CN104133817A - Online community interaction method and device and online community platform - Google Patents

Online community interaction method and device and online community platform Download PDF

Info

Publication number
CN104133817A
CN104133817A CN201310159377.7A CN201310159377A CN104133817A CN 104133817 A CN104133817 A CN 104133817A CN 201310159377 A CN201310159377 A CN 201310159377A CN 104133817 A CN104133817 A CN 104133817A
Authority
CN
China
Prior art keywords
quizmaster
answer
similarity
historical
user
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
Application number
CN201310159377.7A
Other languages
Chinese (zh)
Inventor
程刚
林松
姜爱荣
潘璇
庄子明
赵艳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Shiji Guangsu Information Technology Co Ltd
Original Assignee
Shenzhen Shiji Guangsu Information Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Shiji Guangsu Information Technology Co Ltd filed Critical Shenzhen Shiji Guangsu Information Technology Co Ltd
Priority to CN201310159377.7A priority Critical patent/CN104133817A/en
Publication of CN104133817A publication Critical patent/CN104133817A/en
Pending legal-status Critical Current

Links

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
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/332—Query formulation
    • G06F16/3329—Natural language query formulation or dialogue systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Business, Economics & Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Primary Health Care (AREA)
  • Marketing (AREA)
  • Human Resources & Organizations (AREA)
  • Tourism & Hospitality (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Human Computer Interaction (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention relates to an online community interaction method and device and an online community platform. The method comprises the following steps: digging and classifying a current question raised by a questioner on the online community platform to obtain classification information; according to the classification information, obtaining historical questions corresponding to classification; calculating a similarity between the historical questions and the current question, calculating to obtain the similarity between answers corresponding a preset quantity of historical question which exhibit the higher similarity with the question and the current question, and taking the answer corresponding to the historical question which exhibits the highest similarity as a smart recommendation answer; and pushing the smart recommendation answer to the questioner. The questioner can quickly obtain the answer of the question, and the question matching accuracy of an ask-answer community and user activeness are improved. In addition, the user can obtain the answer of the question through on-line help seeking so as to further improve question answering efficiency, interaction among users can be enhanced, and the user activeness of ask-answer community is further improved.

Description

Web Community's exchange method, device and Web Community's platform
Technical field
The present invention relates to Internet technical field, relate in particular to a kind of Web Community exchange method, device and Web Community's platform.
Background technology
In internet, some Ask-Answer Communities (such as qq asks, microblogging community, friend's net etc.) can provide the service of question and answer class, and user can propose the problem of oneself in Ask-Answer Community, provides corresponding answer by Ask-Answer Community.For the Internet community, it is asked theme and contains every fields such as women, amusement, automobile, physical culture, culture, life, society, current events, history, literature, emotion, tourism, constellation.
Conventionally user wishes, after asking a question, can be answered fast.At present, although some Ask-Answer Communities can be recommended answer fast, be mostly to carry out keyword coupling according to problem stem, its matching accuracy rate is low, make user satisfaction also lower, and problem coverage rate is also lower.
Summary of the invention
The fundamental purpose of the embodiment of the present invention is to provide a kind of Web Community exchange method, device and Web Community's platform, is intended to improve Ask-Answer Community problem answers matching accuracy rate and user's liveness.
In order to achieve the above object, the embodiment of the present invention proposes a kind of Web Community exchange method, comprising:
The current problem that quizmaster is proposed at Web Community's platform is excavated classification, obtains classified information;
According to described classified information, obtain the historical problem of corresponding classification;
The similarity of described historical problem and current problem is obtained in calculating, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer;
Described intelligent recommendation answer is pushed to quizmaster.
The embodiment of the present invention also proposes a kind of Web Community interactive device, comprising:
Sort module, excavates classification for the current problem that quizmaster is proposed at Web Community's platform, obtains classified information;
Historical problem acquisition module, for according to described classified information, obtains the historical problem of corresponding classification;
Calculate acquisition module, for calculating the similarity of obtaining described historical problem and current problem, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer;
Pushing module, for being pushed to quizmaster by described intelligent recommendation answer.
The embodiment of the present invention also proposes a kind of Web Community platform, comprises device as above.
A kind of Web Community exchange method, device and Web Community's platform that the embodiment of the present invention proposes, excavate classification by the current problem that quizmaster is proposed at Web Community's platform, obtains classified information; According to described classified information, obtain the historical problem of corresponding classification; The similarity of described historical problem and current problem is obtained in calculating, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, answer corresponding historical problem the highest correlativity is pushed to quizmaster as intelligent recommendation answer, quizmaster's quick obtaining problem answers be can make thus, Ask-Answer Community problem answers matching accuracy rate and user's liveness improved; In addition, can also make user by the acquisition problem answers of seeking help online, further improve and putd question to the answer efficiency of problem, and can promote the interaction between user, further improve Ask-Answer Community user's liveness.
Brief description of the drawings
The schematic flow sheet of exchange method the first embodiment of Tu1Shi Web Community of the present invention;
Fig. 2 a is that the embodiment of the present invention is shown an instantiation schematic diagram recommending answer to user;
Fig. 2 b calculates the similarity of obtaining described historical problem and current problem in the embodiment of the present invention, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, the schematic flow sheet using answer corresponding historical problem the highest correlativity as intelligent recommendation answer;
The schematic flow sheet of exchange method the second embodiment of Tu3Shi Web Community of the present invention;
Fig. 4 a is the schematic flow sheet that pushes online recourse in the embodiment of the present invention to quizmaster;
Fig. 4 b is a kind of example schematic of selecting recourse in the embodiment of the present invention;
The schematic flow sheet of exchange method the 3rd embodiment of Tu5Shi Web Community of the present invention;
The schematic flow sheet of exchange method the 4th embodiment of Tu6Shi Web Community of the present invention;
The structural representation of interactive device the first embodiment of Tu7Shi Web Community of the present invention;
Fig. 8 is a kind of structural representation that calculates acquisition module in the embodiment of the present invention;
Fig. 9 is the another kind of structural representation that calculates acquisition module in the embodiment of the present invention;
Figure 10 is a kind of structural representation of pushing module in the embodiment of the present invention;
Figure 11 is the structural representation of interactive device the second embodiment of Web Community of the present invention.
In order to make technical scheme of the present invention clearer, clear, be described in further detail below in conjunction with accompanying drawing.
Embodiment
Solution for embodiment of the invention is mainly: excavate classification by the current problem that quizmaster is proposed at Web Community's platform, obtain classified information; According to classified information, obtain the historical problem of corresponding classification; The similarity of historical problem and current problem is obtained in calculating, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of problem similarity rank and the correlativity of current problem, answer corresponding historical problem the highest correlativity is pushed to quizmaster as intelligent recommendation answer, quizmaster's quick obtaining problem answers be can make thus, Ask-Answer Community problem answers matching accuracy rate and user's liveness improved; In addition, can also make user by the acquisition problem answers of seeking help online, further improve the answer efficiency of enquirement problem, and can promote the interaction between user.
As shown in Figure 1, first embodiment of the invention proposes a kind of Web Community exchange method, comprising:
Step S101, the current problem that quizmaster is proposed at Web Community's platform is excavated classification, obtains classified information;
Wherein, Web Community's platform refers to the community platform that question and answer class service is provided for user, such as qq asks or other have question and answer class community or the product of " put question to-answer " feature.
Quizmaster refers to the people who asks a question in community asking.
For the problem that quizmaster is proposed is recommended answer fast, the current problem that first the present embodiment proposes for quizmaster, according to the problem stem of problem (being problem title) and problem content classify (quizmaster also can oneself amendment classification), to obtain classified information, this classified information comprises the type of current problem, such as toiletries, amusement class etc.
Wherein, the mode that the current problem that quizmaster is proposed is classified can include but not limited to following several: adopt SVM(SupportVectorMachine, vector machine support vector machine), KNN(k-Nearest Neighbor algorithm, the neighbouring node algorithm of K) or the model such as linear classification classify, or the word comprising according to current problem is classified.
Wherein, the process that the word comprising according to current problem is classified is as follows:
In advance various known problems are carried out to participle, and the categorical attribute of the word comprising based on known problem and each word is set up dictionary.Afterwards, when current problem is classified, from current problem, extract the word that it comprises, each word included current problem is mated with dictionary, thereby obtain the corresponding categorical attribute of each word, then carry out corresponding weighted calculation and obtain the classification type of current problem according to the categorical attribute of each word, thus the classified information of obtaining.
Classify by models such as SVM, KNN or linear classifications, prior art has ripe scheme, as follows in this brief introduction:
The basic ideas of KNN algorithm are: after given new text, consider K section text concentrated at training text and this new text nearest (the most similar), according to the classification under the new text of kind judging under this K section text, concrete algorithm steps is as follows:
One, according to characteristic item, training text vector is redescribed in set;
Two, after new text arrives, according to the new text of Feature Words participle, determine the vector representation of new text;
Three, concentrate and select K the text the most similar to new text at training text;
Wherein, determining at present of K value generally adopts and first determines an initial value, then adjusts K value according to the result of experiment test, and general initial value is decided to be hundreds of and arrives between several thousand.
Four, in K the text the most similar to new text, calculate successively the weight of every class, computing formula is as follows:
p ( x ‾ , C j ) = Σ d ‾ i ∈ KNN Sim ( x ‾ , d ‾ i ) y ( d ‾ i , C j )
Wherein, the proper vector that x is new text, Sim (x, di) is similarity, and y (di, Cj) is category attribute function, if di belongs to class Cj, functional value is 1 so, otherwise is 0.
Five, the weight of comparing class, assigns to new text in that classification of weight maximum.
In addition, support vector machine and neural network algorithm are applied also comparatively extensively in Text Classification System, and the basic thought of support vector machine is to use simple linear sorter to divide sample space.For linear inseparable pattern in current feature space, use a kernel function that sample is mapped in a higher dimensional space, sample can linearly be divided.
And neural network algorithm adopts perception algorithm to classify.In this model, classificating knowledge is implicitly stored on the weights of connection, determines weight vector with iterative algorithm.In the time that network output differentiation is correct, weight vector remains unchanged, otherwise the adjustment that increases or reduce, therefore also referred to as rewards and punishments method.
Concrete mode classification is not here described in detail.
Step S102, according to described classified information, obtains the historical problem of corresponding classification;
Obtaining after the classified information of current problem, searching the historical problem of corresponding classification, from historical problem, picking out identical with current problem classification and there is the historical problem that is satisfied with class answer.
Wherein, being satisfied with class answer and referring to the answer of being approved and adopting by the quizmaster of historical problem, or the answer of being set by Web Community's platform, can also be the answer that adopts other modes to obtain, such as the better answer of obtaining from other servers etc.
Above-mentioned historical problem can be obtained from the background server of Web Community's platform, also can obtain from other webservers.
Step S103, the similarity of described historical problem and current problem is obtained in calculating, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer;
Step S104, is pushed to quizmaster by described intelligent recommendation answer.
Finding after the historical problem of corresponding classification, calculate the similarity of historical problem and current problem, the answer that the historical problem of the forward front predetermined quantity of computational problem similarity rank is corresponding and the correlativity of current problem, and find in the satisfied answers of historical problem and current problem correlativity is the highest one as recommending answer to push away to quizmaster.As shown in Figure 2 a, it is for adopting the present embodiment scheme to show an instantiation schematic diagram recommending answer to user.
Particularly, as shown in Figure 2 b, as a kind of embodiment, above-mentioned steps S103 calculates and obtains the similarity of described historical problem and current problem, and obtains intelligent recommendation answer based on described similarity and can comprise:
Step S1031, calculates the similarity between the stem of current problem and the stem of historical problem, obtains stem similarity;
The present embodiment by stem for similarity title_score (question_id) represent, wherein, the unique identification that question_id is current problem.
Step S1032, calculates the similarity between the content of current problem and the content of historical problem, obtains content similarity;
The present embodiment represents the content_score for similarity (question_id) between problem content.
Step S1033, obtains problem similarity according to described stem similarity and the calculating of content similarity;
Here by problem for similarity Question_like_score (question_id) represent,
Problem similarity formula is as follows:
Question_like_score (question_id)=α * title_socre (question_id)+β * content_score (question_id), wherein, α and β are coefficient, and α >0, β >0, alpha+beta=1.
Step S1034, filters out the historical problem of the forward front predetermined quantity of problem similarity rank and corresponding is satisfied with class answer, as Candidate Set;
Select the individual historical problem of front n (n>0) and the answer that Question_like_score (question_id) is the highest and enter Candidate Set, n is that constant is determined on a case-by-case basis.
Step S1035, calculates the correlativity that is satisfied with class answer of historical problem in described current problem and described Candidate Set, obtains correlation results;
In current problem and described Candidate Set, the question_answer_score for correlativity (question_id) that is satisfied with class answer of historical problem represents, the form of expression of its computing formula can have a variety of.
Step S1036, obtains overall relevancy mark according to described problem similarity and the calculating of described correlation results;
Step S1037, is satisfied with class answer as intelligent recommendation answer using corresponding historical problem the highest described overall relevancy mark.
Overall relevancy is Total_score (question_id) expression for mark, and its computing formula is as follows:
Total_score (question_id)=Question_like_score (question_id)+γ * question_answer_score (question_id), wherein, γ is coefficient, and γ ∈ (0,1].
Then, the class answer that is satisfied with of Total_score (question_id) historical problem that score is the highest is recommended to quizmaster as intelligent recommendation answer.
The method of more than calculating title_score (question_id), content_score (question_id), question_answer_score (question_id) Shi Junke employing calculating correlativity obtains.
Wherein, the method for calculating text relevant has a lot, for example, sentence can be carried out to participle, then extract proper vector, cosine value between calculated characteristics vector, calculates current problem stem, content and historical problem stem, content thus, and correlativity between satisfied answers; Or the mode of the public substring that employing calculating text packets contains is calculated correlativity; Or can also utilize tfidf algorithm to come computational problem stem, content and historical problem stem, content, and correlativity between satisfied answers.Be not described in detail at this.
In addition, it should be noted that, above-mentioned overall relevancy mark Total_score (question_id) can recommend to user need to be greater than certain threshold tau time, be if there is no greater than threshold tau be satisfied with answer time, do not carry out intelligent recommendation.The τ >0 here, concrete value is determined on a case-by-case basis.
The present embodiment is by such scheme, and the current problem that quizmaster is proposed at Web Community's platform is excavated classification, obtains classified information; According to described classified information, search the historical problem of corresponding classification; The similarity of described historical problem and current problem is obtained in calculating, and obtains intelligent recommendation answer based on described similarity; Described intelligent recommendation answer is pushed to quizmaster, can makes thus quizmaster's quick obtaining problem answers, improve Ask-Answer Community problem answers matching accuracy rate and user's liveness.
As shown in Figure 3, second embodiment of the invention proposes a kind of Web Community exchange method, on the basis of above-mentioned the first embodiment, after above-mentioned steps S103 or step S104, also comprises:
Step S105, when obtaining less than described intelligent recommendation answer, or when described quizmaster is dissatisfied to the intelligent recommendation answer of recommending, push online recourse to described quizmaster and select for described quizmaster, the online recourse of being selected by described quizmaster provides the answer of current problem to described quizmaster.
Wherein Fig. 3 illustrates the intelligent recommendation answer of recommending is dissatisfied with quizmaster.
The difference of the present embodiment and above-mentioned the first embodiment is, the present embodiment is obtaining less than intelligent recommendation answer, or quizmaster is when dissatisfied to the intelligent recommendation answer of recommending, and quizmaster can seek help and obtain problem answers to online friend or other people.
Particularly, in the present embodiment, described online recourse is for the current community platform of logging in network and have and be satisfied with class answer and maybe can provide and recommend the user of answer, this user can be described quizmaster's affiliated person, such as qq good friend, microblogging good friend etc., can be also other unconnected people.
When Web Community's platform can not find can recommend to quizmaster be satisfied with answer or find be satisfied with answer and do not allow quizmaster be satisfied with time, can allow quizmaster select to seek help to the current online people who is associated.
Particularly, as shown in Fig. 4 a, the step that pushes online recourse to quizmaster in above-mentioned steps S105 can comprise:
Step S1051, the login user mark according to described quizmaster at described Web Community platform, affiliated person's list of obtaining described quizmaster;
Wherein, user ID user is user's unique identification id, can be No. qq etc.
Step S1052 according to described quizmaster's affiliated person's online situation, obtains online association list from described affiliated person's list;
Step S1053, mates described online association list with the user tag of obtaining in advance and classified information, obtain described quizmaster's affiliated person's label and classified information;
Step S1054, mates described quizmaster's affiliated person's label and classified information with described current problem, obtain mating mark;
Step S1055, front M the corresponding affiliated person who described coupling mark is greater than to predetermined value is pushed to quizmaster as online recourse; Described M is natural number.
The present embodiment is sought help and is obtained question answering as example taking the friend such as qq good friend, microblogging good friend to quizmaster, and said process is described in detail.
First, according to logging in No. id(qq or community's account number etc. of quizmaster), obtain quizmaster's list of friends, including but not limited to qq good friend, microblogging good friend etc.
Then screen according to the online situation of quizmaster's friend, obtain online list of friends, and mate by this online list of friends and the user tag obtaining in advance and classified information, obtain friend's quizmaster label and classified information.
Then quizmaster's friend's label and classified information and quizmaster's current problem is mated, and mated mark accordingly.
Wherein, friend's quizmaster label and classified information are as follows with the computation process that quizmaster's current problem Questionk mates mark Match_score:
First, described quizmaster's friend's label is mated with described current problem, calculates the tag match value of obtaining affiliated person (friend):
Tag_score (Question_id)= a mistake! Do not find Reference source., tag (i) represents i label, match (tag (i)) represents the mark that the tag (i) in friend's quizmaster who filters out label can match with Questionk.
Described quizmaster's affiliated person's (friend) classified information is mated with described current problem, calculate the classification and matching value of obtaining affiliated person (friend);
Category _ score (Question_id)= a mistake! Do not find Reference source., category (i) represents i classification, match (category (i)) represents the mark that the category (i) in friend's quizmaster who filters out classification can match with Questionk.
Then obtain described coupling mark Match_score (Question_idk) according to described tag match value and the calculating of classification and matching value:
Match_score (Question_idk)=A*tag_score (Question_id)+B*category_score (Question_id), wherein, A and B are coefficient, and A>0, B>0, A+B=1.
In the present embodiment, setting Match_score (Question_id) mark is not that 0 user thinks answerer.In the time having multiple answerers can answer a question Question_id, with set expression be A1(Question_id), A2(Question_id), A3(Question_id) ....
In the time having multiple answerer in quizmaster's list of friends, preferentially select the maximum front top_n name answerer of Match_score (Question_id) to be pushed to quizmaster's (top_n is positive integer, can adjust according to actual conditions).Quizmaster selects recourse from top_n answerer.
Adopt a kind of example schematic of selecting recourse that the such scheme of the present embodiment realizes as shown in Figure 4 b.
Wherein, quizmaster can select anonymity to seek help to friend, and the friend who is sought help also can select anonymous answer.
In addition, it should be noted that, at every turn seeking help of quizmaster need to consume certain integration, therefore, can limit in some time period of user, can only seek help to the friend of some.For example, each user can only seek help with interior friend to N is individual every day, and wherein, N is natural number, and the present embodiment preferred value is set as N ∈ [5,10].
The present embodiment is by such scheme, and the current problem that quizmaster is proposed at Web Community's platform is excavated classification, obtains classified information; According to described classified information, obtain the historical problem of corresponding classification; The similarity of described historical problem and current problem is obtained in calculating, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer; Described intelligent recommendation answer is pushed to quizmaster, can makes thus quizmaster's quick obtaining problem answers, improve Ask-Answer Community problem answers matching accuracy rate and user's liveness; In addition, obtaining less than intelligent recommendation answer, or when quizmaster is dissatisfied to the intelligent recommendation answer of recommending, can also make user by the acquisition problem answers of seeking help online, further improve the answer efficiency of puing question to problem, and can promote the interaction between user, further improve Ask-Answer Community user's liveness.
As shown in Figure 5, third embodiment of the invention proposes a kind of Web Community exchange method, on the basis of above-mentioned the second embodiment, before above-mentioned steps 101, also comprises:
Step 100, the current problem that quizmaster is proposed at Web Community's platform is carried out advertisement filter processing.
The difference of the present embodiment and above-mentioned the second embodiment is, the present embodiment is before recommending answer for quizmaster, also need the current problem that quizmaster is proposed at Web Community's platform to carry out advertisement filter processing, if judge that user's enquirement is with advertisement character, do not enter the flow process of seeking help, to prevent that commercial paper enquirement from causing user's harassing and wrecking, and reduce the harm such as advertisement.
As shown in Figure 6, fourth embodiment of the invention proposes a kind of Web Community exchange method, on the basis of above-mentioned the 3rd embodiment, before above-mentioned steps 101, also comprises:
Step 90, carries out label and classified excavation according to user's historical information to user, obtains described user's label and classified information.
The difference of the present embodiment and above-mentioned the 3rd embodiment is, the present embodiment is before recommending answer for quizmaster, need to carry out label and classified excavation to user according to user's historical information, obtain all users' label and classified information, obtaining less than intelligent recommendation answer so that follow-up, or quizmaster is dissatisfied to the intelligent recommendation answer of recommending, quizmaster seeks help while obtaining problem answers to online friend or other people, online association list is mated with the above-mentioned all users' that obtain label and classified information, obtain quizmaster's affiliated person's label and classified information, and then obtain online recourse and be pushed to quizmaster.
Particularly, according to user's historical information, user is carried out to label and classified excavation, can adopt and include but not limited to: the historical problem of answering of user, user's micro-blog information, user's search daily record, user's group chat information etc.
The common practices that user is added to label and classified information is: first the content with user id is carried out to participle, then word is carried out to analysis statistically, find out and there is the label of word definite meaning, can representative of consumer as user, such as the field that user is interested or be good at, topic, as cosmetics etc., also such as specific to some brands etc., and according to these contents to user annotation classified information.Concrete label and mode classification are not here described in detail.
The present embodiment, by such scheme, carries out label and classified excavation according to user's historical information to user, obtains user's label and classified information; The current problem that quizmaster is proposed at Web Community's platform is excavated classification, obtains classified information; According to described classified information, obtain the historical problem of corresponding classification; The similarity of described historical problem and current problem is obtained in calculating, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer; Described intelligent recommendation answer is pushed to quizmaster, can makes thus quizmaster's quick obtaining problem answers, improve Ask-Answer Community problem answers matching accuracy rate and user's liveness; In addition, can also be in conjunction with user's label and classified information, make user by the acquisition problem answers of seeking help online, further improve and putd question to the answer efficiency of problem, and can promote the interaction between user, further improve Ask-Answer Community user's liveness.
As shown in Figure 7, first embodiment of the invention proposes a kind of Web Community interactive device, comprising: sort module 201, historical problem acquisition module 202, calculating acquisition module 203 and pushing module 204, wherein:
Sort module 201, excavates classification for the current problem that quizmaster is proposed at Web Community's platform, obtains classified information;
Historical problem acquisition module 202, for according to described classified information, obtains the historical problem of corresponding classification;
Calculate acquisition module 203, for calculating the similarity of obtaining described historical problem and current problem, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer;
Pushing module 204, for being pushed to quizmaster by described intelligent recommendation answer.
Wherein, Web Community's platform refers to the community platform that question and answer class service is provided for user, such as qq asks or other have question and answer class community or the product of " put question to-answer " feature.
Quizmaster refers to the people who asks a question in community asking.
For the problem that quizmaster is proposed is recommended answer fast, the current problem that first the present embodiment proposes for quizmaster, according to the problem stem of problem (being problem title) and problem content classify (quizmaster also can oneself amendment classification), to obtain classified information, this classified information comprises the type of current problem, such as toiletries, amusement class etc.
Wherein, the mode that the current problem that quizmaster is proposed is classified can include but not limited to following several: adopt the models such as SVM, KNN or linear classification to classify, or the word comprising according to current problem is classified.
Wherein, the process that the word comprising according to current problem is classified is as follows:
In advance various known problems are carried out to participle, and the categorical attribute of the word comprising based on known problem and each word is set up dictionary.Afterwards, when current problem is classified, from current problem, extract the word that it comprises, each word included current problem is mated with dictionary, thereby obtain the corresponding categorical attribute of each word, then carry out corresponding weighted calculation and obtain the classification type of current problem according to the categorical attribute of each word, thus the classified information of obtaining.
Classify by models such as SVM, KNN or linear classifications, prior art has ripe scheme, as follows in this brief introduction:
The basic ideas of KNN algorithm are: after given new text, consider K section text concentrated at training text and this new text nearest (the most similar), according to the classification under the new text of kind judging under this K section text, concrete algorithm steps is as follows:
One, according to characteristic item, training text vector is redescribed in set;
Two, after new text arrives, according to the new text of Feature Words participle, determine the vector representation of new text;
Three, concentrate and select K the text the most similar to new text at training text;
Wherein, determining at present of K value generally adopts and first determines an initial value, then adjusts K value according to the result of experiment test, and general initial value is decided to be hundreds of and arrives between several thousand.
Four, in K the text the most similar to new text, calculate successively the weight of every class, computing formula is as follows:
p ( x ‾ , C j ) = Σ d ‾ i ∈ KNN Sim ( x ‾ , d ‾ i ) y ( d ‾ i , C j )
Wherein, the proper vector that x is new text, Sim (x, di) is similarity, and y (di, Cj) is category attribute function, if di belongs to class Cj, functional value is 1 so, otherwise is 0.
Five, the weight of comparing class, assigns to new text in that classification of weight maximum.
In addition, support vector machine and neural network algorithm are applied also comparatively extensively in Text Classification System, and the basic thought of support vector machine is to use simple linear sorter to divide sample space.For linear inseparable pattern in current feature space, use a kernel function that sample is mapped in a higher dimensional space, sample can linearly be divided.
And neural network algorithm adopts perception algorithm to classify.In this model, classificating knowledge is implicitly stored on the weights of connection, determines weight vector with iterative algorithm.In the time that network output differentiation is correct, weight vector remains unchanged, otherwise the adjustment that increases or reduce, therefore also referred to as rewards and punishments method.
Concrete mode classification is not here described in detail.
Obtaining after the classified information of current problem, searching the historical problem of corresponding classification, from historical problem, picking out identical with current problem classification and there is the historical problem that is satisfied with class answer.
Wherein, being satisfied with class answer and referring to the answer of being approved and adopting by the quizmaster of historical problem, or the answer of being set by Web Community's platform, can also be the answer that adopts other modes to obtain, such as the better answer of obtaining from other servers etc.
Above-mentioned historical problem can be obtained from the background server of Web Community's platform, also can obtain from other webservers.
Finding after the historical problem of corresponding classification, calculate the similarity of historical problem and current problem, the answer that the historical problem of the forward front predetermined quantity of computational problem similarity rank is corresponding and the correlativity of current problem, and find in satisfied answers and current problem correlativity is the highest one as recommending answer to push away to quizmaster.
Particularly, as shown in Figure 8, described calculating acquisition module 203 can comprise: the first similarity calculated 2031, the second similarity calculated 2032, third phase are seemingly spent computing unit 2033, the first screening unit 2034, the first correlativity computing unit 2035, the second correlativity computing unit 2036, the second screening unit 2037, wherein:
The first similarity calculated 2031, for calculating the similarity between the stem of current problem and the stem of historical problem, obtains stem similarity;
The present embodiment by stem for similarity title_score (question_id) represent, wherein, the unique identification that question_id is current problem.
The second similarity calculated 2032, for calculating the similarity between the content of current problem and the content of historical problem, obtains content similarity;
The present embodiment represents the content_score for similarity (question_id) between problem content.
Third phase is seemingly spent computing unit 2033, for obtaining problem similarity according to described stem similarity and the calculating of content similarity;
Here by problem for similarity Question_like_score (question_id) represent.
Problem similarity formula is as follows:
Question_like_score (question_id)=α * title_socre (question_id)+β * content_score (question_id), wherein, α and β are coefficient, and α >0, β >0, alpha+beta=1.
The first screening unit 2034, for filtering out the historical problem of the forward front predetermined quantity of problem similarity rank and corresponding being satisfied with class answer, as Candidate Set;
Select the individual historical problem of front n (n>0) and the answer that Question_like_score (question_id) is the highest and enter Candidate Set, n is that constant is determined on a case-by-case basis.
The first correlativity computing unit 2035, for calculating the correlativity that is satisfied with class answer of described current problem and described Candidate Set historical problem, obtains correlation results;
In current problem and described Candidate Set, the question_answer_score for correlativity (question_id) that is satisfied with class answer of historical problem represents, the form of expression of its computing formula can have a variety of.
The second correlativity computing unit 2036, for obtaining overall relevancy mark according to described problem similarity and the calculating of described correlation results;
The second screening unit 2037, for being satisfied with class answer as intelligent recommendation answer using corresponding historical problem the highest described overall relevancy mark.
Overall relevancy is Total_score (question_id) expression for mark, and its computing formula is as follows:
Total_score (question_id)=Question_like_score (question_id)+γ * question_answer_score (question_id), wherein, γ is coefficient, and γ ∈ (0,1].
Then, the class answer that is satisfied with of Total_score (question_id) historical problem that score is the highest is recommended to quizmaster as intelligent recommendation answer.
The method of more than calculating title_score (question_id), content_score (question_id), question_answer_score (question_id) Shi Junke employing calculating correlativity obtains.
Wherein, the method for calculating text relevant has a lot, for example, sentence can be carried out to participle, then extract proper vector, cosine value between calculated characteristics vector, calculates current problem stem, content and historical problem stem, content thus, and correlativity between satisfied answers; Or the mode of the public substring that employing calculating text packets contains is calculated correlativity; Or can also utilize tfidf algorithm to come computational problem stem, content and historical problem stem, content, and correlativity between satisfied answers.Be not described in detail at this.
As shown in Figure 9, as another kind of embodiment, described calculating acquisition module 203 also comprises:
Judging unit 2038, for judging that whether overall relevancy mark corresponding to described intelligent recommendation answer is greater than preset threshold values, if so, carries out propelling movement step by described pushing module 204; Otherwise do not recommend.
Particularly, above-mentioned overall relevancy mark Total_score (question_id) can recommend to user need to be greater than certain threshold tau time, be if there is no greater than threshold tau be satisfied with answer time, do not carry out intelligent recommendation.The τ >0 here, concrete value is determined on a case-by-case basis.
The present embodiment is by such scheme, and the current problem that quizmaster is proposed at Web Community's platform is excavated classification, obtains classified information; According to described classified information, search the historical problem of corresponding classification; The similarity of described historical problem and current problem is obtained in calculating, and obtains intelligent recommendation answer based on described similarity; Described intelligent recommendation answer is pushed to quizmaster, can makes thus quizmaster's quick obtaining problem answers, improve Ask-Answer Community problem answers matching accuracy rate and user's liveness.
Further, described pushing module 204 also obtains less than described intelligent recommendation answer for working as, or when described quizmaster is dissatisfied to the intelligent recommendation answer of recommending, push online recourse to described quizmaster and select for described quizmaster, the online recourse of being selected by described quizmaster provides the answer of current problem to described quizmaster.
That is to say, obtaining less than intelligent recommendation answer, or quizmaster is when dissatisfied to the intelligent recommendation answer of recommending, quizmaster can seek help and obtain problem answers to online friend or other people.
Particularly, in the present embodiment, described online recourse is for the current community platform of logging in network and have and be satisfied with class answer and maybe can provide and recommend the user of answer, this user can be described quizmaster's affiliated person, such as qq good friend, microblogging good friend etc., can be also other unconnected people.
When Web Community's platform can not find can recommend to quizmaster be satisfied with answer or find be satisfied with answer and do not allow quizmaster be satisfied with time, can allow quizmaster select to seek help to the current online people who is associated.
More specifically, as shown in figure 10, described pushing module 204 can comprise: the first list acquiring unit 2041, the second list acquiring unit 2042, the first matching unit 2043, the second matching unit 2044 and push unit 2045, wherein:
The first list acquiring unit 2041, for the login user mark at described Web Community platform according to described quizmaster, affiliated person's list of obtaining described quizmaster;
The second list acquiring unit 2042 for according to described quizmaster's affiliated person's online situation, obtains online association list from described affiliated person's list;
The first matching unit 2043, for described online association list is mated with the user tag of obtaining in advance and classified information, obtains described quizmaster's affiliated person's label and classified information;
The second matching unit 2044, for described quizmaster's affiliated person's label and classified information are mated with described current problem, obtains mating mark;
Push unit 2045, is pushed to quizmaster for front M the corresponding affiliated person who described coupling mark is greater than to predetermined value as online recourse; Described M is natural number.
Wherein, user ID user is user's unique identification id, can be No. qq etc.
The present embodiment is sought help and is obtained question answering as example taking the friend such as qq good friend, microblogging good friend to quizmaster, and said process is described in detail.
First, according to logging in No. id(qq or community's account number etc. of quizmaster), obtain quizmaster's list of friends, including but not limited to qq good friend, microblogging good friend etc.
Then screen according to the online situation of quizmaster's friend, obtain online list of friends, and mate by this online list of friends and the user tag obtaining in advance and classified information, obtain friend's quizmaster label and classified information.
Then quizmaster's friend's label and classified information and quizmaster's current problem is mated, and mated mark accordingly.
Wherein, friend's quizmaster label and classified information are as follows with the computation process that quizmaster's current problem Questionk mates mark Match_score:
First, described quizmaster's friend's label is mated with described current problem, calculates the tag match value of obtaining affiliated person (friend):
Tag_score (Question_id)= a mistake! Do not find Reference source., tag (i) represents i label, match (tag (i)) represents the mark that the tag (i) in friend's quizmaster who filters out label can match with Questionk.
Described quizmaster's affiliated person's (friend) classified information is mated with described current problem, calculate the classification and matching value of obtaining affiliated person (friend);
Category _ score (Question_id)= a mistake! Do not find Reference source., category (i) represents i classification, match (category (i)) represents the mark that the category (i) in friend's quizmaster who filters out classification can match with Questionk.
Then obtain described coupling mark Match_score (Question_idk) according to described tag match value and the calculating of classification and matching value:
Match_score (Question_idk)=A*tag_score (Question_id)+B*category_score (Question_id), wherein, A and B are coefficient, and A>0, B>0, A+B=1;
In the present embodiment, setting Match_score (Question_id) mark is not that 0 user thinks answerer.In the time having multiple answerers can answer a question Question_id, with set expression be A1(Question_id), A2(Question_id), A3(Question_id) ....
In the time having multiple answerer in quizmaster's list of friends, preferentially select the maximum front top_n name answerer of Match_score (Question_id) to be pushed to quizmaster's (top_n is positive integer, can adjust according to actual conditions).Quizmaster selects recourse from top_n answerer.
In such scheme, quizmaster can select anonymity to seek help to friend, and the friend who is sought help also can select anonymous answer.
In addition, it should be noted that, at every turn seeking help of quizmaster need to consume certain integration, therefore, can limit in some time period of user, can only seek help to the friend of some.For example, each user can only seek help with interior friend to N is individual every day, and wherein, N is natural number, and the present embodiment preferred value is set as N ∈ [5,10].
The present embodiment is by such scheme, and the current problem that quizmaster is proposed at Web Community's platform is excavated classification, obtains classified information; According to described classified information, obtain the historical problem of corresponding classification; The similarity of described historical problem and current problem is obtained in calculating, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer; Described intelligent recommendation answer is pushed to quizmaster, can makes thus quizmaster's quick obtaining problem answers, improve Ask-Answer Community problem answers matching accuracy rate and user's liveness; In addition, obtaining less than intelligent recommendation answer, or when quizmaster is dissatisfied to the intelligent recommendation answer of recommending, can also make user by the acquisition problem answers of seeking help online, further improve the answer efficiency of puing question to problem, and can promote the interaction between user, further improve Ask-Answer Community user's liveness.
As shown in figure 11, second embodiment of the invention proposes a kind of Web Community interactive device, on the basis of above-mentioned the first embodiment, also comprises:
Filtering module 205, carries out advertisement filter processing for the current problem that quizmaster is proposed at Web Community's platform.
The difference of the present embodiment and above-mentioned the second embodiment is, the present embodiment is before recommending answer for quizmaster, also need the current problem that quizmaster is proposed at Web Community's platform to carry out advertisement filter processing, if judge that user's enquirement is with advertisement character, do not enter the flow process of seeking help, to prevent that commercial paper enquirement from causing user's harassing and wrecking, and reduce the harm such as advertisement.
Further, described sort module 201, also for user being carried out to label and classified excavation according to user's historical information, obtains described user's label and classified information.
Before recommending answer for quizmaster, need to carry out label and classified excavation to user according to user's historical information, obtain all users' label and classified information, obtaining less than intelligent recommendation answer so that follow-up, or quizmaster is dissatisfied to the intelligent recommendation answer of recommending, quizmaster seeks help while obtaining problem answers to online friend or other people, online association list is mated with the above-mentioned all users' that obtain label and classified information, obtain quizmaster's affiliated person's label and classified information, and then obtain online recourse and be pushed to quizmaster.
Particularly, according to user's historical information, user is carried out to label and classified excavation, can adopt and include but not limited to: the historical problem of answering of user, user's micro-blog information, user's search daily record, user's group chat information etc.
The common practices that user is added to label and classified information is: first the content with user id is carried out to participle, then word is carried out to analysis statistically, find out and there is the label of word definite meaning, can representative of consumer as user, such as the field that user is interested or be good at, topic, as cosmetics etc., also such as specific to some brands etc., and according to these contents to user annotation classified information.Concrete label and mode classification are not here described in detail.
The present embodiment, by such scheme, carries out label and classified excavation according to user's historical information to user, obtains user's label and classified information; The current problem that quizmaster is proposed at Web Community's platform is excavated classification, obtains classified information; According to described classified information, obtain the historical problem of corresponding classification; The similarity of described historical problem and current problem is obtained in calculating, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer; Described intelligent recommendation answer is pushed to quizmaster, can makes thus quizmaster's quick obtaining problem answers, improve Ask-Answer Community problem answers matching accuracy rate and user's liveness; In addition, can also be in conjunction with user's label and classified information, make user by the acquisition problem answers of seeking help online, further improve the answer efficiency of puing question to problem, and can promote the interaction between user, between good friend, show the achievement (including but not limited to grade, title, honor etc.) in Ask-Answer Community, to strengthen user's sense of accomplishment, thereby further improve Ask-Answer Community user's liveness.
In addition, the embodiment of the present invention also proposes a kind of Web Community platform, and this Web Community's platform comprises the device described in the various embodiments described above, does not repeat them here.
It should be noted that, in this article, term " comprises ", " comprising " or its any other variant are intended to contain comprising of nonexcludability, thereby the process, method, article or the device that make to comprise a series of key elements not only comprise those key elements, but also comprise other key elements of clearly not listing, or be also included as the intrinsic key element of this process, method, article or device.The in the situation that of more restrictions not, the key element being limited by statement " comprising ... ", and be not precluded within process, method, article or the device that comprises this key element and also have other identical element.
The invention described above embodiment sequence number, just to describing, does not represent the quality of embodiment.
Through the above description of the embodiments, those skilled in the art can be well understood to the mode that above-described embodiment method can add essential general hardware platform by software and realize, can certainly pass through hardware, but in a lot of situation, the former is better embodiment.Based on such understanding, the part that technical scheme of the present invention contributes to prior art in essence in other words can embody with the form of software product, this computer software product is stored in a storage medium (as ROM/RAM, magnetic disc, CD), comprise that some instructions (can be mobile phones in order to make a station terminal equipment, computing machine, server, or the network equipment etc.) carry out the method described in each embodiment of the present invention.
The foregoing is only the preferred embodiments of the present invention; not thereby limit the scope of the claims of the present invention; every equivalent structure or flow process conversion that utilizes instructions of the present invention and accompanying drawing content to do; or be directly or indirectly used in other relevant technical field, be all in like manner included in scope of patent protection of the present invention.

Claims (22)

1. Web Community's exchange method, is characterized in that, comprising:
The current problem that quizmaster is proposed at Web Community's platform is excavated classification, obtains classified information;
According to described classified information, obtain the historical problem of corresponding classification;
The similarity of described historical problem and current problem is obtained in calculating, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer;
Described intelligent recommendation answer is pushed to described quizmaster.
2. method according to claim 1, is characterized in that, described according to classified information, and the step of obtaining the historical problem of corresponding classification comprises:
From historical problem, pick out identical with current problem classification and there is the historical problem that is satisfied with class answer; Describedly be satisfied with class answer and refer to the answer of being approved and adopting by the quizmaster of historical problem, or the answer of being set by Web Community's platform.
3. method according to claim 2, is characterized in that, the step that the similarity of described historical problem and current problem is obtained in described calculating comprises:
Calculate the similarity between the stem of current problem and the stem of historical problem, obtain stem similarity;
Calculate the similarity between the content of current problem and the content of historical problem, obtain content similarity;
Obtain problem similarity according to described stem similarity and the calculating of content similarity.
4. method according to claim 3, it is characterized in that, answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem are obtained in described calculating, and answer corresponding historical problem the highest correlativity is comprised as the step of intelligent recommendation answer:
Filter out the historical problem of the forward front predetermined quantity of problem similarity rank and corresponding be satisfied with class answer, as Candidate Set;
The correlativity that is satisfied with class answer of calculating historical problem in described current problem and described Candidate Set, obtains correlation results;
Obtain overall relevancy mark according to described problem similarity and the calculating of described correlation results;
Corresponding historical problem the highest described overall relevancy mark is satisfied with to class answer as intelligent recommendation answer.
5. method according to claim 4, is characterized in that, before the described step that intelligent recommendation answer is pushed to quizmaster, also comprises:
Judge whether overall relevancy mark corresponding to described intelligent recommendation answer is greater than preset threshold values, if so, carry out and push step; Otherwise do not recommend.
6. according to the method described in any one in claim 1-5, it is characterized in that, also comprise:
When obtaining less than described intelligent recommendation answer, or when described quizmaster is dissatisfied to the intelligent recommendation answer of recommending, push online recourse to described quizmaster and select for described quizmaster, the online recourse of being selected by described quizmaster provides the answer of current problem to described quizmaster; Described online recourse for current logined described Web Community platform and had be satisfied with class answer and maybe can provide and recommend the user of answer, this user at least comprises described quizmaster's affiliated person.
7. method according to claim 6, is characterized in that, the affiliated person that described recourse is described quizmaster, and the described step that pushes online recourse to quizmaster comprises:
Login user mark according to described quizmaster at described Web Community platform, affiliated person's list of obtaining described quizmaster;
According to described quizmaster's affiliated person's online situation, from described affiliated person's list, obtain online association list;
Described online association list is mated with the user tag of obtaining in advance and classified information, obtain described quizmaster's affiliated person's label and classified information;
Described quizmaster's affiliated person's label and classified information are mated with described current problem, obtain mating mark;
Front M the corresponding affiliated person who described coupling mark is greater than to predetermined value is pushed to quizmaster as online recourse; Described M is natural number.
8. method according to claim 7, is characterized in that, described quizmaster's affiliated person's label and classified information is mated with described current problem, and the step that obtains mating mark comprises:
Described quizmaster's affiliated person's label is mated with described current problem, calculate the tag match value of obtaining affiliated person;
Described quizmaster's affiliated person's classified information is mated with described current problem, calculate the classification and matching value of obtaining affiliated person;
Obtain described coupling mark according to described tag match value and the calculating of classification and matching value.
9. method according to claim 7, is characterized in that, the described step that pushes online recourse to quizmaster further comprises:
The quantity of the online recourse pushing to described quizmaster is within a predetermined period of time no more than N, and wherein, N is natural number.
10. method according to claim 1, is characterized in that, the described current problem that quizmaster is proposed at Web Community's platform is excavated classification, and the step of obtaining classified information also comprises before:
The current problem that quizmaster is proposed at Web Community's platform is carried out advertisement filter processing.
11. methods according to claim 7, is characterized in that, the described current problem that quizmaster is proposed at Web Community's platform is excavated classification, and the step of obtaining classified information also comprises before:
According to user's historical information, user is carried out to label and classified excavation, obtain described user's label and classified information.
12. 1 kinds of Web Community's interactive devices, is characterized in that, comprising:
Sort module, excavates classification for the current problem that quizmaster is proposed at Web Community's platform, obtains classified information;
Historical problem acquisition module, for according to described classified information, obtains the historical problem of corresponding classification;
Calculate acquisition module, for calculating the similarity of obtaining described historical problem and current problem, and calculate and obtain answer corresponding to the historical problem of the forward front predetermined quantity of described problem similarity rank and the correlativity of current problem, using answer corresponding historical problem the highest correlativity as intelligent recommendation answer;
Pushing module, for being pushed to described quizmaster by described intelligent recommendation answer.
13. devices according to claim 12, is characterized in that,
Described historical problem acquisition module, also for picking out identical with current problem classification and having a historical problem that is satisfied with class answer from historical problem; Describedly be satisfied with class answer and refer to the answer of being approved and adopting by the quizmaster of historical problem, or the answer of being set by Web Community's platform.
14. devices according to claim 12, is characterized in that, described calculating acquisition module comprises:
The first similarity calculated, for calculating the similarity between the stem of current problem and the stem of historical problem, obtains stem similarity;
The second similarity calculated, for calculating the similarity between the content of current problem and the content of historical problem, obtains content similarity;
Third phase is seemingly spent computing unit, for obtaining problem similarity according to described stem similarity and the calculating of content similarity;
The first screening unit, for filtering out the historical problem of the forward front predetermined quantity of problem similarity rank and corresponding being satisfied with class answer, as Candidate Set;
The first correlativity computing unit, for calculating the correlativity that is satisfied with class answer of described current problem and described Candidate Set historical problem, obtains correlation results;
The second correlativity computing unit, for obtaining overall relevancy mark according to described problem similarity and the calculating of described correlation results;
The second screening unit, for being satisfied with class answer as intelligent recommendation answer using corresponding historical problem the highest described overall relevancy mark.
15. devices according to claim 14, is characterized in that, described calculating acquisition module also comprises:
Judging unit, for judging whether overall relevancy mark corresponding to described intelligent recommendation answer is greater than preset threshold values, is if so, carried out and is pushed step by described pushing module; Otherwise do not recommend.
16. according to the device described in any one in claim 12-15, it is characterized in that,
Described pushing module, also obtain less than described intelligent recommendation answer for working as, or when described quizmaster is dissatisfied to the intelligent recommendation answer of recommending, push online recourse to described quizmaster and select for described quizmaster, the online recourse of being selected by described quizmaster provides the answer of current problem to described quizmaster; Described online recourse for current logined described Web Community platform and had be satisfied with class answer and maybe can provide and recommend the user of answer, this user at least comprises described quizmaster's affiliated person.
17. devices according to claim 16, is characterized in that, described pushing module comprises:
The first list acquiring unit, for the login user mark at described Web Community platform according to described quizmaster, affiliated person's list of obtaining described quizmaster;
The second list acquiring unit for according to described quizmaster's affiliated person's online situation, obtains online association list from described affiliated person's list;
The first matching unit, for described online association list is mated with the user tag of obtaining in advance and classified information, obtains described quizmaster's affiliated person's label and classified information;
The second matching unit, for described quizmaster's affiliated person's label and classified information are mated with described current problem, obtains mating mark;
Push unit, is pushed to quizmaster for front M the corresponding affiliated person who described coupling mark is greater than to predetermined value as online recourse; Described M is natural number.
18. devices according to claim 17, is characterized in that, described the second matching unit, also for described quizmaster's affiliated person's label is mated with described current problem, calculates the tag match value of obtaining affiliated person; Described quizmaster's affiliated person's classified information is mated with described current problem, calculate the classification and matching value of obtaining affiliated person; Obtain described coupling mark according to described tag match value and the calculating of classification and matching value.
19. devices according to claim 17, is characterized in that, described push unit is also no more than N for the quantity of the online recourse that pushes to described quizmaster within a predetermined period of time, and wherein, N is natural number.
20. devices according to claim 12, is characterized in that, also comprise:
Filtering module, carries out advertisement filter processing for the current problem that quizmaster is proposed at Web Community's platform.
21. devices according to claim 12, is characterized in that, described sort module, also for user being carried out to label and classified excavation according to user's historical information, obtains described user's label and classified information.
22. 1 kinds of Web Community's platforms, is characterized in that, comprise the device described in any one in claim 12-21.
CN201310159377.7A 2013-05-02 2013-05-02 Online community interaction method and device and online community platform Pending CN104133817A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310159377.7A CN104133817A (en) 2013-05-02 2013-05-02 Online community interaction method and device and online community platform

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310159377.7A CN104133817A (en) 2013-05-02 2013-05-02 Online community interaction method and device and online community platform

Publications (1)

Publication Number Publication Date
CN104133817A true CN104133817A (en) 2014-11-05

Family

ID=51806496

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310159377.7A Pending CN104133817A (en) 2013-05-02 2013-05-02 Online community interaction method and device and online community platform

Country Status (1)

Country Link
CN (1) CN104133817A (en)

Cited By (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104462064A (en) * 2014-12-15 2015-03-25 陈包容 Method and system for prompting content input in information communication of mobile terminals
CN104933125A (en) * 2015-06-10 2015-09-23 百度在线网络技术（北京）有限公司 Recommendation method and device based on trust relationship
CN104991887A (en) * 2015-06-18 2015-10-21 北京京东尚科信息技术有限公司 Information providing method and apparatus
CN105117398A (en) * 2015-06-25 2015-12-02 扬州大学 Software development problem automatic answering method based on crowdsourcing
CN105260478A (en) * 2015-11-10 2016-01-20 沈文策 Network interactive method and device
CN105550361A (en) * 2015-12-31 2016-05-04 上海智臻智能网络科技股份有限公司 Log processing method and apparatus, and ask-answer information processing method and apparatus
WO2016101765A1 (en) * 2014-12-25 2016-06-30 北京奇虎科技有限公司 Question-and-answer page related question recommendation method and device
CN106095941A (en) * 2016-06-14 2016-11-09 华南师范大学 Recommend method and system based on the solution in big data knowledge storehouse
CN106126541A (en) * 2016-06-14 2016-11-16 华南师范大学 Personalized solution based on big data knowledge storehouse recommends method and system
CN106716933A (en) * 2016-11-10 2017-05-24 深圳达闼科技控股有限公司 Message processing method and apparatus, and electronic device
CN106803092A (en) * 2015-11-26 2017-06-06 阿里巴巴集团控股有限公司 A kind of determination method and device of typical problem data
CN107180062A (en) * 2016-03-11 2017-09-19 中国石油化工股份有限公司 Any active ues method for digging and device based on seismic interpretation system
CN107423440A (en) * 2017-08-04 2017-12-01 逸途（北京）科技有限公司 A kind of question and answer context switching based on sentiment analysis is with strengthening system of selection
CN107451199A (en) * 2017-07-05 2017-12-08 阿里巴巴集团控股有限公司 Method for recommending problem and device, equipment
CN107704580A (en) * 2017-09-30 2018-02-16 百度在线网络技术（北京）有限公司 Question and answer method for pushing, device, server and storage medium based on user's period
CN108073664A (en) * 2016-11-11 2018-05-25 北京搜狗科技发展有限公司 A kind of information processing method, device, equipment and client device
CN108920654A (en) * 2018-06-29 2018-11-30 泰康保险集团股份有限公司 A kind of matched method and apparatus of question and answer text semantic
CN109272319A (en) * 2018-08-14 2019-01-25 阿里巴巴集团控股有限公司 Community's mapping and transaction violation community identification method, device, electronic equipment
CN109492076A (en) * 2018-09-20 2019-03-19 西安交通大学 A kind of network-based community's question and answer website answer credible evaluation method
CN109684475A (en) * 2018-11-21 2019-04-26 斑马网络技术有限公司 Processing method, device, equipment and the storage medium of complaint
CN109753561A (en) * 2019-01-16 2019-05-14 长安汽车金融有限公司 A kind of generation method automatically replied and device
CN110019733A (en) * 2017-12-28 2019-07-16 中兴通讯股份有限公司 A kind of intelligent answer method and device of Community-oriented
CN110737754A (en) * 2018-07-02 2020-01-31 中兴通讯股份有限公司 message response method, device, equipment and storage medium
CN111382244A (en) * 2018-12-29 2020-07-07 深圳市优必选科技有限公司 Deep retrieval matching classification method and device and terminal equipment
CN112069299A (en) * 2020-09-03 2020-12-11 三一专用汽车有限责任公司 Response method for question-answer interaction, server, engineering vehicle and readable storage medium
CN112735420A (en) * 2019-10-14 2021-04-30 百度在线网络技术（北京）有限公司 Question and answer method and device based on intelligent sound box, intelligent sound box and medium
CN114125043A (en) * 2020-09-01 2022-03-01 上海智臻智能网络科技股份有限公司 Information pushing method and information pushing device

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1821991A (en) * 2005-02-18 2006-08-23 上海赢思软件技术有限公司 Knowledge question-and-answer quick processing system based on artificial intelligence
CN1991829A (en) * 2005-12-29 2007-07-04 陈亚斌 Searching method of search engine system
CN102654859A (en) * 2011-03-01 2012-09-05 北京彩云在线技术开发有限公司 Method and system for recommending songs
CN102654860A (en) * 2011-03-01 2012-09-05 北京彩云在线技术开发有限公司 Personalized music recommendation method and system
CN103049528A (en) * 2012-12-24 2013-04-17 北京信息科技大学 Personalized web page searching and sorting method on basis of interest vectors of user

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1821991A (en) * 2005-02-18 2006-08-23 上海赢思软件技术有限公司 Knowledge question-and-answer quick processing system based on artificial intelligence
CN1991829A (en) * 2005-12-29 2007-07-04 陈亚斌 Searching method of search engine system
CN102654859A (en) * 2011-03-01 2012-09-05 北京彩云在线技术开发有限公司 Method and system for recommending songs
CN102654860A (en) * 2011-03-01 2012-09-05 北京彩云在线技术开发有限公司 Personalized music recommendation method and system
CN103049528A (en) * 2012-12-24 2013-04-17 北京信息科技大学 Personalized web page searching and sorting method on basis of interest vectors of user

Cited By (44)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104462064B (en) * 2014-12-15 2017-11-03 陈包容 A kind of method and system of information of mobile terminal communication prompt input content
CN104462064A (en) * 2014-12-15 2015-03-25 陈包容 Method and system for prompting content input in information communication of mobile terminals
WO2016101765A1 (en) * 2014-12-25 2016-06-30 北京奇虎科技有限公司 Question-and-answer page related question recommendation method and device
CN104933125B (en) * 2015-06-10 2019-05-03 百度在线网络技术（北京）有限公司 Recommended method and device based on trusting relationship
CN104933125A (en) * 2015-06-10 2015-09-23 百度在线网络技术（北京）有限公司 Recommendation method and device based on trust relationship
CN104991887A (en) * 2015-06-18 2015-10-21 北京京东尚科信息技术有限公司 Information providing method and apparatus
CN104991887B (en) * 2015-06-18 2018-01-19 北京京东尚科信息技术有限公司 The method and device of information is provided
CN105117398B (en) * 2015-06-25 2018-10-26 扬州大学 A kind of software development problem auto-answer method based on crowdsourcing
CN105117398A (en) * 2015-06-25 2015-12-02 扬州大学 Software development problem automatic answering method based on crowdsourcing
CN105260478A (en) * 2015-11-10 2016-01-20 沈文策 Network interactive method and device
CN106803092A (en) * 2015-11-26 2017-06-06 阿里巴巴集团控股有限公司 A kind of determination method and device of typical problem data
CN105550361A (en) * 2015-12-31 2016-05-04 上海智臻智能网络科技股份有限公司 Log processing method and apparatus, and ask-answer information processing method and apparatus
CN105550361B (en) * 2015-12-31 2018-11-09 上海智臻智能网络科技股份有限公司 Log processing method and device and question and answer information processing method and device
CN107180062A (en) * 2016-03-11 2017-09-19 中国石油化工股份有限公司 Any active ues method for digging and device based on seismic interpretation system
CN106126541B (en) * 2016-06-14 2019-07-19 华南师范大学 Personalized solution recommended method and system based on big data knowledge base
CN106126541A (en) * 2016-06-14 2016-11-16 华南师范大学 Personalized solution based on big data knowledge storehouse recommends method and system
CN106095941A (en) * 2016-06-14 2016-11-09 华南师范大学 Recommend method and system based on the solution in big data knowledge storehouse
CN106095941B (en) * 2016-06-14 2020-03-20 华南师范大学 Big data knowledge base-based solution recommendation method and system
CN106716933B (en) * 2016-11-10 2020-11-06 深圳达闼科技控股有限公司 Message processing method and device and electronic equipment
WO2018086040A1 (en) * 2016-11-10 2018-05-17 深圳达闼科技控股有限公司 Message processing method, device, and electronic apparatus
CN106716933A (en) * 2016-11-10 2017-05-24 深圳达闼科技控股有限公司 Message processing method and apparatus, and electronic device
CN108073664A (en) * 2016-11-11 2018-05-25 北京搜狗科技发展有限公司 A kind of information processing method, device, equipment and client device
CN108073664B (en) * 2016-11-11 2021-08-31 北京搜狗科技发展有限公司 Information processing method, device, equipment and client equipment
CN107451199A (en) * 2017-07-05 2017-12-08 阿里巴巴集团控股有限公司 Method for recommending problem and device, equipment
CN107451199B (en) * 2017-07-05 2020-06-26 阿里巴巴集团控股有限公司 Question recommendation method, device and equipment
CN107423440B (en) * 2017-08-04 2020-09-01 逸途（北京）科技有限公司 Question-answer context switching and reinforced selection method based on emotion analysis
CN107423440A (en) * 2017-08-04 2017-12-01 逸途（北京）科技有限公司 A kind of question and answer context switching based on sentiment analysis is with strengthening system of selection
CN107704580A (en) * 2017-09-30 2018-02-16 百度在线网络技术（北京）有限公司 Question and answer method for pushing, device, server and storage medium based on user's period
CN110019733A (en) * 2017-12-28 2019-07-16 中兴通讯股份有限公司 A kind of intelligent answer method and device of Community-oriented
CN108920654A (en) * 2018-06-29 2018-11-30 泰康保险集团股份有限公司 A kind of matched method and apparatus of question and answer text semantic
CN110737754A (en) * 2018-07-02 2020-01-31 中兴通讯股份有限公司 message response method, device, equipment and storage medium
CN109272319B (en) * 2018-08-14 2022-05-31 创新先进技术有限公司 Community mapping and transaction violation community identification method and device, and electronic equipment
CN109272319A (en) * 2018-08-14 2019-01-25 阿里巴巴集团控股有限公司 Community's mapping and transaction violation community identification method, device, electronic equipment
CN109492076A (en) * 2018-09-20 2019-03-19 西安交通大学 A kind of network-based community's question and answer website answer credible evaluation method
CN109684475B (en) * 2018-11-21 2021-03-30 斑马网络技术有限公司 Complaint processing method, complaint processing device, complaint processing equipment and storage medium
CN109684475A (en) * 2018-11-21 2019-04-26 斑马网络技术有限公司 Processing method, device, equipment and the storage medium of complaint
CN111382244A (en) * 2018-12-29 2020-07-07 深圳市优必选科技有限公司 Deep retrieval matching classification method and device and terminal equipment
CN111382244B (en) * 2018-12-29 2023-04-14 深圳市优必选科技有限公司 Deep retrieval matching classification method and device and terminal equipment
CN109753561B (en) * 2019-01-16 2021-04-27 长安汽车金融有限公司 Automatic reply generation method and device
CN109753561A (en) * 2019-01-16 2019-05-14 长安汽车金融有限公司 A kind of generation method automatically replied and device
CN112735420A (en) * 2019-10-14 2021-04-30 百度在线网络技术（北京）有限公司 Question and answer method and device based on intelligent sound box, intelligent sound box and medium
CN112735420B (en) * 2019-10-14 2022-11-11 百度在线网络技术（北京）有限公司 Question and answer method and device based on intelligent sound box, intelligent sound box and medium
CN114125043A (en) * 2020-09-01 2022-03-01 上海智臻智能网络科技股份有限公司 Information pushing method and information pushing device
CN112069299A (en) * 2020-09-03 2020-12-11 三一专用汽车有限责任公司 Response method for question-answer interaction, server, engineering vehicle and readable storage medium

Similar Documents

Publication Publication Date Title
CN104133817A (en) Online community interaction method and device and online community platform
CN103064945B (en) Based on the Situational searching method of body
CN106570708B (en) Management method and system of intelligent customer service knowledge base
US20160048754A1 (en) Classifying resources using a deep network
CN105808590B (en) Search engine implementation method, searching method and device
CN106126582A (en) Recommend method and device
CN113806630B (en) Attention-based multi-view feature fusion cross-domain recommendation method and device
CN112307351A (en) Model training and recommending method, device and equipment for user behavior
CN106844530A (en) Training method and device of a kind of question and answer to disaggregated model
CN103020845A (en) Mobile application pushing method and system
CN109992674B (en) Recommendation method fusing automatic encoder and knowledge graph semantic information
CN106294787A (en) Information pushing method and device and electronic equipment
CN102576438A (en) Method and apparatus for executing a recommendation
CN110110225A (en) Online education recommended models and construction method based on user behavior data analysis
CN111191099B (en) User activity type identification method based on social media
CN112148994B (en) Information push effect evaluation method and device, electronic equipment and storage medium
CN113821592B (en) Data processing method, device, equipment and storage medium
CN109949174A (en) A kind of isomery social network user entity anchor chain connects recognition methods
CN115408618B (en) Point-of-interest recommendation method based on social relation fusion position dynamic popularity and geographic features
CN113011172A (en) Text processing method and device, computer equipment and storage medium
CN104572915B (en) One kind is based on the enhanced customer incident relatedness computation method of content environment
CN113741759B (en) Comment information display method and device, computer equipment and storage medium
Arai et al. Predicting quality of answer in collaborative Q/A community
CN110489665B (en) Microblog personalized recommendation method based on scene modeling and convolutional neural network
Zhu A book recommendation algorithm based on collaborative filtering

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
EXSB Decision made by sipo to initiate substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20141105

RJ01 Rejection of invention patent application after publication