CN106844649A - A kind of waste and old industry commending system based on mixing various modes and its method - Google Patents
A kind of waste and old industry commending system based on mixing various modes and its method Download PDFInfo
- Publication number
- CN106844649A CN106844649A CN201710047693.3A CN201710047693A CN106844649A CN 106844649 A CN106844649 A CN 106844649A CN 201710047693 A CN201710047693 A CN 201710047693A CN 106844649 A CN106844649 A CN 106844649A
- Authority
- CN
- China
- Prior art keywords
- recommendation
- user
- data
- gqinfo
- various modes
- 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of waste and old industry commending system based on mixing various modes and its method, the system mainly includes：A module Main Functions are that correlation recommendation, focus recommendation, TopN recommend three kinds of mixing recommendation patterns；B module Main Functions are, initializing recommendation result；C module Main Functions are, filtering, sort result, recommend to explain, final recommendation results.Wherein to the normal use user in website, correlation recommendation and TopN recommend to set higher weights, and focus recommendation is taken second place；To without any record user, focus recommendation being set into weight higher, and by analyzing the adaptation population belonging to user, correlation recommendation sets corresponding weight.According to embodiments of the present invention, enable to recommendation results more accurate so that the commodity purchasing rate of recommendation increases, so as to improve the order conversion ratio of commodity.
Description
Technical field
Pushed away the present invention relates to Computer Applied Technology field, more particularly to a kind of waste and old industry based on mixing various modes
Recommend system and its method.
Background technology
At present, with network information explosion type be incremented by, consumer face numerous selections, unknown field, overload
It is often at a loss as to what to do during information；But at the same time, suitable user is also earnestly seeking in the businessman of product, and it is most convenient to find
Channel, and the best tool for solving this two classes contradiction is exactly commending system.
Data are the bases of all commending systems.Accurate data for good recommendation effect, such as a piece of article
Effect of the title for article content.Collaborative filtering recommending based on model, the user preference information based on sample trains one
Recommended models, then the information according to real-time user preferences be predicted, calculate recommend, this method is for some specialty goods
The user of taste can not give recommendation well；Content-based recommendation, its core concept is according to recommendation article or content
Metadata, finds the correlation of article or content, is then based on the conventional hobby record of user, recommends the similar thing of user
Product, the criterion of this method article similarity has only taken into account article in itself, there is certain one-sidedness.
Therefore, merely using some proposed algorithm in terms of the precision and diversity of recommendation results Shortcomings, for
The problem of presence, the present invention proposes a kind of waste and old industry commending system based on mixing various modes and its method, can make
Obtain recommendation results more accurate, can more meet the demand of user.
The content of the invention
In view of this, system is recommended it is a primary object of the present invention to provide a kind of waste and old industry based on mixing various modes
System and its method, when user buys commodity, merchandise news interested are recommended to user, can more meet the individual character of user
Change demand.
To reach above-mentioned purpose, the technical proposal of the invention is realized in this way：
A modules：Correlation recommendation, focus recommendation, TopN recommend three kinds of mixing recommendation patterns；
B modules：Initializing recommendation result；
C modules：Filtering, sort result, recommendation explanation, final recommendation results.
Correlation recommendation in wherein described modules A, the similitude according to content of good is recommended, wherein according in commodity
Appearance carries out recommending to need to import the original initial data for preserving in the index in database till now；Described TopN is pushed away
Recommend, that is, browse history recommendation, hits TopN commercial product recommendings in each user a period of time；Described focus is pushed away
Recommend, real-time hot item is recommended into user, possible data interested are recommended to each user.
To the normal use user in website, correlation recommendation and TopN recommend to set higher weights, and focus recommendation is taken second place；To without any note
Family is employed, focus recommendation is set into weight higher, by analyzing the adaptation population belonging to user, can also be associated recommendation.
Further, described TopN recommends mainly to include following two strategies：
TopN strategies：
1）The click logs of nearest 30 days；
2）The data of total hits TopN；
3）Total hits are no less than certain threshold value；
4）Averagely everyone number of clicks is no less than certain threshold value；
It is continuous to be incremented by strategy：
1）The click logs of nearest 30 days；
2）Continuous some days click is presented increasing trend；
3）Continuous number of days is no less than certain threshold value；
4）The number of clicks of average every day is no less than certain threshold value.
The module B initializing recommendation results, the recommendation results of return are recommendation explanation：ResysExplain, supply and demand are compiled
Number：M_gqinfo.gqid, product classification:M_gqinfo.classid, supply and demand classification：M_gqinfo.type, information static page
Face address:M_gqinfo.htmlurl, picture：M_gqinfo.photo, title：M_gqinfo.title, newness degree: M_
Gqinfo.xjcd, quantity of supplying：M_gqinfo.pronum, model specification：M_gqinfo.proxh, transaction value：M_
Gqinfo.proprice products location：M_gqinfo.province, location mode：M_gqinfo.cffs, authority：sys_
User.rankid, membership number：Sys_user.uid, user name：Sys_user.uname, enterprise name：sys_
User.comname, contact person：Sys_user.linkman, sex：Sys_user.sex, company location: sys_
User.comaddress, Debao index:Czizhi_rz.frrz, whether by Debao certification：czizhi_rz；
The module C key steps are as follows：
Step C1, the filtering of recommending data mainly include：The page that user had accessed is filtered out, user is filtered out and is not visited
The page of authority is asked, repeated data is filtered out；
Step C2, sort result be mainly to recommend result be ranked up, according to webpage pouplarity sort, that is,
The number of times that webpage was clicked；
Mainly total number of clicks is explained in step C3, recommendation：totalClickNum；The click data of each user：List
<Entry<User, Integer>> userClickNumForAll；The click time of first day：Date sDate；
Step C4, final recommendation results are mainly by step C1, C2, C3 treated recommending data, with apparent accurate
Mode be presented to user.
Further, the present invention is estimated to recommendation results, mainly passes through three below evaluation criteria：
1）Training data and scoring：In commending system of the present invention, extract a bit of True Data and emulated as test data；
2）Precision ratio：It is the ratio for having " good " result in the middle of top recommends；
3）Recall ratio：It is the ratio during " good " result appears in top recommendations.
Mixing various modes personalized recommendation method provided by the present invention, with advantages below：
1）Correlation recommendation, focus recommendation, TopN recommendation Three models are mixed with, the precision of recommendation results is improve；
2）User's request can more be met so that the commodity purchasing rate of recommendation increases, so as to improve the order conversion ratio of commodity；
3）Increase the diversity recommended, recommended the user of particular preferences.
Brief description of the drawings
Fig. 1 is the waste and old industry commending system schematic flow sheet of present invention mixing various modes；
Fig. 2 is that the waste and old industry commending system of present invention mixing various modes makes training set schematic diagram；
Fig. 3 is the recommendation behavior interaction schematic diagram of present invention mixing various modes.
Specific embodiment
Below in conjunction with the accompanying drawings and embodiments of the invention to it is of the invention mixing various modes personalized recommendation method make
Further details of explanation.
The method that correlation recommendation mainly uses the similar commodity of cluster calculation in system A modules, mainly comprising following
Flow：
1）An entity class SimilarityData is created, three field row are set（OK）、 column （Row）、
similarityValue（Similarity）, wherein the entity class SimilarityData Main Functions are triples, in matrix
Some element, for preserving sparse matrix；Wherein described similarity was tested before cluster calculation is carried out by training set
Draw and minimum similarity degree is set into minSimilarity as 0.8 is more suitable；
2）Example SimilarityData [] [] allSimilarityData is created, is all of similarity number for what is preserved
According to；
3）Array int [] countArray is created, for preserving the sum of the similar data of each data；
4）Initialization matrix；
5）Similarity between any two is calculated, wherein described similarity, using the similar of Jaccard coefficients two vectors of calculating
Degree；
6）The sum of the data similar with row datas；
7）Give set of metadata of similar data assignment；It is wherein described to set of metadata of similar data assignment when, when having a new data：
If the data being expert at are not filled up also, it is directly inserted into behind last element；If the data being expert at are
Through filling up, then it is compared with current data and the least member being expert at, is then replaced if greater than minimum element, otherwise not
Do any operation.
Further, data source needed for similar commodity is calculated in training set, training set Making programme as shown in Fig. 2 main
To include procedure below：
1）The data that database reads carry out participle, and the data after participle are had into default folder in space form
Resys, wherein also needing to judge whether to need to update training set before data are read；
2）Read the training data after participle；
3）Feature extraction is carried out using TF-IDF and LDA mixed models, for clustering；
Wherein described this language models of TF-IDF are mainly and are used as feature set with the statistical nature of vocabulary, and each feature can
Enough saying draws physical significance, extracts effect pretty good, however, a key issue of these features, is not special to sample
Levy and significantly compressed, do not extract the information of key.The grader for namely training only is training its number
Effective according to concentrating, changing a data set effect will be very poor；
The LDA is exactly the rarefaction representation of text, and this class language model of representative is called Topic Model.Think word amount
Big text again, its article theme is just so several.One LDA model of K theme, can tie up a Text compression into K
Vector：Each dimension is exactly the probability that the text belongs to the theme, and this vector is also referred to as Topic Proportion.So
After K dimension data collection after being compressed afterwards, any grader, or even simplest cosine similarity index are reused, all may be used
To obtain extraordinary classifying quality；
Therefore, the present invention combines both that to carry out feature extraction better.
4）Vectorization is carried out to commodity using TF, while initializing LDA topic models, and each theme correspondence is exported
All words；The feature of its Chinese version, it may be possible to topic, it is also possible to word；
5）Cluster, calculates similar commodity.
The present invention mainly employs following technology to solve the problems, such as personalized recommendation, carries out letter to these technologies below
It is single to introduce.
1）Participle technique.It is main in the present invention to use IK participles, mainly used when training set is made, wherein IK participles
" forward iteration most fine granularity segmentation algorithm " is used, has been exactly in brief：Segmenter can word for word recognize lemma, the present invention
IK is written over, useSmart true will be set into IKAnalyzerSegmenter classes, wherein described useSmart
When its value is that false is non intelligent participle, fine granularity exports all possible cutting result；When its value is that true is intelligence point
Word, merges number and measure word, and ambiguity judgement is carried out to word segmentation result.
2）Clustering technique.Present invention is mainly used in clustering technique has been used when calculating similar commodity, phase knowledge and magnanimity meter is clustered
The algorithm for calculating commodity creates an entity class SimilarityData first, sets three field row（OK）、 column
（Row）、similarityValue（Similarity）, wherein the entity class SimilarityData Main Functions are triple, square
Some element in battle array, for preserving sparse matrix, secondly initialization matrix, finally calculates the similarity between commodity.
3）Text feature extraction technique.The present invention has mainly used two kinds of feature extraction modes, and one kind is to use TF-IDF
Feature extraction is carried out with LDA mixed models, for clustering.
4）Recommendation results ordering techniques.The present invention arranges recommendation results according to totalClickNum descendings,
TotalClickNum is exactly that is, the pouplarity according to web hit.
With reference to Fig. 3, personalized recommendation method embodiment of the invention is as follows.
The several typical application scenarios of the method are described below：
Application scenarios one：
The interaction of recommendation behavior of the invention, it is as shown in table 1 below.
Table 1
| Metadata | Implication | Citing | In api interface Equivalent |
| User name | Perform the user name this time recommended | User name is " Zhang San " | username |
| Recommend the moment | System performs the moment recommended | " Zhang San " is in moment " 2016-12-25 15:55:00 " login system, " 15:55:00 " it is that system is performed The moment of recommendation | recommendTi me |
| The recommended page | For this recommendation, system is given The recommendation results page ID | " Zhang San " is in moment " 2016-12-2515:55:00 " login system, be recommended page ID for " 1,3,47, 556,1007 " etc. | recommendPa geID |
| The recommended page Sequence sequence number | When recommendation results are presented to user, The sequence sequence number of each page | No. ID is for sequence when 5 commodity of " 1,3,47,556,1007 " are showed as recommendation results " 556,3,47,1,1007 " so their corresponding row's serial numbers " 1,2,3,4,5 " | recommendPa geRankID |
| The page is clicked Moment | Click on the recommendation results page when when Carve | User clicks on No. ID for the moment of " 556 " is " 2012-08-05 15:56:24” | clickedTime |
| When the page is resident Between | It is clicked on the page at each and is resident Time | User checks the commodity that No. ID is " 556 ", in moment " 2016-12-25 15:57:26 " leave, when being resident Between be " 62 " second |
User logs in call setRecommendInfo () every time, clicks on and recommends to call updateRecommendInfo during the page
(),
For example：Zhang San is in moment " 2016-12-25 16:00:00 " login system, system recommends 3 pages, its page to him
Serial number " 1,3,47 ", sequence serial number " 2,3,1 ", now call setRecommendInfo (String userName,
Date recommendTime, long [] recommendPageID, long [] recommendPageRankID),
RecommendPageID stores 3 page sequence numbers, and recommendPageRankID stores 3 sequence sequence numbers of the page.
Application scenarios two：
Recommendation method of the invention is applied in certain waste and old Industry system details page displaying correlation recommendation, wherein, the details page is
In the waste and old site search frame input keyword, search is clicked on, into the product list page of waste and old net, then click on some business
Product, into commodity details page.
Input：Scrap iron and steel is reclaimed, and associated recommendation result is presented below：
{"isLoolApply":"","applyStatus":1,"lookApply":1,"auctionStatus":"","
code":"9cdddb1a9a3146a5984510e7057613e6","pmCode":null,"name":" steel-making steel scrap bucket bid
Bulletin ", " imgUrl ":null,"price":null,"valuation":" nothing ", " status ":null,"time":"2016-
11-11","releaseTime":"2016-11-08","num":0,"address":" Hebei province-Xingtai City ", "
endTime":"2016-11-11 00:00:00","bidCompany":""},{"isLoolApply":"","
applyStatus":1,"lookApply":1,"auctionStatus":"","code":"
d9e18ea7fbb94a19adf5d8b470cc0d43","pmCode":null,"name":" scrap house trailer steel scrap to vie for selling public affairs
Accuse ", " imgUrl ":null,"price":null,"valuation":" nothing ", " status ":null,"time":"2016-
11-16","releaseTime":"2016-11-08","num":0,"address":" Beijing-districts under city administration ", "
endTime":"2016-11-16 00:00:00","bidCompany":""},{"isLoolApply":"","
applyStatus":1,"lookApply":1,"auctionStatus":"","code":"
c74d9959c6f64b71af8e532107ed0714","pmCode":null,"name":" scrap lorry steel scrap to vie for selling public affairs
Accuse ", " imgUrl ":null,"price":null,"valuation":" nothing ", " status ":null,"time":"2016-
11-15","releaseTime":"2016-11-08","num":0,"address":" Beijing-districts under city administration ", "
endTime":"2016-11-15 00:00:00","bidCompany":""},{"isLoolApply":"","
applyStatus":1,"lookApply":1,"auctionStatus":"","code":"
303c762d475d4bb28053ccb270ab00e6","pmCode":null,"name":" 320 tons of steel scrap cords of coal industry company
Core conveyer belt transfer is announced ", " imgUrl ":null,"price":null,"valuation":" nothing ", " status ":null,"
time":"2016-11-21","releaseTime":"2016-11-08","num":0,"address":" Hui Nationality in Ningxia Hui Nationality Autonomy is autonomous
Area-Yinchuan City ", " endTime ":"2016-11-21 00:00:00","bidCompany":""},{"
isLoolApply":"","applyStatus":1,"lookApply":1,"auctionStatus":"","code":"
5721a3dad2bb4fac94e6c5359bf5cd44","pmCode":null,"name":" 300 tons of steel scrap cords transfer the possession of public
Accuse ", " imgUrl ":null,"price":null,"valuation":" nothing ", " status ":null,"time":"2016-
11-21","releaseTime":"2016-11-08","num":0,"address":" Ningxia Hui Autonomous Region-Yinchuan City ", "
endTime":"2016-11-21 00:00:00","bidCompany":""},{"isLoolApply":"","
applyStatus":1,"lookApply":1,"auctionStatus":"","code":"
a3dad8ab200c42ff8d734a2f7ed61b10","pmCode":null,"name":" 2000 tons of steel scraps（It is medium-sized）Transfer the possession of
Bulletin ", " imgUrl ":null,"price":null,"valuation":" 50-100 ten thousand ", " status ":null,"
time":"2016-11-21","releaseTime":"2016-11-08","num":0,"address":" Hui Nationality in Ningxia Hui Nationality Autonomy is autonomous
Area-Yinchuan City ", " endTime ":"2016-11-21 00:00:00","bidCompany":""},{"
isLoolApply":"","applyStatus":1,"lookApply":1,"auctionStatus":"","code":"
31e00b9286454f35b70b2f4bce1adbd8","pmCode":null,"name":" 630 tons of reports of the miscellaneous steel scrap of logistics company
The useless a collection of disposal of equipment goods and materials is announced ", " imgUrl ":null,"price":null,"valuation":" nothing ", " status ":
null,"time":"2016-11-16","releaseTime":"2016-11-08","num":0,"address":" Hubei
Province-Wuhan City ", " endTime ":"2016-11-15 00:00:00","bidCompany":""}
The above, only presently preferred embodiments of the present invention is not intended to limit the scope of the present invention.
The technical staff in the field can be understood that, for convenience of description and succinctly, foregoing description is
The specific work process of system, device and unit, may be referred to the corresponding process in preceding method embodiment, will not be repeated here.
In several embodiments provided by the present invention, it should be understood that disclosed system, apparatus and method, can be with
Realize by another way.For example, it is described above to device embodiment be only schematical, for example, the unit
Division, only a kind of division of logic function can have other dividing mode when actually realizing, such as multiple units or group
Part can be combined or be desirably integrated into another system, or some features can be ignored, or not performed.It is another, it is shown or
The coupling each other for discussing or direct-coupling or communication connection can be the indirect couplings of device or unit by some interfaces
Close or communicate to connect, can be electrical, mechanical or other forms.
The unit as separating component explanation can be or can also be physically separate, be shown as unit
Part can be or may not be physical location, you can with positioned at a place, or multiple nets can also be distributed to
On network unit.Some or all of unit therein can be according to the actual needs selected to realize the mesh of this embodiment scheme
's.
In addition, during each functional unit in each embodiment of the invention can be integrated in a processing unit, it is also possible to
It is that unit is individually physically present, it is also possible to which two or more units are integrated in a unit.Above-mentioned integrated list
Unit can both be realized in the form of hardware, can be realized in the form of SFU software functional unit.
It should be noted that one of ordinary skill in the art will appreciate that whole or portion in realizing above-described embodiment method
Split flow, can be by computer program to instruct the hardware of correlation to complete, and described program can be stored in a computer
In read/write memory medium, the program is upon execution, it may include such as the flow of the embodiment of above-mentioned each method.Wherein, it is described
Storage medium can be magnetic disc, CD, read-only memory（Read-Only Memory, ROM）Or random access memory
（Random Access Memory, RAM）Deng.
The waste and old industry commending system based on mixing various modes provided by the present invention and its method are carried out above
It is discussed in detail, specific embodiment used herein is set forth to principle of the invention and implementation method, and the above is implemented
The explanation of example is only intended to help and understands the method for the present invention and its core concept；Simultaneously for the general technology people of this area
Member, according to thought of the invention, will change in specific embodiments and applications, in sum, this explanation
Book content should not be construed as limiting the invention.
Claims (10)
1. it is a kind of based on the personalized recommendation method for mixing various modes, it is characterised in that the similar commodity process master of cluster calculation
Including：
A, one entity class SimilarityData of establishment, set three field row（OK）、 column （Row）、
similarityValue（Similarity）, wherein the entity class SimilarityData Main Functions are triples, in matrix
Some element, for preserving sparse matrix；
B, establishment example SimilarityData [] [] allSimilarityData, are all of similarity numbers for what is preserved
According to；
C, establishment array int [] countArray, for preserving the sum of the similar data of each data；
D, initialization matrix；
E, calculating similarity between any two；
The sum of F and the similar data of row datas；
G, give set of metadata of similar data assignment.
2. it is according to claim 1 based on the personalized recommendation method for mixing various modes, it is characterised in that the step
The similarity of A, specially：Before cluster calculation is carried out, drawn by training set test and set minimum similarity degree
MinSimilarity is 0.8 more suitable.
3. it is according to claim 1 based on the personalized recommendation method for mixing various modes, it is characterised in that the step
E, specially：Similarity must be calculated and calculate two similarities of vector using Jaccard coefficients.
4. it is according to claim 1 based on the personalized recommendation method for mixing various modes, it is characterised in that the step
G, specially：When there are a new data, if the data being expert at are not filled up also, last is directly inserted into
Behind individual element；If the data being expert at have been filled up, with current data and the minimum being expert at.
5. it is according to claim 1 based on the personalized recommendation method for mixing various modes, it is characterised in that the step
Also need to make set of metadata of similar data training set before A, specially：
The data read from database carry out participle, and the data after participle are had into default folder in space form
Resys, wherein also needing to judge whether to need to update training set before data are read；
Read the training data after participle；
Feature extraction is carried out using TF-IDF and LDA mixed models, for clustering；
Vectorization is carried out to commodity using TF, while LDA topic models are initialized, and it is corresponding all to export each theme
Word；The feature of its Chinese version, it may be possible to topic, it is also possible to word.
6. it is a kind of based on the waste and old industry commending system for mixing various modes, it is characterised in that the system mainly includes following several
Individual module：A modules, correlation recommendation, focus recommendation, three kinds of mixing recommendation patterns of TopN recommendations, B modules, initializing recommendation result,
C modules, filtering, sort result, recommendation explanation, final recommendation results.
7. it is according to claim 5 based on the waste and old industry commending system for mixing various modes, it is characterised in that the A
Module, specially：To the normal use user in website, correlation recommendation and TopN recommend to set higher weights, and focus recommendation is taken second place；Nothing is appointed
What record user, weight higher is set by focus recommendation, by analyzing the adaptation population belonging to user, can be also associated and be pushed away
Recommend.
8. it is according to claim 5 based on the waste and old industry commending system for mixing various modes, it is characterised in that the B
Module, specially：The recommendation results of return are explained for recommendation：ResysExplain, supply and demand numbering：M_gqinfo.gqid, product
Classification:M_gqinfo.classid, supply and demand classification：M_gqinfo.type, information static page address: M_
Gqinfo.htmlurl, picture：M_gqinfo.photo, title：M_gqinfo.title, newness degree: M_
Gqinfo.xjcd, quantity of supplying：M_gqinfo.pronum, model specification：M_gqinfo.proxh, transaction value：M_
Gqinfo.proprice products location：M_gqinfo.province, location mode：M_gqinfo.cffs, authority：sys_
User.rankid, membership number：Sys_user.uid, user name：Sys_user.uname, enterprise name：sys_
User.comname, contact person：Sys_user.linkman, sex：Sys_user.sex, company location: sys_
User.comaddress, Debao index:Czizhi_rz.frrz, whether by Debao certification：czizhi_rz.
9. it is according to claim 5 based on the waste and old industry commending system for mixing various modes, it is characterised in that the C
Module, specially：The filtering of recommending data mainly includes：The page that user had accessed is filtered out, filtering out user does not have
The page of access rights, filters out repeated data.
10. it is according to claim 5 based on the waste and old industry commending system for mixing various modes, it is characterised in that the C
Module, specially：Sort result be mainly to recommend result be ranked up, according to webpage pouplarity sort, also
It is number of times that webpage was clicked.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710047693.3A CN106844649A (en) | 2017-01-22 | 2017-01-22 | A kind of waste and old industry commending system based on mixing various modes and its method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710047693.3A CN106844649A (en) | 2017-01-22 | 2017-01-22 | A kind of waste and old industry commending system based on mixing various modes and its method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106844649A true CN106844649A (en) | 2017-06-13 |
Family
ID=59119335
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710047693.3A Pending CN106844649A (en) | 2017-01-22 | 2017-01-22 | A kind of waste and old industry commending system based on mixing various modes and its method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106844649A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107391687A (en) * | 2017-07-24 | 2017-11-24 | 华中师范大学 | A kind of mixing commending system towards local chronicle website |
| CN107688628A (en) * | 2017-08-21 | 2018-02-13 | 北京金堤科技有限公司 | The conventional packet construction method of relation group data and device |
| CN108427774A (en) * | 2017-11-23 | 2018-08-21 | 国网技术学院 | A kind of method and apparatus for commending contents |
| CN109493172A (en) * | 2018-10-23 | 2019-03-19 | 广州致轩服饰有限公司 | A kind of commodity method for pushing and device based on user tag |
| CN112000945A (en) * | 2020-08-24 | 2020-11-27 | 平安国际智慧城市科技股份有限公司 | Artificial intelligence based authorization method, device, equipment and medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103678672A (en) * | 2013-12-25 | 2014-03-26 | 北京中兴通软件科技股份有限公司 | Method for recommending information |
| CN104933239A (en) * | 2015-06-09 | 2015-09-23 | 江苏大学 | Hybrid model based personalized position information recommendation system and realization method therefor |
| CN105488662A (en) * | 2016-01-07 | 2016-04-13 | 北京歌利沃夫企业管理有限公司 | Bi-directional recommendation-based online recruitment system |
-
2017
- 2017-01-22 CN CN201710047693.3A patent/CN106844649A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103678672A (en) * | 2013-12-25 | 2014-03-26 | 北京中兴通软件科技股份有限公司 | Method for recommending information |
| CN104933239A (en) * | 2015-06-09 | 2015-09-23 | 江苏大学 | Hybrid model based personalized position information recommendation system and realization method therefor |
| CN105488662A (en) * | 2016-01-07 | 2016-04-13 | 北京歌利沃夫企业管理有限公司 | Bi-directional recommendation-based online recruitment system |
Non-Patent Citations (2)
| Title |
|---|
| 邓雄杰: "基于Hadoop的推荐系统的设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 马晓姝: "基于LDA模型的新闻话题发现研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107391687A (en) * | 2017-07-24 | 2017-11-24 | 华中师范大学 | A kind of mixing commending system towards local chronicle website |
| CN107688628A (en) * | 2017-08-21 | 2018-02-13 | 北京金堤科技有限公司 | The conventional packet construction method of relation group data and device |
| CN108427774A (en) * | 2017-11-23 | 2018-08-21 | 国网技术学院 | A kind of method and apparatus for commending contents |
| CN109493172A (en) * | 2018-10-23 | 2019-03-19 | 广州致轩服饰有限公司 | A kind of commodity method for pushing and device based on user tag |
| CN109493172B (en) * | 2018-10-23 | 2021-02-19 | 广州致轩服饰有限公司 | Commodity pushing method and device based on user tags |
| CN112000945A (en) * | 2020-08-24 | 2020-11-27 | 平安国际智慧城市科技股份有限公司 | Artificial intelligence based authorization method, device, equipment and medium |
| CN112000945B (en) * | 2020-08-24 | 2023-12-29 | 平安国际智慧城市科技股份有限公司 | Authorization method, device, equipment and medium based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106844649A (en) | A kind of waste and old industry commending system based on mixing various modes and its method | |
| Hsu | A personalized English learning recommender system for ESL students | |
| CN111784455B (en) | Article recommendation method and recommendation equipment | |
| CN110222272A (en) | A kind of potential customers excavate and recommended method | |
| CN105574003B (en) | A kind of information recommendation method based on comment text and scoring analysis | |
| Wang et al. | A personalized recommender system for the cosmetic business | |
| US8781813B2 (en) | Intent management tool for identifying concepts associated with a plurality of users' queries | |
| CN102982042B (en) | A kind of personalization content recommendation method, platform and system | |
| CN111898031B (en) | Method and device for obtaining user portrait | |
| CN104391883B (en) | A kind of online advertisement audient's sort method based on transfer learning | |
| CN106611344A (en) | Method and device for mining potential customers | |
| CN104268292A (en) | Label word library update method of portrait system | |
| CN106326318B (en) | Searching method and device | |
| CN105468649B (en) | Method and device for judging matching of objects to be displayed | |
| CN105740382A (en) | Aspect classification method for short comment texts | |
| CN108614832A (en) | A kind of user individual commercial articles searching implementation method and device | |
| CN107895303A (en) | A kind of method of the personalized recommendation based on OCEAN models | |
| CN107423335A (en) | A kind of negative sample system of selection for single class collaborative filtering problem | |
| CN105069129A (en) | Self-adaptive multi-label prediction method | |
| KR101013942B1 (en) | Apparatus and Method for Offering Item Recommendation | |
| CN105550892A (en) | On-line member gender recognition method and system, and information recommendation system | |
| CN110781399A (en) | Cross-platform information pushing method and device | |
| Clarizia et al. | Sentiment analysis in social networks: A methodology based on the latent dirichlet allocation approach | |
| Shelar et al. | TRAVELMATE travel package recommendation system | |
| CN110035111A (en) | A kind of method for pushing and device dispensing entity |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170613 |
|
| WD01 | Invention patent application deemed withdrawn after publication |