing content ranking acquiring section 16 ranks all marketing sessions within a period of time (i.e., the acquisition interval time of the dialogue information acquiring section 11) according to their respective marketing session emotion scores (e.g., ranks from small to large according to emotion scores), and eliminates marketing sessions with a small number of user replies (e.g., less than three-person replies) in the ranked sequence, and further acquires recommended marketing content and non-recommended marketing content from the resulting sequence.
In this embodiment, the recommended marketing content and the non-recommended marketing content are dialogue texts of dialogue messages sent by employees in the marketing session (i.e., belonging to a clustered prototype of the marketing session).
In this embodiment, after the marketing content order obtaining section 16 obtains the recommended marketing content and the non-recommended marketing content, the service side communication section 19 transmits the recommended marketing content and the non-recommended marketing content to all employee terminals, thereby allowing the employees to understand the marketing content popular with the users and unpopular with the users and improving the quality of the marketing service.
The average emotion score calculating section 17 is configured to calculate an average value of service session emotion scores corresponding to the employee accounts according to the service session emotion scores, thereby obtaining service session average emotion scores.
In this embodiment, the average emotion of the service session is an average value of emotion scores of service sessions of all service sessions transmitted by each employee account in the community.
The quality of service ordering unit 18 is configured to order employee accounts according to the average emotion score of the service session to form a quality of service sequence.
In this embodiment, the service quality sequence is composed of names (or ids) of a plurality of employee accounts in text form, and after the generation, the service quality sequence is transmitted to all employee terminals 2 through the service side communication unit 19, so that the employee can know the service quality of the employee and other employees.
In the present embodiment, the service-side communication unit 19 transmits the quality of service sequence, the recommended marketing content, and the non-recommended marketing content to the employee terminal 2 via the administrator terminal 4.
Fig. 3 is a block diagram of the structure of an employee terminal in an embodiment of the invention.
As shown in fig. 3, the employee terminal 2 includes an employee side screen storage 21, an employee side input display 22, an employee side communication unit 23, and an employee side control unit 24.
The employee-side communication unit 23 is configured to perform data communication between the respective constituent parts of the employee terminal 2 and between the employee terminal 2 and another terminal, and the employee-side control unit 24 includes a computer program for controlling operations of the respective constituent parts of the employee terminal 2.
The employee side screen storage 21 stores a community display screen, a marketing content display screen, and a service sequence display screen. In this embodiment, the community display screen, the prompt message display screen, and the marketing content display screen are all display screens of the community module.
The community information display picture is used for displaying the dialogue information in the community when the staff logs in the community and displaying the dialogue information in the community in the picture for the staff to check. In this embodiment, the employee may also perform the editing and sending operation of the message through the screen.
The marketing content display screen is used for displaying and allowing staff to view when the staff terminal receives the recommended marketing content and does not recommend the marketing content. In this embodiment, the marketing content display screen is a session interface formed by employee accounts and administrator accounts, and the recommended marketing content is displayed in the session interface in the form of a dialogue message.
The service sequence display screen is used for displaying the service quality sequence when the staff terminal receives the service quality sequence and displaying the service quality sequence in the screen for the staff to view. In this embodiment, the service sequence display screen is also a session interface formed by employee accounts and administrator accounts, and the prompt information is displayed in the session interface in the form of a dialogue message. In other embodiments, the service sequence display can also be displayed in the form of a web page.
The employee-side input display 22 is configured to display the aforementioned images, and allow the employee to perform corresponding human-computer interaction through the images.
In other embodiments, the employee terminal 2 may further have a reminder module for reminding the employee to view when receiving the recommended marketing content, the non-recommended marketing content or the quality of service sequence by displaying a reminder box, generating a reminder tone, etc.
Fig. 4 is a block diagram of a user terminal in an embodiment of the present invention.
As shown in fig. 4, the user terminal 3 includes a user side screen storage unit 31, a user side input display unit 32, a user side communication unit 33, and a user side control unit 34.
The user-side communication unit 33 is for performing data communication between the respective components of the user terminal 3 and between the user terminal 3 and another terminal, and the user-side control unit 34 includes a computer program for controlling the operations of the respective components of the user terminal 3.
The user-side screen storage unit 31 stores a community display screen. In this embodiment, the community display screen is a display screen of the community module.
The community information display picture is used for displaying the dialogue information in the community when the user logs in the community and displaying the dialogue information in the community in the picture for the user to view. In this embodiment, the user may also perform editing and sending operations of the message through the screen.
The user-side input display unit 32 is configured to display the above-mentioned images, and to allow the employee to perform corresponding man-machine interaction through the images.
Fig. 5 is a block diagram of the structure of an administrator terminal in an embodiment of the present invention.
As shown in fig. 5, the administrator terminal 4 includes a login acquisition unit 41, a management-side communication unit 42, and a management-side control unit 43.
The management-side communication unit 42 is for performing data communication between the respective components of the administrator terminal 4 and between the administrator terminal 4 and other terminals, and the management-side control unit 43 includes a computer program for controlling the operations of the respective components of the administrator terminal 4.
In the present embodiment, when the management-side communication unit 42 receives the acquisition request sent by the dialogue information acquisition unit 11, the management-side control unit 43 controls the login acquisition unit 41 to log in each community through the community module and the administrator account, and sequentially uses all dialogue messages in each community in the past period as one dialogue message stream, and further controls the management-side communication unit 42 to send all dialogue message streams to the service management device 1.
In this embodiment, the value of the preset time period is the same as the periodic setting value of the dialogue information acquisition unit 11, that is, the login acquisition unit 41 acquires all dialogue messages within one week.
Fig. 6 is a flow chart of a quality of service management process in an embodiment of the invention.
As shown in fig. 6, the quality of service management process includes the steps of:
step S1, a dialogue information acquisition part 11 accesses a community module through an administrator account and acquires a dialogue message stream of a corresponding community, and then the step S2 is entered;
step S2, the dialogue recognizing unit 12 sequentially performs dialogue recognition on each piece of dialogue text in the dialogue message stream acquired in step S1 to acquire a plurality of sessions of each community, and then proceeds to step S3;
step S3, the emotion score calculation section 13 sequentially calculates emotion scores of each dialogue message in the dialogue message stream obtained in step S1, obtains message emotion scores corresponding to each dialogue message, and then proceeds to step S4;
step S4, the session emotion calculating section 14 calculates a service session emotion score corresponding to each service session and a marketing session emotion score corresponding to each marketing session from the dialogue message corresponding to the user account and the corresponding message emotion score in each session, and then proceeds to step S5;
Step S5, the user message statistics part 15 counts the number of user accounts for sending dialogue messages in each marketing session so as to obtain the number of user replies corresponding to each marketing session, and then the step S6 is carried out;
step S6, the marketing content ordering and acquiring part 16 orders all marketing sessions according to the marketing session emotion scores calculated in the step S4 and the user reply quantity calculated in the step S5, acquires a plurality of marketing sessions with higher ordering as recommended marketing content and a plurality of marketing sessions with lower ordering as non-recommended marketing content, and then proceeds to the step S7;
step S7, the average emotion score calculation section 17 calculates the average emotion score of the service session corresponding to each employee account based on the service session emotion score calculated in step S4, and then proceeds to step S8;
step S8, the service quality sorting part 18 sorts the employee accounts according to the average emotion scores of the service session calculated in the step S7 to form a service quality sequence, and then the step S9 is performed;
in step S9, the service-side communication unit 19 transmits the recommended marketing content and the non-recommended marketing content acquired in step S6 and the quality of service sequence acquired in step S8 to all employee terminals, and then enters an end state.
In the present embodiment, the time judging unit of the session information acquiring unit 11 judges that the periodic setting value is reached, and then acquires the session information stream of each community again and executes the above steps again.
In this embodiment, the steps S5 to S18 are sequentially performed. In other embodiments, the steps S5 to S6 and the steps S7 to S8 may be executed separately and simultaneously by parallel threads, and the result may be transmitted to the employee terminal by the service-side communication unit 19.
Through the process, the organization can monitor and manage the service quality of staff and remind and improve the staff regularly and better.
Example operation and Effect
According to the community system-based service quality optimization management system provided by the embodiment, the dialogue information acquisition part can identify the dialogue information flow acquired from the communities to acquire the service session and the marketing session due to the dialogue identification part. The conversation emotion calculating part is further provided with an emotion score calculating part for carrying out emotion recognition on each conversation message in the conversation message stream, so that the conversation emotion calculating part can further calculate conversation emotion scores of each conversation according to the message emotion scores of the conversation messages, and analysis of emotion of users and staff in communities is realized. And the marketing content ordering and acquiring part is used for ordering the marketing session according to the session emotion score of the marketing session, so that the marketing content popular with users in a period of time can be obtained more simply and accurately. By extracting the popular marketing contents, the invention enables staff and enterprises to better know the user preference, thereby improving the service quality of the staff and being beneficial to the enterprises to grasp the whole market to a certain extent through communities.
In the embodiment, the user message statistics part is provided, so that the number of the users replied in each marketing session can be counted, the marketing content sorting acquisition part can sort the marketing sessions further according to the number of the replies of the users, and therefore misjudgment of some marketing sessions with smaller reply amounts is avoided, and identification of marketing content popular by users is more accurate.
In the embodiment, the service quality sorting part is arranged, so that the service quality of the staff can be monitored according to the reply of the staff, and the enterprise is helped to better know the reply attitude of the staff when the staff provides the service, thereby helping the enterprise to better manage the service quality of the staff.
The above examples are only for illustrating the specific embodiments of the present invention, and the present invention is not limited to the description scope of the above examples.
In an embodiment, the administrator terminal is a stand-alone computer. In other embodiments, each constituent element of the administrator terminal may also be part of the service management server, that is, the login acquisition section is a part of the service management server, and the functions of the management-side communication section and the management-side control section are completed by the service-side communication section and the service-side control section of the service management server.
Claims (7)
1. A community system-based quality of service optimization management system, comprising:
a service management device;
each employee terminal is provided with an employee account for accessing the community module, each employee account corresponds to one community, and the employee accounts are used for enabling employees to log into the communities through the employee accounts and provide inquiry and answer marketing service information for users;
the user terminals are provided with the community module and are provided with user accounts for accessing the community module, and the user terminals are used for enabling the user to log in the community through the user accounts and obtaining question-answer marketing service information provided by the staff;
an administrator terminal provided with the community module and having an administrator account for accessing all the community modules, the service management apparatus logging in the community through the administrator account and periodically acquiring dialogue messages in the community,
wherein the service management device comprises a dialogue information acquisition part, a dialogue recognition part, an emotion score calculation part, a dialogue emotion calculation part, a marketing content order acquisition part and a service side communication part,
The dialogue information acquisition part periodically accesses the community module through the administrator account and acquires a dialogue message stream which corresponds to each community and consists of a plurality of dialogue messages,
the dialogue identification part performs dialogue identification on each dialogue message in the dialogue message stream so as to acquire a plurality of conversations which are related to content types and contain a plurality of corresponding dialogue messages, the conversations are divided into service conversations and marketing conversations, the service conversations are active service conversations and request service conversations, the active service conversations are conversations generated by staff users after the staff actively sends the service message, the request service conversations are conversations generated by staff users after the user sends the question message, the marketing conversations are conversations generated by staff users after the staff sends the marketing message,
the emotion score calculation section sequentially performs emotion score calculation on each of the dialogue messages and obtains message emotion scores corresponding to each of the dialogue messages,
the session emotion calculating section calculates marketing session emotion scores corresponding to the marketing sessions according to the dialogue messages corresponding to the user account numbers in the sessions and the corresponding message emotion scores,
The marketing content ordering and acquiring part orders all the marketing sessions according to the marketing session emotion scores corresponding to the marketing sessions, acquires a plurality of marketing sessions with the top ordering as recommended marketing content and a plurality of marketing sessions with the bottom ordering as non-recommended marketing content,
the service-side communication section transmits the recommended marketing content and the non-recommended marketing content to all the employee terminals,
the conversation message contains conversation text and a timestamp,
the dialogue recognizing unit performs the dialogue recognition by:
identifying the start of three types of said sessions, and contextual text clustering said dialog text based on a time window, thereby obtaining a plurality of said sessions,
wherein a context dialog text m j Is set to be a time window W (m j ) Is that;
in the formula (1), t i For dialogue text m i T j For dialogue text m j The value of D is 6 or 12 hours;
each dialog text m contains three contexts, namely an author context C A (m) refer to context C C (m) and temporal context C T (m)，
The word vector of the word wordbedding of the dialog text m, denoted v (m), is used with the dialog text m i Is expanded by the three said contexts of the dialog text m i The dialog text m i Is expressed m' i The method comprises the following steps:
in the formula (2), the meaning of each parameter is:
P A (d ij ) Dialog text m j Belonging to the author context C A (m i ) Time dialogue text m i With dialog text m j Probability of belonging to the same session;
P C (d ij ) Dialog text m j Belonging to reference context C C (m i ) Time of dayDialog text m i With dialog text m j Probability of belonging to the same session;
P T (d ij ) Dialog text m j Belonging to time context C T (m i ) Time dialogue text m i With dialog text m j Probability of belonging to the same session;
alpha-the weight of the content of the dialog text m in the expanded representation;
λ A -relative weight of the author context;
λ C -said relative weights referring to contexts;
λ T -relative weights of the temporal contexts, λ A +λ C +λ T ＝1；
Further, the text clustering is carried out on the dialogue text m, the start of the three types of the sessions is marked as a clustering prototype T, and the dialogue text m is calculated i And dialog text m j Text approximation sim (m i ，m j )：
For the remaining other dialog text m j Calculating similarity sim (m j ，T)：
In the formula, if
Dialog text m j The belonging cluster is->
If->
Then m is j Not belonging to the dialog to be identified herein, t thresh Is m j Belonging to->
Is used to determine the minimum similarity threshold of (c),
obtaining dialogue texts m corresponding to each clustering prototype T through formulas (4) and (5), combining the dialogue texts m according to each clustering prototype T in turn to form each session,
the session emotion calculating section calculates a total session emotion score, test_senti, of each session when calculating the service session emotion score and the marketing session emotion it The method comprises the following steps:
wherein C represents the total number of each session, J represents the number of users, and Cu_Senti ijtc The message emotion score of the user in the conversation process c of employee i and user j once in week t, wherein i represents an employee, j represents a consumer, t represents a week, and c represents a certain conversation in the week.
2. The community system-based quality of service optimization management system of claim 1, wherein:
wherein the employee terminal has an employee side screen storage unit and an employee side input display unit,
the employee side screen storage stores a marketing content viewing screen,
the staff side input display part is used for displaying the marketing content viewing picture and enabling the staff to view and push the recommended marketing content when the staff terminal receives the recommended marketing content and the non-recommended marketing content.
3. The community system-based quality of service optimization management system of claim 1, wherein:
wherein the service management device further has a user message statistics section,
the user message statistics part sequentially counts the number of the user account numbers for transmitting the dialogue message in each marketing session to obtain the number of user replies corresponding to each marketing session,
and when the marketing content ordering acquisition part orders the marketing sessions, the marketing sessions with the user reply quantity lower than a certain quantity are removed from the ordered sequence according to the user reply quantity.
4. The community system-based quality of service optimization management system of claim 1, wherein:
wherein the service management device further comprises an average emotion score calculation section and a service quality ranking section,
the conversation emotion calculating section further calculates a service conversation emotion score corresponding to each service conversation based on the dialogue message corresponding to the user account in each conversation and the corresponding message emotion score,
the average emotion score calculating part calculates the average value of the service session emotion scores corresponding to the employee accounts according to the service session emotion scores to obtain service session average emotion scores,
The service quality sorting part sorts the employee accounts according to the average emotion scores of the service sessions to form a service quality sequence,
the service side communication unit transmits the service quality sequence to all employee terminals.
5. The community system-based quality of service optimization management system of claim 1, wherein:
wherein the author context C A (m) is made up of all the dialog texts of which the dialog text m belongs to the same author a as the dialog text m:
dialog text m j Belonging to the author context C A (m) at the time of dialogue text m i With the dialog text m j Probability P belonging to the same session A (d ij ) The method comprises the following steps:
wherein N represents a probability density function of normal distribution, μ a Mean value of probability density function representing normal distribution, set to 0, sigma a Represents the standard deviation of the normal distribution probability density function belonging to the author context,
the reference context C C (m) is constituted by all other dialog texts of author a mentioned by said dialog text m and all other dialog texts of author a mentioned by the dialog text m:
wherein M is a All dialog text representing author a,
in the dialogue text m j Belonging to the reference context C C (m) at the time of dialogue text m i With the dialog text m j Probability P belonging to the same session C (d ij ) The method comprises the following steps:
wherein mu is c Mean value of probability density function representing normal distribution, set to 0, sigma c Represents the standard deviation of the normal distribution probability density function belonging to the reference context,
the time context C T (m) is made up of all other sentences except the dialog text m:
C T (m)＝M\m (11)
where M represents all the dialog text,
in the dialogue text m j Belonging to the time context C T (m) dialog text m i With the dialog text m j Probability P belonging to the same session T (d ij ) The method comprises the following steps:
wherein mu is T Mean value of probability density function representing normal distribution, set to 0, sigma T Representing the standard deviation of the normal distribution probability density function belonging to the temporal context.
6. The community system-based quality of service optimization management system of claim 1, wherein:
wherein the dialogue message at least contains dialogue text,
the emotion score calculation section includes:
a text word segmentation unit for segmenting the dialogue text of each dialogue message;
the word segmentation emotion segmentation calculation unit is used for carrying out emotion analysis on each word segment according to a pre-stored emotion dictionary so as to obtain emotion scores of each word segment;
And the message emotion score setting unit is used for setting the message emotion score of each dialogue message according to the sum of the emotion scores of all the word segmentation in the dialogue message.
7. The community system-based quality of service optimization management system of claim 6, wherein:
the emotion dictionary comprises a word emotion dictionary and an expression emotion dictionary.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910664648.1A CN110427620B (en) | 2019-07-23 | 2019-07-23 | Service quality optimization management system based on community system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910664648.1A CN110427620B (en) | 2019-07-23 | 2019-07-23 | Service quality optimization management system based on community system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110427620A CN110427620A (en) | 2019-11-08 |
| CN110427620B true CN110427620B (en) | 2023-04-28 |
Family
ID=68411872
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910664648.1A Active CN110427620B (en) | 2019-07-23 | 2019-07-23 | Service quality optimization management system based on community system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110427620B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112184299A (en) * | 2020-09-23 | 2021-01-05 | 中国建设银行股份有限公司 | Arbitrage user identification method, apparatus, electronic device and medium |
| CN112163075A (en) * | 2020-09-27 | 2021-01-01 | 北京乐学帮网络技术有限公司 | Information recommendation method and device, computer equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110991189A (en) * | 2019-11-27 | 2020-04-10 | 广州快决测信息科技有限公司 | Method and system for generating decision result according to data acquired by acquisition module |
| CN111666502A (en) * | 2020-07-08 | 2020-09-15 | 腾讯科技（深圳）有限公司 | Abnormal user identification method and device based on deep learning and storage medium |
| CN114511374A (en) * | 2021-12-31 | 2022-05-17 | 青岛海尔工业智能研究院有限公司 | Processing method and device for set of customized household appliances, electronic equipment and storage medium |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9536269B2 (en) * | 2011-01-19 | 2017-01-03 | 24/7 Customer, Inc. | Method and apparatus for analyzing and applying data related to customer interactions with social media |
| US11599731B2 (en) * | 2019-10-02 | 2023-03-07 | Oracle International Corporation | Generating recommendations by using communicative discourse trees of conversations |
-
2019
- 2019-07-23 CN CN201910664648.1A patent/CN110427620B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110991189A (en) * | 2019-11-27 | 2020-04-10 | 广州快决测信息科技有限公司 | Method and system for generating decision result according to data acquired by acquisition module |
| CN111666502A (en) * | 2020-07-08 | 2020-09-15 | 腾讯科技（深圳）有限公司 | Abnormal user identification method and device based on deep learning and storage medium |
| CN114511374A (en) * | 2021-12-31 | 2022-05-17 | 青岛海尔工业智能研究院有限公司 | Processing method and device for set of customized household appliances, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110427620A (en) | 2019-11-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Han et al. | The importance of interactions between content characteristics and creator characteristics for studying virality in social media | |
| US10445351B2 (en) | Customer support solution recommendation system | |
| Serrano et al. | The rise of Germany's AfD: A social media analysis | |
| Barberá | Less is more? How demographic sample weights can improve public opinion estimates based on Twitter data | |
| US10147037B1 (en) | Method and system for determining a level of popularity of submission content, prior to publicizing the submission content with a question and answer support system | |
| CN109308357B (en) | Method, device and equipment for obtaining answer information | |
| EP3168801A1 (en) | Electronic meeting intelligence | |
| EP3169060A1 (en) | Electronic meeting intelligence | |
| WO2021104385A1 (en) | Method and system for online data collection | |
| US20080059447A1 (en) | System, method and computer program product for ranking profiles | |
| KR20160055930A (en) | Systems and methods for actively composing content for use in continuous social communication | |
| CN112733042B (en) | Recommendation information generation method, related device and computer program product | |
| CN107918618B (en) | Data processing method and device | |
| CN110474944B (en) | Network information processing method, device and storage medium | |
| CN106921795B (en) | Contact data management method and system | |
| Fox et al. | Olympians on Twitter: a linguistic perspective of the role of authenticity, clout, and expertise in social media advertising | |
| WO2019226375A1 (en) | Personalized query formulation for improving searches | |
| WO2023287910A1 (en) | Intelligent task completion detection at a computing device | |
| CN110427620B (en) | Service quality optimization management system based on community system | |
| CN110414999B (en) | Service quality management system based on community system | |
| CN112995719B (en) | Bullet screen text-based problem set acquisition method and device and computer equipment | |
| Schlogl | Digital activism and the global middle class: Generation hashtag | |
| US20230334314A1 (en) | Content recommendation method and apparatus, device, storage medium, and program product | |
| CN111933128A (en) | Method and device for processing question bank of questionnaire and electronic equipment | |
| Lee et al. | Anchoring in the past, tweeting from the present: Cognitive bias in journalists’ word choices |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |