CN109003122A - A kind of user classification method and server launched based on advertisement - Google Patents
A kind of user classification method and server launched based on advertisement Download PDFInfo
- Publication number
- CN109003122A CN109003122A CN201810685006.5A CN201810685006A CN109003122A CN 109003122 A CN109003122 A CN 109003122A CN 201810685006 A CN201810685006 A CN 201810685006A CN 109003122 A CN109003122 A CN 109003122A
- Authority
- CN
- China
- Prior art keywords
- keyword
- evaluation
- user
- target user
- content
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0242—Determining effectiveness of advertisements
- G06Q30/0244—Optimization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0257—User requested
- G06Q30/0258—Registration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0277—Online advertisement
Abstract
The present invention provides a kind of user classification method and server launched based on advertisement, method includes: that the word frequency of multiple evaluation keywords relevant to content is launched and each evaluation keyword is counted from information bank；Target user is obtained about the objective appraisal keyword for launching content；According to the word frequency of the objective appraisal keyword and each evaluation keyword, the target user is calculated about the evaluation of estimate for launching content；According to the target user about the evaluation of estimate for launching content, classify to target user.The present invention realizes the exact classification to target user, improves the precision of advertisement dispensing, reduces invalid dispensing, reduces advertisement and launches cost.
Description
Technical field
The present invention relates to advertisement putting field, espespecially a kind of user classification method and server launched based on advertisement.
Background technique
With the fast development of Internet technology, the web advertisement has been increasingly becoming a kind of mainstream advertisement putting mode, with biography
The features such as system advertising media compares, and Internet advertising wide coverage is at low cost, interactive strong, and it is properer present
People's lives mode.
In the scenes applications such as customer analysis, recommendation marketing analysis and forecast analysis, if it is possible to carry out user accurately
Classification, then carry out accurate user's marketing and advertisement promotion, the growth rate of customer group can be greatlyd improve, while can
Invalid popularization and cost of marketing are reduced, the income of accurate user marketing and advertisement promotion is improved.
However, used method is all to directly acquire the browsing of user during extracting to target user
Data, the associative key by launching main body carry out coarse user and extract, so as to cause the user extracted, much simultaneously
Be not really to launch the interested user of main body, or even be much to launch main body detest user.
To solve the above-mentioned problems, the precision for improving dispensing, the present invention provides a kind of users launched based on advertisement
Classification method and server.
Summary of the invention
The object of the present invention is to provide a kind of user classification methods and server launched based on advertisement, realize and use target
The exact classification at family improves the precision of advertisement dispensing, reduces invalid dispensing, reduces advertisement and launches cost.
Technical solution provided by the invention is as follows:
The present invention provides a kind of user classification methods launched based on advertisement, comprising steps of counting from information bank
The word frequency of multiple evaluation keywords relevant to content is launched and each evaluation keyword；Target user is obtained about the throwing
Put the objective appraisal keyword of content；According to the word frequency of the objective appraisal keyword and each evaluation keyword, calculate
The target user is about the evaluation of estimate for launching content；According to the target user about the evaluation for launching content
Value, classifies to target user.
Preferably, it is counted from information bank and launches the relevant multiple evaluation keywords of content and each evaluation key
The word frequency of word, this step specifically include: evaluation keyword relevant to content is launched is counted from information bank；To the evaluation
Keyword is divided into front evaluation keyword and unfavorable ratings keyword according to classifying rules；Count each front evaluation keyword
The word frequency of word frequency and each unfavorable ratings keyword.
Preferably, according to the word frequency of the objective appraisal keyword and each evaluation keyword, the target is calculated
User specifically includes about the evaluation of estimate for launching content, this step: according to the word frequency of each front evaluation keyword
With the word frequency of each unfavorable ratings keyword, the front evaluation keyword and the unfavorable ratings keyword are returned
One change processing calculates the weighted value of each front evaluation keyword and the weighted value of each unfavorable ratings keyword；According to institute
State the weighted value of objective appraisal keyword and each front evaluation keyword, the weighted value of each unfavorable ratings keyword, meter
Target user is calculated about the evaluation of estimate for launching content.
Preferably, according to the weighted value of the objective appraisal keyword and each front evaluation keyword, each negative
The weighted value for evaluating keyword calculates method of the target user about the evaluation of estimate meter for launching content are as follows:
Wherein, V is target user about the evaluation of estimate for launching content；B is offset parameter；wiIndicate i-th of front
The weight of keyword is evaluated, andIf i-th of front evaluation keyword occurs in the objective appraisal keyword,
By xiIt is denoted as 1, is otherwise denoted as 0；w0jIndicate the weight of j-th of unfavorable ratings keyword, andIf j-th is negatively commented
Valence keyword occurs in the objective appraisal keyword, then by xjIt is denoted as 1, is otherwise denoted as 0.
Preferably, classified about the evaluation of estimate for launching content to target user according to the target user, this
Step specifically includes: if the target user is greater than preset value about the evaluation of estimate for launching content, the target being used
Family is classified as first kind user, and the target user is otherwise classified as the second class user；Or；By all target users according to each
Self-corresponding evaluation of estimate is ranked up, and takes the target user of preset order section as first kind user, other target users are made
For the second class user.
The present invention also provides a kind of servers for user's classification, and server includes: statistical module, are used for from information
The word frequency of multiple evaluation keywords relevant to content is launched and each evaluation keyword is counted in library；Module is obtained, is used
In acquisition target user about the objective appraisal keyword for launching content；Computing module, with the statistical module, described obtain
The electrical connection of modulus block calculates the mesh for the word frequency according to the objective appraisal keyword and each evaluation keyword
User is marked about the evaluation of estimate for launching content；User's categorization module is electrically connected with the computing module, for according to
Target user classifies to target user about the evaluation of estimate for launching content.
Preferably, the statistical module specifically includes: keyword statistic unit, is also used to count and throw from information bank
Put the relevant evaluation keyword of content；Keyword classification unit is electrically connected with the keyword statistic unit, is closed to the evaluation
Keyword is divided into front evaluation keyword and unfavorable ratings keyword according to classifying rules；Word frequency statistics unit, with the keyword
Taxon electrical connection is also used to count the word frequency of each front evaluation keyword and the word frequency of each unfavorable ratings keyword.
Preferably, the computing module is also used to according to the word frequency of each front evaluation keyword and each described
The front evaluation keyword and the unfavorable ratings keyword is normalized in the word frequency of unfavorable ratings keyword,
Calculate the weighted value of each front evaluation keyword and the weighted value of each unfavorable ratings keyword；The computing module, also
For weighted value, each unfavorable ratings keyword according to the objective appraisal keyword and each front evaluation keyword
Weighted value, calculate target user about it is described launch content evaluation of estimate.
Preferably, the computing module evaluates the power of keyword according to the objective appraisal keyword and each front
The weighted value of weight values, each unfavorable ratings keyword calculates method of the target user about the evaluation of estimate for launching content
Are as follows:
Wherein, V is target user about the evaluation of estimate for launching content；B is offset parameter；wiIndicate i-th of front
The weight of keyword is evaluated, andIf i-th of front evaluation keyword occurs in the objective appraisal keyword,
By xiIt is denoted as 1, is otherwise denoted as 0；w0jIndicate the weight of j-th of unfavorable ratings keyword, andIf j-th is negatively commented
Valence keyword occurs in the objective appraisal keyword, then by xjIt is denoted as 1, is otherwise denoted as 0.
Preferably, user's categorization module, if being also used to the target user about the evaluation of estimate for launching content
Greater than preset value, then the target user is classified as first kind user, the target user is otherwise classified as the second class and is used
Family；Or；User's categorization module is also used to for all target users being ranked up according to corresponding evaluation of estimate, take pre-
If the target user of sequential segments is as first kind user, using other target users as the second class user.
There is provided through the invention it is a kind of based on advertisement launch user classification method and server, can bring with down toward
It is few a kind of the utility model has the advantages that
1, the present invention evaluates the analysis of keyword by multipair target user, can accurately classify to user, sentence
Whether disconnected user out is to be appropriate for advertisement dispensing, to realize the accurate dispensing of advertisement, reduces invalid popularization and marketing
Cost improves the income of accurate user marketing and advertisement promotion.
2, the present invention will be divided into front evaluation keyword and unfavorable ratings keyword to the evaluation keyword for launching content, and
It is analyzed by the objective appraisal keyword to target user, the evaluation of estimate being calculated is able to reflect user to dispensing content
Fancy grade, realize to the exact classification of user.
Detailed description of the invention
Below by clearly understandable mode, preferred embodiment is described with reference to the drawings, to it is a kind of based on advertisement launch
Above-mentioned characteristic, technical characteristic, advantage and its implementation of user classification method and server are further described.
Fig. 1 is a kind of flow chart of the one embodiment for the user classification method launched based on advertisement of the present invention；
Fig. 2 is a kind of flow chart of another embodiment of the user classification method launched based on advertisement of the present invention；
Fig. 3 is a kind of structural schematic diagram of one embodiment of the server for user's classification of the present invention.
Drawing reference numeral explanation:
1- statistical module, 11- keyword statistic unit, 12- keyword classification unit, 13- word frequency statistics unit, 2- are obtained
Module, 3- computing module, 4- user's categorization module.
Specific embodiment
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, Detailed description of the invention will be compareed below
A specific embodiment of the invention.It should be evident that drawings in the following description are only some embodiments of the invention, for
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing, and obtain other embodiments.
To make simplified form, part related to the present invention is only schematically shown in each figure, they are not represented
Its practical structures as product.In addition, there is identical structure or function in some figures so that simplified form is easy to understand
Component only symbolically depicts one of those, or has only marked one of those.Herein, "one" is not only indicated
" only this ", can also indicate the situation of " more than one ".
The present invention provides a kind of one embodiment of user classification method launched based on advertisement, as shown in Figure 1, packet
It includes:
S1 counts from information bank multiple evaluation keywords relevant to content is launched and each evaluation keyword
Word frequency；
S2 obtains target user about the objective appraisal keyword for launching content；
S3 calculates the target user according to the word frequency of the objective appraisal keyword and each evaluation keyword
About the evaluation of estimate for launching content；
S4, about the evaluation of estimate for launching content, classifies to target user according to the target user.
Advertisement needs the target user of selected dispensing before dispensing.In the prior art, can by the network data of user,
Judge whether user consults excessively advertisement and launch the relevant information of content, if user browsed data relevant to ad content,
Then the user can be determined as target user.Such putting mode does not distinguish user, has in targeted user population
Many users launch content to advertisement and lose interest in, or even launch content to advertisement and detest very much.If blindly launching, one
It increases many invalid popularizations and cost is launched in advertisement, two is more anti-to the corresponding product of ad content to also result in user
Sense.
The present invention can classify to user during selected target user, such as be divided into interested in dispensing content
User and to launch the uninterested user of content, when carrying out advertisement dispensing, only to dispensing the interested user of content
Advertisement dispensing is carried out, had both achieved the purpose that advertisement was launched in this way, and had decreased invalid dispensing, reduces advertisement dispensing
Cost.
Specifically, can be thrown first by crawler software, related around the business of dispensing when carrying out category filter to user
Content is put, crawls corpus data relevant to business information, and be stored in database.It such as is " flourish for ad placement service
Shine 10 mobile phones " advertisement launch, the relevant evaluation on Baidu's discussion bar about " 10 mobile phone of honor " can be crawled by crawler software,
And these are evaluated and is expected in deposit information bank.
Then multiple evaluation keywords relevant to " 10 mobile phone of honor " are counted in information bank and each evaluation is closed
The word frequency of keyword, such as the data counted on are as follows: favorable comment 2285 times, performance 1032 times, appearance looks elegant 506 times, touching are sensitive
112 times, photographic effect 884 times, difference are commented 66 times, fringe is 32 times too ugly, jitter 12 times, are not had earphone 25 times.
If getting objective appraisal keyword of the target user about " 10 mobile phone of honor " are as follows: favorable comment, performance are good, then root
According to the word frequency of the objective appraisal keyword and each evaluation keyword, the target user is calculated about the dispensing
The evaluation of estimate of content；Evaluation of estimate is that positive number then shows that user is interested in " 10 mobile phone of honor ", if evaluation of estimate is negative, table
Bright user loses interest in " 10 mobile phone of honor ".
After being classified, the relevant advertisements of " 10 mobile phone of honor " can be launched to " 10 mobile phone of honor " interested user.
In this way, the accuracy that can be improved advertisement dispensing, avoids the dispensing of invalid advertisement, the throwing of advertisement is greatly saved
Put cost.
The present invention also provides a kind of one embodiment of user classification method launched based on advertisement, as shown in Fig. 2, packet
It includes:
S11 counts evaluation keyword relevant to content is launched from information bank；
S12 is divided into front evaluation keyword and unfavorable ratings keyword according to classifying rules to the evaluation keyword；
S13 counts the word frequency of each front evaluation keyword and the word frequency of each unfavorable ratings keyword；
S2 obtains target user about the objective appraisal keyword for launching content；
S31 is right according to the word frequency of each front evaluation keyword and the word frequency of each unfavorable ratings keyword
The front evaluation keyword and the unfavorable ratings keyword are normalized, and calculate each front evaluation keyword
The weighted value of weighted value and each unfavorable ratings keyword；
S32 is according to the objective appraisal keyword and weighted value, each unfavorable ratings of each front evaluation keyword
The weighted value of keyword calculates target user about the evaluation of estimate for launching content, method are as follows:
Wherein, V is target user about the evaluation of estimate for launching content；B is offset parameter；wiIndicate i-th of front
The weight of keyword is evaluated, andIf i-th of front evaluation keyword occurs in the objective appraisal keyword,
By xiIt is denoted as 1, is otherwise denoted as 0；wjIndicate the weight of j-th of unfavorable ratings keyword, andIf j-th is negatively commented
Valence keyword occurs in the objective appraisal keyword, then by xjIt is denoted as 1, is otherwise denoted as 0；
If the S41 target user is greater than preset value about the evaluation of estimate for launching content, by the target user
It is classified as first kind user, the target user is otherwise classified as the second class user；
Or；
All target users are ranked up by S42 according to corresponding evaluation of estimate, take the target user of preset order section
As first kind user, using other target users as the second class user.
In the present embodiment, ad content is launched by taking " striking news router " as an example, when classifying to user, can be led to first
Cross crawler software crawl in electric business website with the evaluation keyword of " striking news router " related content, such as: performance is high, it is high-quality,
Signal is strong, signal stabilization, appearance looks elegant, logistics are slow, swinging of signal, too expensive, not cost-effective, by two-value classification analysis, by these
Evaluation keyword is divided into front evaluation keyword: xi={ performance is high, and high-quality, signal is strong, signal stabilization, appearance looks elegant }, and
Unfavorable ratings keyword: xj={ logistics is slow, swinging of signal, too expensive, not cost-effective }.When statistical appraisal keyword, also need
Count the appearance word frequency of each evaluation keyword, such as relevant evaluation data are as follows: performance is 1028 times high, it is 1522 times high-quality,
Signal is 899 times strong, signal stabilization 728 times, appearance looks elegant 255 times, logistics are 66 times slow, jitter 38 times, 72 times too expensive, no
Cost-effective 19 times.
When classifying to target user, need to obtain in the network data of target user about " striking news router "
Objective appraisal keyword, for classifying to target user, such as user is to " striking news router " objective appraisal keyword
Are as follows: performance is high, signal is strong, too expensive.
According to the word frequency of the word frequency of each front evaluation keyword and each unfavorable ratings keyword, to described
Front evaluation keyword and the unfavorable ratings keyword are normalized, and calculate the weight of each front evaluation keyword
The weighted value of value and each unfavorable ratings keyword；Such as weight=1028/ of " performance is high " this front evaluation keyword
(1028+1522+899+728+225), the weight for thus obtaining this five front evaluation keywords are followed successively by wi=0.234,
0.346,0.204,0.165,0.051 }, the weight that can also equally calculate four unfavorable ratings keywords is followed successively by wj=-
0.338, -0.195, -0.369, -0.097 }.
If i-th of front evaluation keyword in front evaluation keyword occurs in the objective appraisal keyword,
By xiIt is denoted as 1, is otherwise denoted as 0；If j-th of unfavorable ratings keyword in unfavorable ratings keyword is crucial in the objective appraisal
Occur in word, then by xjIt is denoted as 1, is otherwise denoted as 0；According to objective appraisal keyword are as follows: performance is high, signal is strong, too expensive, can obtain
To xi={ 1,0,1,0,0 }；xj={ 0,0,1,0 },
Target user can be then calculated by formula 1 about the evaluation of estimate for launching content, preset if evaluation of estimate is greater than
Value, then illustrate that the front evaluation comparison of target user is more, target user be classified as first kind user (can launch user).If
Target user is less than preset value about the evaluation of estimate for launching content, then illustrates that the unfavorable ratings of target user are relatively more, will
This target user is classified as the second class user (can not launch user).By above-mentioned example, it can calculate this target user's
Evaluation of estimate=0.234+0.204-0.369+b, b are constant, therefore can be divided according to the preset value of setting to target user
Class.
In addition, after calculating all target users about the evaluation of estimate for launching content, to each target user's
Evaluation of estimate is ranked up, and goes preceding 30 percent to be used as first kind user, target user in addition is as the second class user.For
First kind user then carries out corresponding advertisement dispensing, for the second class user, then without launching.
The present invention provides a kind of one embodiment of user classification method launched based on advertisement, server includes:
Statistical module 1, for being counted from information bank and launching relevant multiple evaluation keywords of content and each
Evaluate the word frequency of keyword；
Module 2 is obtained, for obtaining target user about the objective appraisal keyword for launching content；
Computing module 3 is electrically connected with the statistical module 1, the acquisition module 2, for being closed according to the objective appraisal
The word frequency of keyword and each evaluation keyword calculates the target user about the evaluation of estimate for launching content；
User's categorization module 4 is electrically connected with the computing module 3, is used for according to the target user about the dispensing
The evaluation of estimate of content, classifies to target user.
Advertisement needs the target user of selected dispensing before dispensing.In the prior art, can by the network data of user,
Judge whether user consults the excessively relevant information of ad content, it, will if user browsed data relevant to ad content
The user is determined as target user.Such putting mode does not distinguish user, there is many use in targeted user population
Family launches content to advertisement and loses interest in, or even launches content to advertisement and detest very much.If blindly launching, one is increased
Cost is launched in many invalid popularizations and advertisement, and two to cause to the corresponding product of ad content but also user dislikes very much
More dislike.
The present invention can classify to user during selected target user, such as be divided into interested in dispensing content
User and to launch the uninterested user of content, when carrying out advertisement dispensing, only to dispensing the interested user of content
Advertisement dispensing is carried out, had both achieved the purpose that advertisement was launched in this way, and had decreased invalid dispensing, reduces advertisement dispensing
Cost.
Specifically, can be thrown first by crawler software, related around the business of dispensing when carrying out category filter to user
Content is put, crawls corpus data relevant to business information, and be stored in database.Such as ad placement service is " honor 10
The advertisement of mobile phone " is launched, and the relevant evaluation on Baidu's discussion bar about " 10 mobile phone of honor " can be crawled by crawler software, and will
These evaluations are expected in deposit information bank.
Then multiple evaluation keywords relevant to " 10 mobile phone of honor " are counted in information bank and each evaluation is closed
The word frequency of keyword, such as the data counted on are as follows: favorable comment 2285 times, performance 1032 times, appearance looks elegant 506 times, touching are sensitive
112 times, photographic effect 884 times, difference are commented 66 times, fringe is 32 times too ugly, jitter 12 times, are not had earphone 25 times.
Getting wherein objective appraisal keyword of the user about " 10 mobile phone of honor " are as follows: favorable comment, performance are good, then
According to the word frequency of the objective appraisal keyword and each evaluation keyword, the target user is calculated about the throwing
Put the evaluation of estimate of content；Evaluation of estimate is that positive number then shows that user is interested in " 10 mobile phone of honor ", if evaluation of estimate is negative,
Show that user loses interest in " 10 mobile phone of honor ".
After being classified, the relevant advertisements of " 10 mobile phone of honor " can be launched to " 10 mobile phone of honor " interested user.
In this way, the accuracy that can be improved advertisement dispensing, avoids the dispensing of invalid advertisement, the throwing of advertisement is greatly saved
Put cost.
As shown in figure 3, the present invention provides a kind of one embodiment of user classification method launched based on advertisement, service
Device includes:
Statistical module 1, for being counted from information bank and launching relevant multiple evaluation keywords of content and each
Evaluate the word frequency of keyword；
The statistical module 1 specifically includes:
Keyword statistic unit 11 is also used to count evaluation keyword relevant to content is launched from information bank；
Keyword classification unit 12 is electrically connected to the evaluation keyword according to classification with the keyword statistic unit 11
Rule is divided into front evaluation keyword and unfavorable ratings keyword；
Word frequency statistics unit 13 is electrically connected with the keyword classification unit 12 and is also used to count each front evaluation key
The word frequency of the word frequency of word and each unfavorable ratings keyword.
Module 2 is obtained, for obtaining target user about the objective appraisal keyword for launching content；
Computing module 3 is electrically connected, for being commented according to each front with the statistical module 1, the acquisition module 2
The word frequency of the word frequency of valence keyword and each unfavorable ratings keyword described is commented to the front evaluation keyword and negatively
Valence keyword is normalized, and calculates the weighted value and each unfavorable ratings keyword of each front evaluation keyword
Weighted value；
Computing module 3 is also used to the weight according to the objective appraisal keyword and each front evaluation keyword
The weighted value of value, each unfavorable ratings keyword calculates target user about the evaluation of estimate for launching content, method
Are as follows:
Wherein, V is target user about the evaluation of estimate for launching content；B is offset parameter；wiIndicate i-th of front
The weight of keyword is evaluated, andIf i-th of front evaluation keyword occurs in the objective appraisal keyword,
By xiIt is denoted as 1, is otherwise denoted as 0；w0jIndicate the weight of j-th of unfavorable ratings keyword, andIf j-th is negatively commented
Valence keyword occurs in the objective appraisal keyword, then by xjIt is denoted as 1, is otherwise denoted as 0.
User's categorization module 4 is greater than in advance if being also used to the target user about the evaluation of estimate for launching content
If value, then be classified as first kind user for the target user, the target user be otherwise classified as the second class user；
Or；
User's categorization module 4 is also used to for all target users being ranked up according to corresponding evaluation of estimate, take
The target user of preset order section is as first kind user, using other target users as the second class user.
In the present embodiment, ad content is launched by taking " striking news router " as an example, when classifying to user, can be led to first
It crosses crawler software to crawl with the evaluation keyword of " striking news router " related content in electric business website, performance is high, high-quality, signal
By force, signal stabilization, appearance looks elegant, logistics be slow, swinging of signal, too expensive, not cost-effective, and by two-value classification analysis, these are evaluated
Keyword is divided into front evaluation keyword: xi={ performance is high, and high-quality, signal is strong, signal stabilization, appearance looks elegant }, and it is negative
Evaluate keyword: xj={ logistics is slow, swinging of signal, too expensive, not cost-effective }.When statistical appraisal keyword, it is also necessary to unite
Count the appearance word frequency of each evaluation keyword, such as relevant evaluation data are as follows: performance is 1028 times high, 1522 times high-quality, signal
Strong 899 times, signal stabilization 728 times, appearance looks elegant 255 times, logistics be 66 times slow, jitter 38 times, 72 times too expensive, not cost-effective
19 times.
When classifying to target user, need to obtain in the network data of target user about " striking news router "
Objective appraisal keyword, for classifying to target user, such as objective appraisal keyword are as follows: jitter, too expensive.
According to the word frequency of the word frequency of each front evaluation keyword and each unfavorable ratings keyword, to described
Front evaluation keyword and the unfavorable ratings keyword are normalized, and calculate the weight of each front evaluation keyword
The weighted value of value and each unfavorable ratings keyword；Such as weight=1028/ of " performance is high " this front evaluation keyword
(1028+1522+899+728+225), the weight for thus obtaining this five front evaluation keywords are followed successively by wi=0.234,
0.346,0.204,0.165,0.051 }, the weight that can also equally calculate four unfavorable ratings keywords is followed successively by wj=-
0.338, -0.195, -0.369, -0.097 }.
If front evaluation keyword xiIn i-th front evaluation keyword occur in the objective appraisal keyword,
Then by xiIt is denoted as 1, is otherwise denoted as 0；If unfavorable ratings keyword xjIn j-th of unfavorable ratings keyword in the objective appraisal
Occur in keyword, then by xjIt is denoted as 1, is otherwise denoted as 0；According to objective appraisal keyword are as follows: performance is high, signal is strong, too expensive, by
This available xi={ 0,0,0,0,0 }；xj={ 0,1,1,0 },
Target user can be then calculated by formula 1 about the evaluation of estimate=- 0.195-0.369+b, b for launching content
For constant, therefore can be classified according to the preset value of setting to target user.If evaluation of estimate is greater than preset value, illustrate mesh
The front evaluation comparison for marking user is more, and target user is classified as first kind user (can launch user).If target user is closed
It is less than preset value in the evaluation of estimate for launching content, then illustrates that the unfavorable ratings of target user are relatively more, by this target user
It is classified as the second class user (user can not be launched).
In addition, after calculating all target users about the evaluation of estimate for launching content, to each target user's
Evaluation of estimate is ranked up, and goes preceding 30 percent to be used as first kind user, target user in addition is as the second class user.For
First kind user then carries out corresponding advertisement dispensing, for the second class user, then without launching.
It should be noted that above-described embodiment can be freely combined as needed.The above is only of the invention preferred
Embodiment, it is noted that for those skilled in the art, in the premise for not departing from the principle of the invention
Under, several improvements and modifications can also be made, these modifications and embellishments should also be considered as the scope of protection of the present invention.
Claims (10)
1. a kind of user classification method launched based on advertisement, which is characterized in that comprising steps of
The word frequency of multiple evaluation keywords relevant to content is launched and each evaluation keyword is counted from information bank；
Target user is obtained about the objective appraisal keyword for launching content；
According to the word frequency of the objective appraisal keyword and each evaluation keyword, the target user is calculated about institute
State the evaluation of estimate for launching content；
According to the target user about the evaluation of estimate for launching content, classify to target user.
2. a kind of user classification method launched based on advertisement according to claim 1, which is characterized in that from information bank
The word frequency of multiple evaluation keywords relevant to content is launched and each evaluation keyword is counted, this step specifically includes:
Evaluation keyword relevant to content is launched is counted from information bank；
Front evaluation keyword and unfavorable ratings keyword are divided into according to classifying rules to the evaluation keyword；
Count the word frequency of each front evaluation keyword and the word frequency of each unfavorable ratings keyword.
3. a kind of user classification method launched based on advertisement according to claim 2, which is characterized in that according to the mesh
The word frequency of mark evaluation keyword and each evaluation keyword calculates the target user about described and launches commenting for content
Value, this step specifically include:
According to the word frequency of the word frequency of each front evaluation keyword and each unfavorable ratings keyword, to the front
Evaluation keyword and the unfavorable ratings keyword be normalized, calculate it is each front evaluation keyword weighted value with
And the weighted value of each unfavorable ratings keyword；
According to the objective appraisal keyword and weighted value, each unfavorable ratings keyword of each front evaluation keyword
Weighted value, calculate target user about it is described launch content evaluation of estimate.
4. a kind of user classification method launched based on advertisement according to claim 3, which is characterized in that according to the mesh
The weighted value of the weighted value of mark evaluation keyword and each front evaluation keyword, each unfavorable ratings keyword, calculates
Method of the target user about the evaluation of estimate meter for launching content are as follows:
Wherein, V is target user about the evaluation of estimate for launching content；B is offset parameter；wiIndicate that i-th of front evaluation is closed
The weight of keyword, andIf i-th of front evaluation keyword occurs in the objective appraisal keyword, by xiNote
It is 1, is otherwise denoted as 0；w0jIndicate the weight of j-th of unfavorable ratings keyword, andIf j-th of unfavorable ratings is crucial
Word occurs in the objective appraisal keyword, then by xjIt is denoted as 1, is otherwise denoted as 0.
5. a kind of user classification method launched based on advertisement described in any one of -4 according to claim 1, which is characterized in that
According to the target user about the evaluation of estimate for launching content, classify to target user, this step specifically includes:
If the target user is greater than preset value about the evaluation of estimate for launching content, the target user is classified as the
Otherwise the target user is classified as the second class user by a kind of user；
Or；
All target users are ranked up according to corresponding evaluation of estimate, take the target user of preset order section as first
Class user, using other target users as the second class user.
6. a kind of server for user's classification, which is characterized in that server includes:
Statistical module, for being counted from information bank and launching the relevant multiple evaluation keywords of content and each evaluation pass
The word frequency of keyword；
Module is obtained, for obtaining target user about the objective appraisal keyword for launching content；
Computing module is electrically connected with the statistical module, the acquisition module, for according to the objective appraisal keyword, with
And the word frequency of each evaluation keyword, the target user is calculated about the evaluation of estimate for launching content；
User's categorization module is electrically connected with the computing module, for launching content about described according to the target user
Evaluation of estimate classifies to target user.
7. a kind of server for user's classification according to claim 6, which is characterized in that the statistical module is specific
Include:
Keyword statistic unit is also used to count evaluation keyword relevant to content is launched from information bank；
Keyword classification unit is electrically connected with the keyword statistic unit, to the evaluation keyword according to classifying rules point
For front evaluation keyword and unfavorable ratings keyword；
Word frequency statistics unit is electrically connected with the keyword classification unit, is also used to count the word of each front evaluation keyword
The word frequency of frequency and each unfavorable ratings keyword.
8. a kind of server for user's classification according to claim 7, it is characterised in that:
The computing module is also used to crucial according to the word frequency of each front evaluation keyword and each unfavorable ratings
The word frequency of word is normalized the front evaluation keyword and the unfavorable ratings keyword, calculates each front
Evaluate the weighted value of keyword and the weighted value of each unfavorable ratings keyword；
The computing module, be also used to according to the objective appraisal keyword and it is each front evaluation keyword weighted value,
The weighted value of each unfavorable ratings keyword calculates target user about the evaluation of estimate for launching content.
9. it is according to claim 8 it is a kind of for user classification server, which is characterized in that the computing module according to
The weighted value of the weighted value of the objective appraisal keyword and each front evaluation keyword, each unfavorable ratings keyword,
Calculate method of the target user about the evaluation of estimate for launching content are as follows:
Wherein, V is target user about the evaluation of estimate for launching content；B is offset parameter；wiIndicate that i-th of front evaluation is closed
The weight of keyword, andIf i-th of front evaluation keyword occurs in the objective appraisal keyword, by xiNote
It is 1, is otherwise denoted as 0；w0jIndicate the weight of j-th of unfavorable ratings keyword, andIf j-th of unfavorable ratings is crucial
Word occurs in the objective appraisal keyword, then by xjIt is denoted as 1, is otherwise denoted as 0.
10. a kind of server for user's classification according to any one of claim 6-10, it is characterised in that:
User's categorization module is greater than preset value about the evaluation of estimate for launching content if being also used to the target user,
The target user is then classified as first kind user, the target user is otherwise classified as the second class user；
Or；
User's categorization module is also used to for all target users being ranked up according to corresponding evaluation of estimate, take default
The target user of sequential segments is as first kind user, using other target users as the second class user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810685006.5A CN109003122A (en) | 2018-06-28 | 2018-06-28 | A kind of user classification method and server launched based on advertisement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810685006.5A CN109003122A (en) | 2018-06-28 | 2018-06-28 | A kind of user classification method and server launched based on advertisement |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109003122A true CN109003122A (en) | 2018-12-14 |
Family
ID=64601330
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810685006.5A Pending CN109003122A (en) | 2018-06-28 | 2018-06-28 | A kind of user classification method and server launched based on advertisement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109003122A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110322319A (en) * | 2019-06-26 | 2019-10-11 | 安徽景徽菜篮子电子商务有限公司 | A kind of electric business platform auto recommending method of user's evaluation |
| CN111400587A (en) * | 2020-03-03 | 2020-07-10 | 网易（杭州）网络有限公司 | User classification method and device, electronic equipment and storage medium |
| CN113935789A (en) * | 2021-12-20 | 2022-01-14 | 广州半城云信息科技有限公司 | User-defined crowd division method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550903A (en) * | 2015-12-25 | 2016-05-04 | 腾讯科技（深圳）有限公司 | Target user determination method and apparatus |
| CN107608980A (en) * | 2016-07-11 | 2018-01-19 | 中国电信股份有限公司 | Information-pushing method and system based on the analysis of DPI big datas |
-
2018
- 2018-06-28 CN CN201810685006.5A patent/CN109003122A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550903A (en) * | 2015-12-25 | 2016-05-04 | 腾讯科技（深圳）有限公司 | Target user determination method and apparatus |
| CN107608980A (en) * | 2016-07-11 | 2018-01-19 | 中国电信股份有限公司 | Information-pushing method and system based on the analysis of DPI big datas |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110322319A (en) * | 2019-06-26 | 2019-10-11 | 安徽景徽菜篮子电子商务有限公司 | A kind of electric business platform auto recommending method of user's evaluation |
| CN111400587A (en) * | 2020-03-03 | 2020-07-10 | 网易（杭州）网络有限公司 | User classification method and device, electronic equipment and storage medium |
| CN111400587B (en) * | 2020-03-03 | 2024-01-30 | 网易（杭州）网络有限公司 | User classification method, device, electronic equipment and storage medium |
| CN113935789A (en) * | 2021-12-20 | 2022-01-14 | 广州半城云信息科技有限公司 | User-defined crowd division method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104298719B (en) | Category division, advertisement placement method and the system of user is carried out based on Social behaviors | |
| CN105574159B (en) | A kind of user's portrait method for building up and user's portrait management system based on big data | |
| CN105183897B (en) | A kind of method and system of video search sequence | |
| CN103116588B (en) | A kind of personalized recommendation method and system | |
| CN110489644A (en) | Information-pushing method, device, computer readable storage medium and computer equipment | |
| CN103678613B (en) | Method and device for calculating influence data | |
| CN102929873B (en) | Method and device for extracting searching value terms based on context search | |
| US8332775B2 (en) | Adaptive user feedback window | |
| US20060248072A1 (en) | System and method for spam identification | |
| CN106033445B (en) | The method and apparatus for obtaining article degree of association data | |
| CN109003122A (en) | A kind of user classification method and server launched based on advertisement | |
| Collins et al. | Fake news types and detection models on social media a state-of-the-art survey | |
| CN106407484A (en) | Video tag extraction method based on semantic association of barrages | |
| CN111324801B (en) | Hot event discovery method in judicial field based on hot words | |
| CN102945246B (en) | The disposal route of network information data and device | |
| CN110737821B (en) | Similar event query method, device, storage medium and terminal equipment | |
| Sandholm et al. | Real-time, location-aware collaborative filtering of web content | |
| CN103605714A (en) | Method and device for identifying abnormal data of websites | |
| CN104809252A (en) | Internet data extraction system | |
| CN103745380A (en) | Advertisement delivery method and apparatus | |
| CN103955480B (en) | A kind of method and apparatus for determining the target object information corresponding to user | |
| CN108460073A (en) | Group recommending method, storage medium and server | |
| CN106919588A (en) | A kind of application program search system and method | |
| CN104834739A (en) | Internet information storage system | |
| CN107741958A (en) | A kind of data processing method and system |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200807 Address after: 201616 Shanghai city Songjiang District Sixian Road No. 3666 Applicant after: Phicomm (Shanghai) Co.,Ltd. Address before: 610100 125 Longquan Street Park Road, Longquanyi District, Chengdu, Sichuan. Applicant before: SICHUAN PHICOMM INFORMATION TECHNOLOGY Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20181214 |
|
| WD01 | Invention patent application deemed withdrawn after publication |