CN108052534A - A kind of real estate based on geographical feature recommends method - Google Patents
A kind of real estate based on geographical feature recommends method Download PDFInfo
- Publication number
- CN108052534A CN108052534A CN201711131295.6A CN201711131295A CN108052534A CN 108052534 A CN108052534 A CN 108052534A CN 201711131295 A CN201711131295 A CN 201711131295A CN 108052534 A CN108052534 A CN 108052534A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- real estate
- mfrac
- project
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 23
- 239000011159 matrix material Substances 0.000 claims abstract description 16
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 15
- 230000006870 function Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000000205 computational method Methods 0.000 claims description 3
- 239000013068 control sample Substances 0.000 claims description 2
- 239000000047 product Substances 0.000 description 3
- 239000000523 sample Substances 0.000 description 3
- 239000013065 commercial product Substances 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000007418 data mining Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- 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
-
- 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
- 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
- G06Q50/10—Services
- G06Q50/16—Real estate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2216/00—Indexing scheme relating to additional aspects of information retrieval not explicitly covered by G06F16/00 and subgroups
- G06F2216/03—Data mining
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Tourism & Hospitality (AREA)
- Accounting & Taxation (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Remote Sensing (AREA)
- Other Investigation Or Analysis Of Materials By Electrical Means (AREA)
Abstract
The invention discloses a kind of real estates based on geographical feature to recommend method.First according to the similarity between real estate project attribute information, each real estate project of building geographical coordinate vector calculating, then according to the similarity between each real estate project, structure includes the regularization term of geographical feature.On the basis of weight matrix decomposition model, with reference to the regularization term comprising geographical feature, using the algorithm of stochastic gradient descent, learn the implicit features vector of user and real estate project, user finally is predicted to not clicking on the click frequency of real estate project using the inner product of user and real estate project implicit features vector, and possible interested real estate information list is provided to the user according to the predicted value.The present invention is on the basis of weight matrix decomposition technique, merge the geographical feature information of real estate project, the implementation procedure that weight matrix decomposes is constrained, can more accurately learn the implicit features vector of real estate project, mitigate the cold start-up problem during real estate is recommended.
Description
Technical field
The invention belongs to data mining technology fields, and in particular to a kind of real estate based on geographical feature recommends method.
Background technology
In recent years, domestic real estate market development is swift and violent, becomes the important force that GDP is pulled to increase.Such as 2015, premises
The proportion that production value added accounts for Chinese entire GDP reaches 6.1%;2016, the GDP proportions of real estate value added further increased,
Rise to 6.5%.It purchases house in order to facilitate house purchaser, occurs some network applications for being absorbed in house property information on internet, such as
House365, room net is searched.In these network applications, service provider has provided building developer, geographical location, valency to the user
The information such as lattice, plot ratio.But in face of substantial amounts of building information, user is difficult that find oneself from building information interested
Building, i.e. user faces serious problem of information overload.Real estate commending system by analyze user history access record,
Attribute information of real estate etc. excavates hiding user preference, provides personalized real estate information service to the user, becomes solution
The certainly important means of real estate information overload problem.Real estate is recommended on the one hand to meet the personalized house-purchase demand of user, subtracts
The problem of information overload that light user faces, on the other hand, real estate commending system helps real estate information service provider to keep
Consumer loyalty degree increases the operating income of enterprise.Therefore, real estate recommendation is played the part of more and more important in real estate network application
Role.
Commending system is furtherd investigate by academia, and is used widely in e-commerce system, such as Amazon
In commercial product recommending, the film in Netflix recommends, the music in Last.fm recommend and Taobao in commercial product recommending etc..Although
Idea intuitively is to apply traditional proposed algorithm to recommend field in real estate, still, different from traditional commending system
Field, real estate commending system have following particular feature:1) implicit feedback click data:In traditional commending system,
User generally expresses the preference of user using explicit scoring.User's scoring is higher, and expression user is more satisfied.It is pushed away different from traditional
System is recommended, in real estate recommendation, the preference of user is the implicit click data expression by user.2) geographic influence:Tradition pushes away
It recommends in system, project (film, commodity, music etc.) does not have geographical attribute, but in real estate recommendation, each real estate has
Specific geographical location, and during user's selection real estate project, the influence in real estate project geographical location is suffered from, is such as selected
Near from urban district, the real estate project of traffic convenience or selection periphery are equipped with high quality primary school, the school district room in junior middle school.3) price
It influences:User is not required to consider the price factor of film and music when to film, music scoring, but purchaser is when buying house,
Whether real estate project price becomes user to the interested key factor of building.Up to the present, seldom research work concern
Real estate recommends problem.
The content of the invention
Existing problem and shortage for the above-mentioned prior art, the present invention propose that the real estate based on geographical feature is recommended to calculate
Method, emphasis calculate the similarity between real estate project, structure includes geographical feature according to the geographical feature of real estate project
Regularization term constrains weight matrix decomposable process, more accurately to learn the implicit features of user and real estate project vector,
So as to provide the performance of real estate project recommendation.
To achieve the above object, the technical solution adopted by the present invention is a kind of real estate recommendation side based on geographical feature
Method specifically comprises the steps of:
Step 1:According between real estate project attribute information, each real estate project of building geographical coordinate vector calculating
Similarity;
Step 2:According to the similarity between each real estate project, structure includes the regularization term of geographical feature;
Step 3:On the basis of weight matrix decomposition model, with reference to the regularization term comprising geographical feature, using boarding steps
The algorithm declined is spent, learns the implicit features vector of user and real estate project;
Step 4:Predict user to not clicking on real estate item using the inner product of user and real estate project implicit features vector
Purpose clicks on the frequency, and provides possible interested real estate information list to the user according to the predicted value.
Further, in above-mentioned steps 1, the similarity calculated between real estate project is to pass through the following formula:
Wherein, xiAnd xfRepresent building hiWith building hfGeographical coordinate vector, i.e., [longitude, latitude], δ is constant, upper
State in building calculating formula of similarity, the distance between similarity sim (i, f) and building between two buildings | | xi-xf||2
Inversely, with distance | | xi-xf||2Increase, the similarity sim (i, f) between building reduces.
Preferably, the value of above-mentioned δ is usually 0.01.
Further, regularization term described in above-mentioned steps 2 be two kinds of regularization terms for including geographical feature, i.e. Average-
Based regularization terms and Individual-based regularization terms, wherein structure Average-based geography regularization terms are roots
According to the following formula:
Wherein, λgIt is regularization term control parameter, N is the quantity of real estate project in commending system,It represents
Frobenius normal forms, column vector qiAnd qfThe implicit features vector of project i and project f is represented respectively, and N (i) is represented and project hi
Similar project set, the similarity between sim (i, f) real estate project i and real estate project f, Average-based are geographical
Regularization term minimizes the distance of the implicit features vector of building implicit features vector sum building similar to its;
It is according to the following formula to build Individual-based geography regularization terms:
Wherein, λgIt is regularization term control parameter, N is the quantity of real estate project in commending system, and N (i) is represented and item
Mesh hiSimilar project set, the similarity between sim (i, f) real estate project i and real estate project f,It represents
Frobenius normal forms, column vector qiAnd qfThe implicit features vector of project i and project f, Individual-based are represented respectively
Geographical regularization term causes the building that distance is near, similarity is high, and the distance of their implicit features vector is small rather than causes every
The implicit features vector of a building is intended to the weighted average of similar building implicit features vector.
Further, it is described on the basis of weight matrix decomposition model combine the regularization term comprising geographical feature use with
The Algorithm Learning user of machine gradient decline and the implicit features vector of real estate project are specially:In weight matrix decomposition model base
The target formula after Average-based geography regularization terms is merged on plinth is:
Wherein λgIt is regularization term control parameter, N (i) is represented and building hiSimilar building set, ruiRepresenting user u is
It is no to access project i, i.e. rui∈ { 0,1 },Represent Frobenius normal forms, regularization termWithFor avoiding plan
It closes, λ1And λ2For regularization term parameter, for controlling influence of the regularization term to implicit features vector, column vector qiAnd qfRespectively
The implicit features of expression project i and project f vector, the similarity between sim (i, f) real estate project i and real estate project f,
wuiConfidence level of the user to project preference is reflected, calculation formula is:
Wherein, a be control sample positive example weight parameter, cuiRepresent click frequencies of the user u on real estate project i,
It is set by the weight of above formula, after merging Individual-based geography regularization terms on the basis of weight matrix decomposition model
Target formula be:
Feature vector is hidden using stochastic gradient descent algorithm study user and building hides feature vector, object function LA
On puAnd qiPartial derivative be:
Wherein N-(i) represent to include building h in similar buildingiSet, using the Algorithm for Solving target of stochastic gradient descent
Function LILocal minimum solution, object function LIOn puAnd qiPartial derivative be:
Further, it is described to predict user to not clicking on room using the inner product of user and real estate project implicit features vector
The click frequency of real estate projects, and according to the predicted value provide to the user may interested real estate information list process
For user u clicks on the frequency to the prediction for not clicking on real estate project jComputational methods it is as follows:
Wherein, puAnd qjThe respectively implicit features of user u and real estate project j vector, for user u, has been calculated all
After the predicted value for not clicking on real estate project, according to predicted value, the k items for recommending predicted value high give user u.
Compared with prior art, the present invention has following technique effect:
1. weight matrix decomposition model regards real estate recommendation problem as OCCF problems, to sample positive example and missing item assignment
Different weights is suitable in the case of sample positive example is only included, and learns the implicit features vector of user and real estate project.
Therefore, WRMF models are more suited to processing real estate recommendation problem, and obtain preferably performance.
2. on the basis of weight matrix decomposition technique, the geographical feature information of real estate project is merged, constrains weight square
The implementation procedure that battle array is decomposed so that closely located real estate project has similar implicit features vector, can be more accurate
Study real estate project implicit features vector, mitigate real estate recommend in cold start-up problem.
Description of the drawings
Fig. 1 is the flow chart that the real estate provided by the invention based on geographical feature recommends method.
Specific embodiment
The specific implementation of the present invention is further described in detail in conjunction with attached drawing.
As shown in Figure 1, the invention discloses the real estates based on geographical feature to recommend method, comprise the steps of:
According to real estate project attribute information, each real estate project is calculated according to building geographical coordinate vector for step 1)
Similarity between (building);
Step 2), according to the similarity between each real estate project (building), structure includes the regularization of geographical feature
;
Step 3), on the basis of weight matrix decomposition model, with reference to the regularization term comprising geographical feature, using boarding steps
The algorithm declined is spent, learns the implicit features vector of user and real estate project;
Step 4) predicts user to not clicking on real estate item using the inner product of user and real estate project implicit features vector
Purpose clicks on the frequency, and provides possible interested real estate information list to the user according to predicted value.
In the step 1), the calculating formula of similarity between real estate project (building) is as follows:
Wherein, xiAnd xfRepresent building hiWith building hfGeographical coordinate vector, i.e., [longitude, latitude].δ is constant, usually
It is arranged to 0.01.In above-mentioned building calculating formula of similarity, between the similarity sim (i, f) and building between two buildings
Distance | | xi-xf||2Inversely, with distance | | xi-xf||2Increase, the similarity sim (i, f) between building subtracts
It is small.
In the step 2), two kinds of regularization terms for including geographical feature are designed, are respectively Average-based regularizations
Item and Individual-based regularization terms.
Average-based geography regularization terms are built according to the following formula:
Wherein, λgIt is regularization term control parameter, N is the quantity of real estate project in commending system,It represents
Frobenius normal forms, column vector qiAnd qfThe implicit features vector of project (building) i and project (building) f is represented respectively.N(i)
It represents and building hiSimilar building set.Between sim (i, f) real estate project (building) i and real estate project (building) f
Similarity.Average-based geography regularization term minimizes the implicit spy of building implicit features vector sum building similar to its
Levy the distance of vector.In other words, Average-based geography regularization term causes the implicit features vector and phase of each building
Weighted average like building implicit features vector is similar as far as possible.
Individual-based geography regularization terms are built according to the following formula:
Wherein, λgIt is regularization term control parameter, N is the quantity of real estate project in commending system, and N (i) is represented and building
Disk hiSimilar building set.Similarity between sim (i, f) real estate project (building) i and real estate project (building) f.Represent Frobenius normal forms, column vector qiAnd qfRespectively represent project (building) i and project (building) f implicit features to
Amount.Individual-based geography regularization terms cause the building that distance is near, similarity is high, their implicit features vector
Apart from small rather than the implicit features vector of each building is caused to be intended to the weighted average of similar building implicit features vector.
In the step 3), after merging Average-based geography regularization terms on the basis of weight matrix decomposition model
Target formula be:
Wherein λgIt is regularization term control parameter, N (i) is represented and building hiSimilar building set.ruiRepresenting user u is
It is no to access project i, i.e. rui∈{0,1}。Represent Frobenius normal forms.Regularization termWithFor avoiding plan
It closes.λ1And λ2For regularization term parameter, for controlling influence of the regularization term to implicit features vector.Column vector qiAnd qfRespectively
The implicit features of expression project (building) i and project (building) f vector.Sim (i, f) real estate project (building) i and real estate
Similarity between project (building) f.wuiConfidence level of the user to project preference is reflected, calculation formula is:
Wherein, a is the weight parameter for controlling sample positive example.cuiRepresent click frequencies of the user u on real estate project i.
It is set by the weight of above formula.
Target formula after merging Individual-based geography regularization terms on the basis of weight matrix decomposition model
For:
Feature vector is hidden using stochastic gradient descent algorithm study user and building hides feature vector.Object function LA
On puAnd qiPartial derivative be:
Wherein N-(i) represent to include building h in similar buildingiSet.
Using the Algorithm for Solving object function L of stochastic gradient descentILocal minimum solution.Object function LIOn puAnd qi's
Partial derivative is:
In the step 4), user u clicks on the frequency to the prediction for not clicking on real estate project jComputational methods such as
Under:
For user u, after all predicted values for not clicking on real estate project have been calculated, according to predicted value, recommend predicted value
High k items give user u.
Those skilled in the art of the present technique are it is understood that unless otherwise defined, all terms used herein are (including skill
Art term and scientific terminology) there is the meaning identical with the general understanding of the those of ordinary skill in fields of the present invention.Also
It should be understood that those terms such as defined in the general dictionary should be understood that with in the context of the prior art
The consistent meaning of meaning, and unless defined as here, will not be explained with the meaning of idealization or overly formal.
Above-described specific embodiment has carried out the purpose of the present invention, technical solution and advantageous effect further
It is described in detail, it should be understood that the foregoing is merely the specific embodiments of the present invention, is not limited to this hair
Bright, within the spirit and principles of the invention, any modification, equivalent substitution, improvement and etc. done should be included in the present invention
Protection domain within.
Claims (6)
1. a kind of real estate based on geographical feature recommends method, which is characterized in that comprises the steps of:
Step 1:According to similar between real estate project attribute information, each real estate project of building geographical coordinate vector calculating
Degree;
Step 2:According to the similarity between each real estate project, structure includes the regularization term of geographical feature;
Step 3:On the basis of weight matrix decomposition model, with reference to the regularization term comprising geographical feature, using under stochastic gradient
The algorithm of drop learns the implicit features vector of user and real estate project;
Step 4:Predict user to not clicking on real estate project using the inner product of user and real estate project implicit features vector
The frequency is clicked on, and possible interested real estate information list is provided to the user according to the predicted value.
2. the real estate according to claim 1 based on geographical feature recommends method, which is characterized in that in the step 1,
The similarity calculated between real estate project is to pass through the following formula:
<mrow>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msup>
<mi>e</mi>
<mfrac>
<mrow>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>x</mi>
<mi>f</mi>
</msub>
<mo>|</mo>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
<msup>
<mi>&delta;</mi>
<mn>2</mn>
</msup>
</mfrac>
</msup>
</mrow>
Wherein, xiAnd xfRepresent building hiWith building hfGeographical coordinate vector, i.e., [longitude, latitude], δ is constant, in above-mentioned building
In disk calculating formula of similarity, the distance between similarity sim (i, f) and building between two buildings | | xi-xf||2Into anti-
Than relation, with distance | | xi-xf||2Increase, the similarity sim (i, f) between building reduces.
3. the real estate according to claim 2 based on geographical feature recommends method, which is characterized in that the value of the δ
For 0.01.
4. the real estate according to claim 1 based on geographical feature recommends method, which is characterized in that in the step 2
The regularization term is two kinds of regularization terms for including geographical feature, i.e. Average-based regularization terms and Individual-
Based regularization terms, wherein structure Average-based geography regularization terms are according to the following formula:
<mrow>
<mfrac>
<msub>
<mi>&lambda;</mi>
<mi>g</mi>
</msub>
<mn>2</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mfrac>
<mrow>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>q</mi>
<mi>f</mi>
</msub>
</mrow>
<mrow>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
</mrow>
Wherein, λgIt is regularization term control parameter, N is the quantity of real estate project in commending system,Represent Frobenius
Normal form, column vector qiAnd qfThe implicit features vector of project i and project f is represented respectively, and N (i) is represented and project hiSimilar project
Set, the similarity between sim (i, f) real estate project i and real estate project f, Average-based geography regularization term is most
The distance of the implicit features vector of smallization building implicit features vector sum building similar to its;
It is according to the following formula to build Individual-based geography regularization terms:
<mrow>
<mfrac>
<msub>
<mi>&lambda;</mi>
<mi>g</mi>
</msub>
<mn>2</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>q</mi>
<mi>f</mi>
</msub>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
<mo>,</mo>
</mrow>
Wherein, λgIt is regularization term control parameter, N is the quantity of real estate project in commending system, and N (i) is represented and project hiPhase
As project set, the similarity between sim (i, f) real estate project i and real estate project f,Represent Frobenius models
Formula, column vector qiAnd qfThe implicit features vector of project i and project f, Individual-based geography regularization terms are represented respectively
So that apart from the high building of near, similarity, the distance of their implicit features vector is small rather than causes the implicit of each building
Feature vector is intended to the weighted average of similar building implicit features vector.
5. the real estate according to claim 4 based on geographical feature recommends method, which is characterized in that described in weight square
Algorithm Learning user and room of the regularization term comprising geographical feature using stochastic gradient descent are combined on the basis of battle array decomposition model
The implicit features vector of real estate projects is specially:Average-based geography is being merged on the basis of weight matrix decomposition model just
Then changing the target formula after item is:
<mrow>
<msup>
<mi>L</mi>
<mi>A</mi>
</msup>
<mo>=</mo>
<munder>
<mi>min</mi>
<mrow>
<mi>P</mi>
<mo>,</mo>
<mi>Q</mi>
</mrow>
</munder>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>u</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>M</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>u</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<msub>
<mi>w</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>r</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>-</mo>
<msubsup>
<mi>p</mi>
<mi>u</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<mfrac>
<msub>
<mi>&lambda;</mi>
<mn>1</mn>
</msub>
<mn>2</mn>
</mfrac>
<mo>|</mo>
<mo>|</mo>
<mi>P</mi>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<mfrac>
<msub>
<mi>&lambda;</mi>
<mn>2</mn>
</msub>
<mn>2</mn>
</mfrac>
<mo>|</mo>
<mo>|</mo>
<mi>Q</mi>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<mfrac>
<msub>
<mi>&lambda;</mi>
<mi>g</mi>
</msub>
<mn>2</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mfrac>
<mrow>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>q</mi>
<mi>f</mi>
</msub>
</mrow>
<mrow>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
</mrow>
Wherein λgIt is regularization term control parameter, N (i) is represented and building hiSimilar building set, ruiRepresent whether user u visits
Asked project i, i.e. rui∈ { 0,1 },Represent Frobenius normal forms, regularization termWithFor avoiding over-fitting,
λ1And λ2For regularization term parameter, for controlling influence of the regularization term to implicit features vector, column vector qiAnd qfIt represents respectively
The implicit features of project i and project f vector, the similarity between sim (i, f) real estate project i and real estate project f, wuiInstead
Confidence level of the user to project preference is reflected, calculation formula is:
<mrow>
<msub>
<mi>w</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mn>1</mn>
<mo>+</mo>
<mi>a</mi>
<mo>&times;</mo>
<msub>
<mi>c</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mtable>
<mtr>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
</mrow>
</mtd>
<mtd>
<mrow>
<msub>
<mi>c</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>></mo>
<mn>0</mn>
<mo>,</mo>
</mrow>
</mtd>
</mtr>
</mtable>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mn>1</mn>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>o</mi>
<mi>t</mi>
<mi>h</mi>
<mi>e</mi>
<mi>r</mi>
<mi>w</mi>
<mi>i</mi>
<mi>s</mi>
<mi>e</mi>
<mo>,</mo>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
Wherein, a be control sample positive example weight parameter, cuiIt represents click frequencies of the user u on real estate project i, passes through
The weight of above formula is set, the mesh after merging Individual-based geography regularization terms on the basis of weight matrix decomposition model
Marking formula is:
<mrow>
<msup>
<mi>L</mi>
<mi>I</mi>
</msup>
<mo>=</mo>
<munder>
<mi>min</mi>
<mrow>
<mi>P</mi>
<mo>,</mo>
<mi>Q</mi>
</mrow>
</munder>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>u</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>M</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<msub>
<mi>w</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>r</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>-</mo>
<msubsup>
<mi>p</mi>
<mi>u</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<mfrac>
<msub>
<mi>&lambda;</mi>
<mn>1</mn>
</msub>
<mn>2</mn>
</mfrac>
<mo>|</mo>
<mo>|</mo>
<mi>P</mi>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<mfrac>
<msub>
<mi>&lambda;</mi>
<mn>2</mn>
</msub>
<mn>2</mn>
</mfrac>
<mo>|</mo>
<mo>|</mo>
<mi>Q</mi>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<mfrac>
<msub>
<mi>&lambda;</mi>
<mi>g</mi>
</msub>
<mn>2</mn>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>q</mi>
<mi>f</mi>
</msub>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
<mo>,</mo>
</mrow>
Feature vector is hidden using stochastic gradient descent algorithm study user and building hides feature vector, object function LAOn
puAnd qiPartial derivative be:
<mrow>
<mfrac>
<mrow>
<mo>&part;</mo>
<msup>
<mi>L</mi>
<mi>A</mi>
</msup>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>p</mi>
<mi>u</mi>
</msub>
</mrow>
</mfrac>
<mo>=</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msubsup>
<mi>p</mi>
<mi>u</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>r</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mn>1</mn>
</msub>
<msub>
<mi>p</mi>
<mi>u</mi>
</msub>
<mo>,</mo>
</mrow>
<mfenced open = "" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mfrac>
<mrow>
<mo>&part;</mo>
<msup>
<mi>L</mi>
<mi>A</mi>
</msup>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
</mrow>
</mfrac>
<mo>=</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msubsup>
<mi>p</mi>
<mi>u</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>r</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<msub>
<mi>p</mi>
<mi>u</mi>
</msub>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mn>1</mn>
</msub>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mi>g</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mfrac>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>q</mi>
<mi>f</mi>
</msub>
</mrow>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mi>g</mi>
</msub>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>g</mi>
<mo>&Element;</mo>
<msup>
<mi>N</mi>
<mo>-</mo>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mfrac>
<mrow>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>g</mi>
<mo>)</mo>
</mrow>
<mrow>
<mo>(</mo>
<msub>
<mi>q</mi>
<mi>g</mi>
</msub>
<mo>-</mo>
<mfrac>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>g</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>g</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>q</mi>
<mi>f</mi>
</msub>
</mrow>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>g</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>g</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msub>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>g</mi>
<mo>)</mo>
</mrow>
</mrow>
</msub>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>g</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>,</mo>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
Wherein N-(i) represent to include building h in similar buildingiSet, using the Algorithm for Solving object function of stochastic gradient descent
LILocal minimum solution, object function LIOn puAnd qiPartial derivative be:
<mrow>
<mfrac>
<mrow>
<mo>&part;</mo>
<msup>
<mi>L</mi>
<mi>I</mi>
</msup>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>p</mi>
<mi>u</mi>
</msub>
</mrow>
</mfrac>
<mo>=</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msubsup>
<mi>p</mi>
<mi>u</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>r</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mn>1</mn>
</msub>
<msub>
<mi>p</mi>
<mi>u</mi>
</msub>
<mo>,</mo>
</mrow>
<mrow>
<mtable>
<mtr>
<mtd>
<mrow>
<mfrac>
<mrow>
<mo>&part;</mo>
<msup>
<mi>L</mi>
<mi>I</mi>
</msup>
</mrow>
<mrow>
<mo>&part;</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
</mrow>
</mfrac>
<mo>=</mo>
<msub>
<mi>w</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msubsup>
<mi>p</mi>
<mi>u</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>r</mi>
<mrow>
<mi>u</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<msub>
<mi>p</mi>
<mi>u</mi>
</msub>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mn>1</mn>
</msub>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mi>g</mi>
</msub>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>f</mi>
<mo>&Element;</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<mrow>
<mo>(</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>q</mi>
<mi>f</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mi>g</mi>
</msub>
<munder>
<mo>&Sigma;</mo>
<mrow>
<mi>g</mi>
<mo>&Element;</mo>
<msup>
<mi>N</mi>
<mo>-</mo>
</msup>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mi>s</mi>
<mi>i</mi>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>g</mi>
<mo>)</mo>
</mrow>
<mrow>
<mo>(</mo>
<msub>
<mi>q</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>q</mi>
<mi>g</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>.</mo>
</mrow>
</mtd>
</mtr>
</mtable>
<mo>.</mo>
</mrow>
6. the real estate according to claim 5 based on geographical feature recommends method, which is characterized in that described to use user
With the inner product of real estate project implicit features vector prediction user to not clicking on the click frequency of real estate project, and it is pre- according to this
The process that measured value provides possible interested real estate information list to the user is that user u was not to clicking on real estate project j's
The frequency is clicked in predictionComputational methods it is as follows:
<mrow>
<msub>
<mover>
<mi>R</mi>
<mo>~</mo>
</mover>
<mrow>
<mi>u</mi>
<mi>j</mi>
</mrow>
</msub>
<mo>=</mo>
<msubsup>
<mi>p</mi>
<mi>u</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>q</mi>
<mi>j</mi>
</msub>
</mrow>
Wherein, puAnd qjThe respectively implicit features of user u and real estate project j vector, for user u, has been calculated all non-points
After hitting the predicted value of real estate project, according to predicted value, the k items for recommending predicted value high give user u.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711131295.6A CN108052534A (en) | 2017-11-15 | 2017-11-15 | A kind of real estate based on geographical feature recommends method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711131295.6A CN108052534A (en) | 2017-11-15 | 2017-11-15 | A kind of real estate based on geographical feature recommends method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108052534A true CN108052534A (en) | 2018-05-18 |
Family
ID=62119116
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711131295.6A Pending CN108052534A (en) | 2017-11-15 | 2017-11-15 | A kind of real estate based on geographical feature recommends method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108052534A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108830507A (en) * | 2018-06-29 | 2018-11-16 | 成都数之联科技有限公司 | A kind of food safety risk method for early warning |
CN109670113A (en) * | 2018-12-20 | 2019-04-23 | 重庆锐云科技有限公司 | A kind of source of houses recommended method, device and server |
CN111949883A (en) * | 2020-08-24 | 2020-11-17 | 贝壳技术有限公司 | House resource recommendation method and device, computer readable storage medium and electronic equipment |
CN111966907A (en) * | 2020-08-21 | 2020-11-20 | 贝壳技术有限公司 | User preference cold start method, device, medium and electronic equipment |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106294859A (en) * | 2016-08-22 | 2017-01-04 | 南京邮电大学盐城大数据研究院有限公司 | A kind of item recommendation method decomposed based on attribute coupling matrix |
CN106503028A (en) * | 2015-09-08 | 2017-03-15 | 北京邮电大学 | Recommend method and system |
-
2017
- 2017-11-15 CN CN201711131295.6A patent/CN108052534A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106503028A (en) * | 2015-09-08 | 2017-03-15 | 北京邮电大学 | Recommend method and system |
CN106294859A (en) * | 2016-08-22 | 2017-01-04 | 南京邮电大学盐城大数据研究院有限公司 | A kind of item recommendation method decomposed based on attribute coupling matrix |
Non-Patent Citations (1)
Title |
---|
余永红等: "基于Ranking的泊松矩阵分解兴趣点推荐算法", 《计算机研究与发展》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108830507A (en) * | 2018-06-29 | 2018-11-16 | 成都数之联科技有限公司 | A kind of food safety risk method for early warning |
CN109670113A (en) * | 2018-12-20 | 2019-04-23 | 重庆锐云科技有限公司 | A kind of source of houses recommended method, device and server |
CN111966907A (en) * | 2020-08-21 | 2020-11-20 | 贝壳技术有限公司 | User preference cold start method, device, medium and electronic equipment |
CN111949883A (en) * | 2020-08-24 | 2020-11-17 | 贝壳技术有限公司 | House resource recommendation method and device, computer readable storage medium and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Crane-Droesch | Machine learning methods for crop yield prediction and climate change impact assessment in agriculture | |
US20180292959A1 (en) | Systems and methods for analyzing performance of an entity | |
US7774227B2 (en) | Method and system utilizing online analytical processing (OLAP) for making predictions about business locations | |
Tiefelsdorf et al. | A variance-stabilizing coding scheme for spatial link matrices | |
Kuo et al. | Building and evaluating a location-based service recommendation system with a preference adjustment mechanism | |
US20130339350A1 (en) | Ranking Search Results Based on Click Through Rates | |
CN108052534A (en) | A kind of real estate based on geographical feature recommends method | |
Yang et al. | Analysis of linkage effects among industry sectors in China’s stock market before and after the financial crisis | |
CN105469263A (en) | Commodity recommendation method and device | |
CN103150667B (en) | A kind of personalized recommendation method based on body construction | |
CN108198045A (en) | The design method of mixing commending system based on e-commerce website data mining | |
Qiao et al. | Can network linkage effects determine return? Evidence from Chinese stock market | |
CN113157752A (en) | Scientific and technological resource recommendation method and system based on user portrait and situation | |
Simlai | Estimation of variance of housing prices using spatial conditional heteroskedasticity (SARCH) model with an application to Boston housing price data | |
Kang et al. | Integration of Internet search data to predict tourism trends using spatial-temporal XGBoost composite model | |
Aggarwal et al. | Examining the factors influencing fintech adoption behaviour of gen Y in India | |
Cohen et al. | Simulating tornado probability and tornado wind speed based on statistical models | |
US20140279053A1 (en) | System and method for applying spatially indexed data to digital advertising bids | |
Buck et al. | Forecasting residential water consumption in California: rethinking model selection | |
Saeed et al. | The poverty impacts of labor heat stress in West Africa under a warming climate | |
Yoon et al. | Item recommendation by predicting bipartite network embedding of user preference | |
CN110413880A (en) | Based on user's personality single classification collaborative filtering method layered | |
CN107274247A (en) | Wisdom based on cloud computing perceives recommendation method | |
CN109978306B (en) | Data processing method and device | |
CN116485503A (en) | Commodity combination recommendation method, device, equipment and medium thereof |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180518 |