CN109658192A - A kind of source of houses recommended method and server - Google Patents
A kind of source of houses recommended method and server Download PDFInfo
- Publication number
- CN109658192A CN109658192A CN201811564897.5A CN201811564897A CN109658192A CN 109658192 A CN109658192 A CN 109658192A CN 201811564897 A CN201811564897 A CN 201811564897A CN 109658192 A CN109658192 A CN 109658192A
- Authority
- CN
- China
- Prior art keywords
- house type
- house
- recommended
- client
- matched
- 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 37
- 239000013598 vector Substances 0.000 claims description 24
- 238000004891 communication Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 230000009286 beneficial effect Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 7
- 230000006399 behavior Effects 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 4
- 238000011002 quantification Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 3
- 238000013139 quantization Methods 0.000 description 3
- 238000012216 screening Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000001737 promoting effect Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000013011 mating Effects 0.000 description 1
- 230000001568 sexual effect Effects 0.000 description 1
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
- 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
Landscapes
- Business, Economics & Management (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides a kind of source of houses recommended method and server, this method comprises: when receiving recommendation enabled instruction, client to be recommended is obtained to the history access record of each house house type, history access record includes at least one of house type access times, number of clicks, access duration;Client to be recommended is calculated respectively to the preference value of each house house type according to history access record, according to each preference value from each house house type, selects a house house type as the benchmark house type of client to be recommended;The similarity of each house type to be matched and benchmark house type in house type library to be matched is calculated, similarity is selected to meet the house type to be matched of preset relation, as recommended candidate house type, the recommendation of recommended candidate house type is sent to client to be recommended.The intelligent recommendation of house house type is realized, client's browsing can be made more targeted, the room time is seen in reduction;In addition, carrying out intelligent recommendation for customer priorities, it is also beneficial to promote house conclusion of the business efficiency.
Description
Technical field
The present invention relates to real estate technical field more particularly to a kind of source of houses recommended method and servers.
Background technique
Displaying live view is carried out to houseclearing for the ease of house buyer to check, is avoided the need for by seeing that room needs to consume on the spot
The problem of taking the plenty of time, current more intermediary service provider, house developer etc. develop network and see room platform, and user can be straight
It connects and houseclearing is browsed on network, such as building position, price, house type, periphery are mating etc., allow user at any time
House multidate information is understood everywhere.
Since network is seen there are a large amount of information of real estate on room platform, user can not carry out browsing to whole house house types and look into
It seeing, user is typically only capable to browsing part house house type, this is possible to the case where leading to miss satisfied house house type, because
This is unfavorable for promoting user experience, is also unfavorable for promoting house conclusion of the business efficiency.In order to enable user's browse efficiency is higher, give
User brings the selection for being more satisfied with house type, how to determine the satisfied house house type of user and carries out the intelligence of house house type
It can recommend, then seem particularly significant.
Summary of the invention
A kind of source of houses recommended method provided by the invention and server, mainly solving the technical problems that: how to realize room
The intelligent recommendation of room house type better meets user to the browsing demand of house house type.
In order to solve the above technical problems, the present invention provides a kind of source of houses recommended method, comprising:
Receive recommend enabled instruction when, obtain client to be recommended to the history access record of each house house type, it is described
History access record includes at least one of house type access times, number of clicks, access duration;
According to the history access record, the client to be recommended is calculated respectively to the preference value of each house house type,
A house house type is selected from each house house type according to each preference value, the benchmark family as the client to be recommended
Type;
The similarity of each house type to be matched and the benchmark house type in house type library to be matched is calculated, similarity is selected to meet
Recommended candidate house type recommendation is sent to described to be recommended by the house type to be matched of preset relation as recommended candidate house type
Client.
Further, described according to the history access record, the client to be recommended is calculated respectively to each house
The preference value of house type includes:
The client to be recommended is added summation to the access times of the house house type, number of clicks, access duration, is obtained
To the client to be recommended to the preference value of the house house type.
Further, described to select a house house type from each house house type according to each preference value, as
The benchmark house type of the client to be recommended includes:
In each preference value, the maximum house house type of preference value, the base as the client to be recommended are selected
Quasi- house type;When the maximum house house type of preference value is used as the client to be recommended there are when two or more, randomly choosing one
Benchmark house type.
It is further, described that recommended candidate house type recommendation is sent to before the client to be recommended, further includes:
It detects in the recommended candidate house type, if there are the house house types that the client to be recommended had browsed, if
In the presence of will exclude other house types outside the house house type browsed in the recommended candidate house type, recommend to be sent to
The client to be recommended;If it does not exist, then recommended candidate house type recommendation is directly sent to the client to be recommended.
Further, it is described receive recommend enabled instruction include:
When the history access record for detecting the client to be recommended updates, determine that receiving recommendation starting refers to
It enables;
And/or when prefixed time interval reaches, determination receives recommendation enabled instruction.
Further, the similarity packet of each house type to be matched calculated in house type library to be matched and the benchmark house type
It includes:
Using cosine similarity formula calculate each house type to be matched in the house type to be matched and the benchmark house type it
Between cosine similarity in following at least one house type feature: place city number, the longitude and latitude of the affiliated building of house type, house type
Area, house type room number, house type unit price, house type direction number;The cosine similarity formula is as follows:
Wherein, the cos θijFor the cosine similarity between house type i and house type j feature vector, the n is house type feature
Number, the x is characterized value;
Each band matching house type in the house type to be matched is calculated using Euclidean distance formula and between the benchmark house type
Distance conformability degree in following at least one house type feature: place city number, the longitude and latitude of the affiliated building of house type, house type face
Product, house type room number, house type unit price, house type direction number;The Euclidean distance degree formula is as follows:
Wherein, the dijFor the distance between house type i and house type j feature vector similarity.
Further, in the similarity of each house type to be matched calculated in house type library to be matched and the benchmark house type
Before, further includes:
Detect whether the client to be recommended has investment intent, if so, for the benchmark house type and it is described to
With each house type to be matched in house type library, its feature vector is weighted processing according to first set weighted value;If it is not, being then directed to
Each house type to be matched in the benchmark house type and the house type library to be matched, second set of weighted value of its feature vector is added
Power processing;The first set weighted value is different from second set of weighted value.
Further, in the similarity of each house type to be matched calculated in house type library to be matched and the benchmark house type
Before, further includes:
It obtains the client to be recommended and often uses position longitude and latitude, by the longitude and latitude of building belonging to the reality of the benchmark house type
Replace with the common position longitude and latitude.
Further, the selection similarity meets the house type to be matched of preset relation, includes: as recommended candidate house type
Each cosine similarity is ranked up according to descending sequence, M before selecting (M is more than or equal to 2)
House type to be matched corresponding to a cosine similarity is as the first house type set to be matched;By each Distance conformability degree according to having
Small to be ranked up to big sequence, house type to be matched corresponding to M Distance conformability degree is as the second house type to be matched before selecting
Set;The intersection between the described first house type set to be matched and the second house type set to be matched is determined, by the intersection
Corresponding house house type is as the recommended candidate house type.
The present invention also provides a kind of server, the server includes processor, memory and communication bus;
The communication bus is for realizing the connection communication between processor and memory;
The processor is for executing one or more program stored in memory, to realize described in any one as above
The source of houses recommended method the step of.
The beneficial effects of the present invention are:
A kind of source of houses recommended method and server provided according to the present invention, this method comprises: recommending starting receiving
When instruction, obtain client to be recommended to the history access record of each house house type, history access record include house type access times,
At least one of number of clicks, access duration;Client to be recommended is calculated respectively to each house house type according to history access record
Preference value select a house house type as the benchmark house type of client to be recommended according to each preference value from each house house type;
Calculate the similarity of each house type to be matched and benchmark house type in house type library to be matched, select similarity meet preset relation to
House type is matched, as recommended candidate house type, the recommendation of recommended candidate house type is sent to client to be recommended.By utilizing customer historical
Access record, and history access record can reflect out user to the preference of each house house type, so as to more accurately true
Determine customized benchmarks house type, based on the similarity of house type to be matched and benchmark house type in house type library to be matched, determines recommended candidate
House type realizes the intelligent recommendation of house house type, does not need client and all carries out browsing to all house types seen on room platform checking,
By server intelligent recommendation, client's browsing can be made more targeted, the room time is seen in reduction;In addition, inclined for client
Intelligent recommendation is carried out well, is also beneficial to promote house conclusion of the business efficiency.
Detailed description of the invention
Fig. 1 is a kind of source of houses recommended method flow diagram of the embodiment of the present invention one;
Fig. 2 is the house type direction quantization schematic diagram of the embodiment of the present invention one;
Fig. 3 is the server architecture schematic diagram of the embodiment of the present invention two;
Fig. 4 is the recommendation results schematic diagram one of the embodiment of the present invention two;
Fig. 5 is the recommendation results schematic diagram two of the embodiment of the present invention two.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, below by specific embodiment knot
Closing attached drawing, invention is further described in detail.It should be appreciated that specific embodiment described herein is only used to explain this
Invention, is not intended to limit the present invention.
Embodiment one:
In order to improve client's usage experience, the efficiency of client's Model of Buying House is promoted, now to see room platform as carrier, with house
House type provides a kind of source of houses recommended method using house type feature and customer information as content for minimum recommended unit.Referring to Figure 1,
Fig. 1 is a kind of source of houses recommended method flow diagram provided in this embodiment, and this method comprises the following steps:
S101, receive recommend enabled instruction when, obtain client to be recommended to the history access record of each house house type,
The history access record includes at least one of house type access times, number of clicks, access duration.
Client when network sees the enterprising having sexual intercourse room information browse of room platform, is used by user terminal (such as mobile phone, PAD etc.)
Family terminal can in real time or customer historical access record is sent to server and stored by timing.User in navigation process,
For oneself satisfied or interested house type, it will usually higher attention rate is shown, to the pass of each house house type
Note degree can be embodied in history access record.This programme is by making full use of customer historical access to record, progress data analysis, from
And realize the process of house house type intelligent recommendation, so that the house house type recommended is more in line with the preference of user, it can be more preferable
Ground meets user and browses demand.
Recommending enabled instruction includes: when the history access record of client to be recommended updates, and determination receives recommendation
Enabled instruction, to trigger the history access record for obtaining the client to be recommended;Optionally, when prefixed time interval reaches,
Determination receives recommendation enabled instruction.For example, in each interval of one day, carrying out house intelligent recommendation to each client to be recommended.
History access record includes but is not limited to house type access times, number of clicks, access duration.Wherein, house type accesses
Number can be understood as the number at interface where other interfaces of room platform in terms of network jump to the house house type.
S102, according to history access record, calculate client to be recommended respectively to the preference value of each house house type, according to it is each partially
Good value selects a house house type, the benchmark house type as client to be recommended from each house house type.
Client to be recommended is added summation to the access times of house house type, number of clicks, access duration, is obtained to be recommended
Preference value of the client to the house house type.For example, according to the following example of the history access record of the client to be recommended of acquisition:
Table 1
House house type | Access times | Number of clicks | Access duration |
A202 | 16 | 22 | 150 |
B223 | 25 | 36 | 360 |
C315 | 3 | 18 | 56 |
The user to be recommended can be obtained and there is access record, respectively A202, B223, C315 to three house house types.Wherein,
History access record to house house type A202 includes: that access times are 16 times, and number of clicks is 22 times, a length of 150 when access
Second, then preference value of the available client to be recommended to house house type A202 accordingly are as follows: 16+22+150=188.Together
Reason, can obtain the client to be recommended to the preference value of house house type B223 are as follows: 25+36+360=421;To the inclined of house house type C315
Good value are as follows: 3+18+56=77.
Then, according to the client to be recommended to the preference value of each house house type (here include A202, B223, C315), from
One house house type of middle selection, the benchmark house type as the client to be recommended.Optionally, in each preference value, preference value is selected
A maximum house house type, the benchmark house type as client to be recommended.Here, the maximum house house type of preference value is B223,
It therefore can be using house house type B223 as the benchmark house type of the client to be recommended.
In practical application, when the maximum house house type of preference value there are when two or more, randomly choose one as to
Recommend the benchmark house type of client.
S103, the similarity for calculating each house type to be matched and benchmark house type in house type library to be matched, select similarity to expire
The recommendation of recommended candidate house type is sent to client to be recommended as recommended candidate house type by the house type to be matched of sufficient preset relation.
The whole house type information seen on room platform are stored in house type library to be matched, the place city including each house type is compiled
Number, the longitude and latitude of the affiliated building of house type, house type area, house type room number, house type unit price, house type towards number etc..
Each house house type in house type library to be matched is matched with the benchmark house type, is calculated between the benchmark house type
Similarity, wherein calculate similarity process specifically use the following two kinds mode:
It is calculated using cosine similarity formula special in following at least one house type between each house type to be matched and benchmark house type
Cosine similarity in sign: place city number, the longitude and latitude of the affiliated building of house type, house type area, house type room number, house type list
Valence, house type direction number;The cosine similarity formula is as follows:
Wherein, x is characterized value, and n is the number of house type feature, cos θijMore than between house type i and house type j feature vector
String similarity.
For example, simultaneously to place city number, the longitude of the affiliated building of house type, latitude, house type area, house type room number,
House type unit price, house type are formed by feature vector towards number and are calculated:
Assuming that the house type feature vector of benchmark house type is (102,108,35,120,5,16500,1), the spy of house type to be matched
Levying vector is (102,110,36,150,5,18500,1.5), then this can be calculated by above-mentioned cosine similarity formula
Cosine similarity between house type to be matched and the benchmark house type.
Before carrying out similarity calculation, city where needing to carry out quantification treatment, such as house type to respective behavior feature,
It needs to carry out quantification treatment to each city name in advance, forms city number, different cities may correspond to different city numbers.
Optionally, the mode of quantization can be according to geographical positional relationship carry out, such as by Chinese city quartile southwest, south China, North China,
The areas such as northwest can use same number to the city for belonging to areal.Optionally, to the multiple cities for belonging to areal
City can also take different numbers, it is first determined the number of this area's central town, according at a distance from central town, away from
From bigger, number difference is away from bigger, and apart from smaller, number difference is away from smaller.Other quantification manners, the present embodiment can certainly be used
It is without limitation.
The longitude value range of the affiliated building of house type can be -180 degree to+180 degree, such as the longitude range of CHINESE REGION
Generally 73 degree to 136 degree of east longitude of east longitude, substantially 3 degree to 54 degree of north latitude of north latitude of latitude scope.
House type area should be greater than 0 (unit can be square metre).
House type room number can flexible quantization, such as the house type of three living rooms and one sitting room a kitchen and a guard according to the actual situation, room
Number can be quantified as 3+1+1+1=6;Such as the house type that four Room, one Room, one kitchen two is defended, room number are 4+1+1+2=8.
House type obtains house type towards number towards that can carry out quantification treatment according to mode as shown in Figure 2.For example, towards just
South to house type, house type towards number be 2;Towards the house type of direct north, house type direction is quantified as -2.
It should be appreciated that in practical applications, more or fewer house type features can be used, form the feature of different dimensions
Vector carries out the calculating of similarity.Such as similarity calculation only is carried out to house type area, house type room number, house type unit price, then
It that is to say that being formed by three-dimensional vector to house type to be matched and benchmark house type carries out similarity calculation.
Using Euclidean distance formula calculate each house type to be matched in house type to be matched and between benchmark house type such as down toward
Distance conformability degree in a kind of few house type feature: place city number, the longitude and latitude of the affiliated building of house type, house type area, house type
Room number, house type unit price, house type direction number;The Euclidean distance degree formula is as follows:
Wherein, dijFor the distance between house type i and house type j feature vector similarity.
In general, house type used by the two is characterized in identical when calculating similarity using cosine similarity and Euclidean distance
's.For example, being formed by when being calculated using cosine similarity using house type area, house type room number, house type unit price three
Three-dimensional feature vector is calculated, then equally using house type area, house type room number, house type list when calculating Euclidean distance
Valence three is formed by three-dimensional feature vector and calculates.
Before the similarity for calculating each house type to be matched and benchmark house type in house type library to be matched, can also first it obtain
Client to be recommended often uses position longitude and latitude, and the longitude and latitude of building belonging to the reality of benchmark house type is replaced with the common position longitude and latitude
Degree.
Recording obtained benchmark house type based on user's history access is necessary being, therefore the benchmark house type has phase
The attributive character answered, such as affiliated building latitude and longitude information.It, can be one despite what is obtained based on customer historical access record
Determine the Behavior preference for reflecting user in degree, compares and meet the needs of users.But it may be with the optimal house house type of user
Still there is certain gap, this gap may be mainly reflected on geographical location.For just needing client, house
Geographical location be particularly important.It, in this way could as far as possible it is therefore desirable to advanced optimize processing to benchmark house type
The accuracy recommended is improved, could more meet user actually sees room demand.
Therefore, this programme also further replaces the actual geographic position of the affiliated building of benchmark house type, is replaced
It is changed to client and often uses position longitude and latitude, obtain a virtual benchmark house type.Although replacing the benchmark house type behind geographical location not
It is necessary being, but the ideal for more meeting user is expected, therefore intelligent recommendation is carried out based on the virtual benchmark house type, it can be with
Preferably meet user sees room demand.
Wherein, obtaining client often uses the mode of position longitude and latitude that can not illustrate using existing any way this.
By the feature vector of each house type to be matched and benchmark house type, above-mentioned cosine similarity formula and Euclidean distance are utilized
Cosine similarity and Euclidean distance between each house type to be matched and benchmark house type is calculated in formula.
Selection similarity meets the house type to be matched of preset relation, includes: as recommended candidate house type
Each cosine similarity is ranked up according to descending sequence, M (M is more than or equal to 2) is a remaining before selecting
House type to be matched corresponding to string similarity is as the first house type set to be matched;Each Distance conformability degree is small to big according to having
Sequence is ranked up, and select house type to be matched corresponding to preceding M Distance conformability degree as the second house type set to be matched;It determines
Intersection between first house type set to be matched and the second house type set to be matched, using the corresponding house house type of intersection as recommendation
Candidate house type.
For example, house type to be matched has 10 in house type library to be matched, then by this 10 house types to be matched respectively with benchmark family
Type carries out similarity calculation, obtains 10 cosine similarities and 10 Distance conformability degrees;By this 10 cosine similarities according to
The descending sequence of numerical value is arranged, select 5 house types to be matched corresponding to preceding 5 cosine similarities as first to
Match house type set;By this 10 Distance conformability degrees according to have it is small arranged to big sequence, select preceding 5 Distance conformability degrees
5 corresponding house types to be matched are as the second house type set to be matched;Determine the first house type set to be matched with second to
With house type intersection of sets collection, and the house type to be matched that the intersection is included is as recommended candidate house type.
Before the similarity for calculating each house type to be matched and benchmark house type in house type library to be matched, can also first it detect
Whether client to be recommended has investment intent, if so, for each house type to be matched in benchmark house type and house type library to be matched,
Its house type feature is corresponded into numerical value and is weighted processing according to first set weighted value;If it is not, then for benchmark house type with it is to be matched
Its house type feature is corresponded to numerical value and is weighted processing according to second set of weighted value by each house type to be matched in house type library;Wherein
First set weighted value is different from second set of weighted value.
For example, the feature vector of benchmark house type is (a1, a2, a3, a4, a5, a6, a7), the feature vector of house type 1 to be matched
For (b1, b2, b3, b4, b5, b6, b7);Assuming that first set weighted value be [0.01,0.01,0.01,0.26,0.32,0.29,
0.10], second set of weighted value is [0.32,0.1,0.1,0.13,0.17,0.15,0.03], city number where respectively corresponding,
Latitude, longitude, house type area, house type room number, house type unit price, house type direction number;If client to be recommended has investment intent,
The feature vector (a1, a2, a3, a4, a5, a6, a7) of benchmark house type is then weighted processing according to first set weighted value, is obtained
(0.01a1,0.01a2,0.01a3,0.26a4,0.32a5,0.29a6,0.10a7), at the same by the feature of house type 1 to be matched to
Amount is weighted processing according to first set weighted value for (b1, b2, b3, b4, b5, b6, b7), obtain (0.01b1,0.01b2,
0.01b3,0.26b4,0.32b5,0.29b6,0.10b7), and then the calculating of similarity is carried out, namely calculate (0.01b1,
0.01b2,0.01b3,0.26b4,0.32b5,0.29b6,0.10b7) with (0.01a1,0.01a2,0.01a3,0.26a4,
0.32a5,0.29a6,0.10a7) between similarity.Similarly, if client to be recommended does not have investment intent, according to second
Set weighted value is handled.
House type recommends intrinsic geographic area limitation, city and position where when recommendation can consider house type.And
Entire customers can then divide into two classes: have investment intent, without investment intent.There is the client of investment intent to where building
City susceptibility is not high, and house type paid close attention to feature itself is directed to this kind of client if appropriate for investment, should weaken
The positional factor of house type;And be directed to the lead referral without investment intent, then need emphasis to consider the influence of positional factor.
Before the recommendation of recommended candidate house type is sent to client to be recommended, it can also first detect in recommended candidate house type, be
The no house house type browsed there are client to be recommended, and if it exists, by exclude to have browsed in recommended candidate house type
Other house types outside house house type, recommendation are sent to client to be recommended;If it does not exist, then directly recommended candidate house type is recommended to send out
Give client to be recommended.Client is recommended to avoid the house house type for browsing client, influences user experience.
The present embodiment provides a kind of source of houses recommended methods, by obtaining visitor to be recommended when receiving recommendation enabled instruction
To the history access record of each house house type, history access record includes house type access times, number of clicks, in access duration at family
At least one;Client to be recommended is calculated respectively to the preference value of each house house type, according to each preference according to history access record
Value selects a house house type as the benchmark house type of client to be recommended from each house house type;It calculates in house type library to be matched
Each house type to be matched and benchmark house type similarity, select similarity to meet the house type to be matched of preset relation, as recommendation
The recommendation of recommended candidate house type is sent to client to be recommended by candidate house type.By being recorded using customer historical access, and history is visited
It asks that record can reflect out the preference that user writes in reply to each house, determines benchmark house type, base so as to more accurately client
The similarity of house type to be matched and benchmark house type in house type library to be matched, determines recommended candidate house type, realizes house family
The intelligent recommendation of type does not need client and all carries out browsing to all house types seen on room platform checking, intelligently pushed away by server
It recommends, client's browsing can be made more targeted, the room time is seen in reduction;In addition, intelligent recommendation is carried out for customer priorities,
Be conducive to promote house conclusion of the business efficiency.
Embodiment two:
The present embodiment on the basis of example 1, provides a kind of server, for realizing described in above-described embodiment one
The step of source of houses recommended method.Fig. 3 is referred to, Fig. 3 is a kind of structural schematic diagram of server provided in this embodiment, the service
Device includes processor 31, memory 32 and communication bus 33;
Communication bus 33 is for realizing the connection communication between processor 31 and memory 32;
Processor 31 is for executing one or more program stored in memory 32, to realize as described in embodiment one
The source of houses recommended method the step of.
Optionally, memory 32 stores five parts: house type information bank, user are to the access record storehouse of house type, user
Positioning record storehouse, house type collection record library, the client store with investment intent.
Wherein, house type information bank contains the longitude and latitude of the unique identifier of house type, house type current state, the affiliated building of house type
Degree, place city (id), house type area, the house type of house type, house type unit price, house type direction.Wherein, " house type unit price " is with house type table
On the basis of, if price belongs to invalid price (for example 0, too small etc.), the reasonable average price of building where use the house type is as being somebody's turn to do
The unit price of house type;" direction " attribute is in the text that is recorded as of database, the herein inhabitation according to Chinese geographic orientation and the people
Habit, is quantified, as shown in Figure 2.
For user to the access record storehouse of house type, it includes user which, which contains the recent access habits and preference of client,
Identifier, user enter the detailed number of house type (such as record is house type access times) and some from building details page
The relevant function number of clicks of house type, access relevant to certain house type record duration (such as minute).
User positions record storehouse, which contains the customer information with house type collection behavior and the house type information of collection,
Item is excluded as the house type of customers used in subsequent recommendation system evaluation and the consequently recommended result of recommender system.
Client store with investment intent, which is the customer information record with investment intent, from the letter of filing of client
It is extracted in breath, contains the voip identifiers that house-purchase purposes contains " investment ", divide different clients in the process as subsequent recommendation
Group and the main foundation matched using different characteristic weight.
This system is the recommender system based on broad context, disclosure be house type feature in house type information bank with
The fusion of building geographic location feature.Recommending benchmark house type used is that feature of preference house type itself is browsed by user
And customer position information combined structure.By the way that combined structure benchmark house type and the house type in house type information bank are compared,
Recommended candidate item is calculated jointly using a variety of similarity measurement means (cosine similarity, Euclidean distance), is then carried out certain
The screening of rule, the final result as recommendation.In the process, it is contemplated that the weight distribution problem of house type feature, for not
With the client of investment intent, different feature weight proportions is formulated, to improve the precision of personalized recommendation.
Feature database data constructed itself have very big magnitude poor, and in recommendation process similarity measurement calculating pair
This is more sensitive, therefore has carried out standardization to used characteristic variable.Using being standardized as mean value in system
It is 0, the normal distribution that variance is 1 meets the value x of each feature: x~N (0,1).
Having carried out standardized data includes: the characteristic other than house type identifier, the use in house type information bank
Family positions record storehouse to the achievement data other than user identifier and house type identifier in house type access record storehouse, user
In the position data other than user identifier.
Benchmark house type is constructed, is the committed step for determining consequently recommended result and the true intention of client and whether meeting.It is logical
For often, the access behavior of client largely reflects the true intention of client.And the intention of client is in a broad sense
Not merely refer to house-purchase intention, it is also the same including client to the browsing intention of some house type of some buildings.Therefore, the visit of client is utilized
Behavior record is asked as the recommendation benchmark for being directed to client, is the strategy that can largely make user obtain intention product.
After the client destination of recommendation has been determined, can according to client's unique identifier from constructed " user is to house type
Access record storehouse " in extract the recent of the client (real-time recommendation strategy can be used when practical application, i.e. client's browsing behavior has
Update just iteration recommended candidate item immediately) access situation, benchmark family is then determined using the overall target that house type access records
Type.
The determination method of currently employed benchmark house type is shown below:
Wherein, indexijIt is j-th of index of i-th of house type, the house type number that n is browsed for the client.Utilize above formula
What is screened is that maximum house type of overall target in house type that the client is browsed, and in this, as the customized benchmarks
House type.
In addition, the location information for positioning longitude and latitude for client equally also joined system considerations.Based on recommendation
Real-time and distance priority etc. consider that introducing Customer Location information has certain meaning.But the position letter for needing client to position
Breath meets: it is identical as city where access house type that client positions city.
It is positioned in record storehouse from user, extracts the location information of selected client, substitute in the benchmark house type of above-mentioned construction
Corresponding position information, and using instead of the house type after location information as finally determining benchmark house type.
From the perspective of client's intention, house type feature selected by system has different influence degrees.Therefore, in root
It is extracted and the highest recommended candidate of benchmark house type similarity from house type information bank according to similitude according to the benchmark house type constructed
Before house type, weight proportion should be carried out to house type feature according to the intention of house purchaser.
In addition, house type recommends intrinsic geographic area limitation, city and position where when recommendation can consider house type
It sets.And entire customers, then it can divide into two classes: have investment intent, without investment intent.There is the client of investment intent to building
Place city susceptibility is not high, and house type paid close attention to feature itself is directed to this kind of client if appropriate for investment, should
Weaken the positional factor of house type;And be directed to the lead referral without investment intent, then need emphasis to consider the influence of positional factor.Cause
This has formulated two sets of weight proportions, has had the weight of investment intent to match for two class clients: [0.01,0.01,0.01,0.26,
0.32,0.29,0.10], the weight proportion of no investment intent: [0.32,0.1,0.1,0.13,0.17,0.15,0.03].
After constructing house type information bank, selected reference house type, determining the distribution of house type feature weight, i.e. progress house type is similar
Property measurement.
The characteristics of recommending in view of house type, it is desirable that between the feature vector of two similar house types, not only two vectors are as far as possible
In parallel, also to make its distance minimum.System uses cosine similarity and Euclidean distance as judging quota, comprehensively considers two kinds
Index screening goes out the recommended candidate item that similarity is best suitable for truth.
System finds out similitude seniority among brothers and sisters Top10 respectively, then asks two recommendations first with two kinds of method for measuring similarity
The intersection of candidate item set, to carry out candidate item screening.Then, recommend index according to two kinds of measure constructions: recommending
Index 1 recommends index 2.Wherein, recommending index 1 is cosine similarity, and recommending index 2 is the reformulations of Euclidean distance: 1/
(1+dij)。dijFor the Euclidean distance between house type i and house type j feature vector.
It is allocated different feature weights for the client of different investment intents, no investment intent then strengthens Customer Location
The influence of factor has the influence of investment intent then weakened locations factor.The recommendation results of two kinds of clients are as shown in Figure 4, Figure 5.Knot
Fruit displaying, the available expected results of recommendation carried out for different clients.
Obviously, those skilled in the art should be understood that each module of aforementioned present invention or each step can be with general
Computing device realizes that they can be concentrated on a single computing device, or be distributed in constituted by multiple computing devices
On network, optionally, they can be realized with the program code that computing device can perform, it is thus possible to be stored in
It is performed by computing device in computer storage medium (ROM/RAM, magnetic disk, CD), and in some cases, it can be with not
The sequence being same as herein executes shown or described step, or they are fabricated to each integrated circuit modules, or
Person makes multiple modules or steps in them to single integrated circuit module to realize.So the present invention is not limited to appoint
What specific hardware and software combines.
The above content is specific embodiment is combined, further detailed description of the invention, and it cannot be said that this hair
Bright specific implementation is only limited to these instructions.For those of ordinary skill in the art to which the present invention belongs, it is not taking off
Under the premise of from present inventive concept, a number of simple deductions or replacements can also be made, all shall be regarded as belonging to protection of the invention
Range.
Claims (10)
1. a kind of source of houses recommended method, which is characterized in that the source of houses recommended method includes:
When receiving recommendation enabled instruction, client to be recommended is obtained to the history access record of each house house type, the history
Access record includes at least one of house type access times, number of clicks, access duration;
According to the history access record, the client to be recommended is calculated respectively to the preference value of each house house type, according to
Each preference value selects a house house type, the benchmark house type as the client to be recommended from each house house type;
The similarity of each house type to be matched and the benchmark house type in house type library to be matched is calculated, selects similarity to meet default
Recommended candidate house type recommendation is sent to the client to be recommended as recommended candidate house type by the house type to be matched of relationship.
2. source of houses recommended method as described in claim 1, which is characterized in that it is described according to the history access record, it calculates
The client to be recommended includes: to the preference value of each house house type respectively
The client to be recommended is added summation to the access times of the house house type, number of clicks, access duration, obtains institute
Client to be recommended is stated to the preference value of the house house type.
3. source of houses recommended method as claimed in claim 2, which is characterized in that it is described according to each preference value from each room
A house house type is selected in room house type, the benchmark house type as the client to be recommended includes:
In each preference value, the maximum house house type of preference value, the benchmark family as the client to be recommended are selected
Type;When there are when two or more, randomly choose a base as the client to be recommended for the maximum house house type of preference value
Quasi- house type.
4. source of houses recommended method as described in claim 1, which is characterized in that described to recommend to send by the recommended candidate house type
Before the client to be recommended, further includes:
It detects in the recommended candidate house type, if there are the house house types that the client to be recommended had browsed, and if it exists,
To exclude other house types outside the house house type browsed in the recommended candidate house type, recommendation be sent to it is described to
Recommend client;If it does not exist, then recommended candidate house type recommendation is directly sent to the client to be recommended.
5. source of houses recommended method as described in claim 1, which is characterized in that it is described receive recommend enabled instruction include:
When the history access record for detecting the client to be recommended updates, determination receives recommendation enabled instruction;
And/or when prefixed time interval reaches, determination receives recommendation enabled instruction.
6. source of houses recommended method as described in any one in claim 1-5, which is characterized in that described to calculate in house type library to be matched
The similarity of each house type to be matched and the benchmark house type include:
Using cosine similarity formula calculate between each house type to be matched and the benchmark house type in the house type to be matched
Cosine similarity in following at least one house type feature: place city number, the longitude and latitude of the affiliated building of house type, house type face
Product, house type room number, house type unit price, house type direction number;The cosine similarity formula is as follows:
Wherein, the cos θijFor the cosine similarity between house type i and house type j feature vector, the n is of house type feature
Number, the x are characterized value;
Each band matching house type in the house type to be matched is calculated and between the benchmark house type such as using Euclidean distance formula
Distance conformability degree in lower at least one house type feature: place city number, the longitude and latitude of the affiliated building of house type, house type area,
House type room number, house type unit price, house type direction number;The Euclidean distance formula is as follows:
Wherein, the dijFor the distance between house type i and house type j feature vector similarity.
7. source of houses recommended method as claimed in claim 6, which is characterized in that it is described calculate in house type library to be matched respectively to
Before the similarity for matching house type and the benchmark house type, further includes:
Detect whether the client to be recommended has investment intent, if so, for the benchmark house type and the family to be matched
Its feature vector is weighted processing according to first set weighted value by each house type to be matched in type library;If it is not, then for described
Each house type to be matched in benchmark house type and the house type library to be matched, its feature vector is added according to second set of weighted value
Power processing;The first set weighted value is different from second set of weighted value.
8. source of houses recommended method as claimed in claim 6, which is characterized in that it is described calculate in house type library to be matched respectively to
Before the similarity for matching house type and the benchmark house type, further includes:
It obtains the client to be recommended and often uses position longitude and latitude, the longitude and latitude of building belonging to the reality of the benchmark house type is replaced
For the common position longitude and latitude.
9. source of houses recommended method as claimed in claim 6, which is characterized in that it is described selection similarity meet preset relation to
House type is matched, includes: as recommended candidate house type
Each cosine similarity is ranked up according to descending sequence, M (M is more than or equal to 2) is a remaining before selecting
House type to be matched corresponding to string similarity is as the first house type set to be matched;By each Distance conformability degree according to there is small arrive
Big sequence is ranked up, and house type to be matched corresponding to M Distance conformability degree is as the second house type set to be matched before selecting;
Determine the intersection between the described first house type set to be matched and the second house type set to be matched, the intersection is corresponding
House house type is as the recommended candidate house type.
10. a kind of server, which is characterized in that the server includes processor, memory and communication bus;
The communication bus is for realizing the connection communication between processor and memory;
The processor is for executing one or more program stored in memory, to realize as appointed in claim 1 to 9
The step of source of houses recommended method described in one.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811564897.5A CN109658192A (en) | 2018-12-20 | 2018-12-20 | A kind of source of houses recommended method and server |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811564897.5A CN109658192A (en) | 2018-12-20 | 2018-12-20 | A kind of source of houses recommended method and server |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109658192A true CN109658192A (en) | 2019-04-19 |
Family
ID=66115976
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811564897.5A Pending CN109658192A (en) | 2018-12-20 | 2018-12-20 | A kind of source of houses recommended method and server |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109658192A (en) |
Cited By (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110210564A (en) * | 2019-05-30 | 2019-09-06 | 贝壳技术有限公司 | Similar house type detection method and device |
CN110222621A (en) * | 2019-05-30 | 2019-09-10 | 贝壳技术有限公司 | Similar house type detection method, electronic equipment and storage medium |
CN110245160A (en) * | 2019-06-03 | 2019-09-17 | 贝壳技术有限公司 | A kind of method and system of determining house decoration scheme |
CN110764665A (en) * | 2019-09-19 | 2020-02-07 | 深圳思为科技有限公司 | Information processing method, information processing device and terminal equipment |
CN110930217A (en) * | 2019-11-05 | 2020-03-27 | 贝壳技术有限公司 | Decoration recommendation method and device, readable storage medium and processor |
CN111028044A (en) * | 2019-10-22 | 2020-04-17 | 贝壳技术有限公司 | Method and device for sharing lease, electronic equipment and storage medium |
CN111047394A (en) * | 2019-11-20 | 2020-04-21 | 泰康保险集团股份有限公司 | Commodity recommendation method and device, electronic equipment and computer readable medium |
CN111080407A (en) * | 2019-12-04 | 2020-04-28 | 青梧桐有限责任公司 | House information recommendation method and device, electronic equipment and readable storage medium |
CN111143689A (en) * | 2019-12-31 | 2020-05-12 | 青梧桐有限责任公司 | Method for constructing recommendation engine according to user requirements and user portrait |
CN111159561A (en) * | 2019-12-31 | 2020-05-15 | 青梧桐有限责任公司 | Method for constructing recommendation engine according to user behaviors and user portrait |
CN111222960A (en) * | 2020-01-17 | 2020-06-02 | 青梧桐有限责任公司 | Room source recommendation method and system based on public traffic zone |
CN111292170A (en) * | 2020-02-18 | 2020-06-16 | 重庆锐云科技有限公司 | Method, device and storage medium for recommending intention customers for appointed building |
CN111639988A (en) * | 2020-05-22 | 2020-09-08 | 贝壳技术有限公司 | Broker recommendation method, device, electronic equipment and storage medium |
CN112330387A (en) * | 2020-09-29 | 2021-02-05 | 重庆锐云科技有限公司 | Virtual broker applied to house-watching software |
CN112365031A (en) * | 2020-10-22 | 2021-02-12 | 汪永强 | House renting system and method based on block chain |
CN112381453A (en) * | 2020-12-04 | 2021-02-19 | 五八同城信息技术有限公司 | Information processing method, information processing apparatus, electronic device, and medium |
CN113034224A (en) * | 2021-03-16 | 2021-06-25 | 重庆锐云科技有限公司 | Similarity-based house source recommendation method, system, equipment and storage medium |
CN113159997A (en) * | 2021-03-03 | 2021-07-23 | 杭州萝卜智能技术有限公司 | Online house-selection transaction system and method |
CN113205383A (en) * | 2021-05-07 | 2021-08-03 | 中山市云经纪网络科技有限公司 | Intelligent house property recommendation method and system |
CN113316778A (en) * | 2019-06-24 | 2021-08-27 | 深圳市欢太科技有限公司 | Equipment recommendation method and related product |
CN113763029A (en) * | 2021-07-20 | 2021-12-07 | 北京房江湖科技有限公司 | Client information recommendation method and device |
CN114298796A (en) * | 2021-12-30 | 2022-04-08 | 江苏冲浪软件科技有限公司 | Housing estate service management method based on block chain |
CN114494742A (en) * | 2021-12-20 | 2022-05-13 | 上海品览数据科技有限公司 | Building professional house type comparison method based on deep learning |
CN115239400A (en) * | 2022-09-21 | 2022-10-25 | 广州越创智数信息科技有限公司 | House purchasing user intention degree calculation method and system |
CN117495515B (en) * | 2023-12-29 | 2024-04-05 | 优材优建(青岛)供应链科技有限公司 | Bid intelligent matching method, system, computer equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106384246A (en) * | 2016-09-09 | 2017-02-08 | 珠海幸福家网络科技有限公司 | House purchase intention determination device and method |
CN108256067A (en) * | 2018-01-16 | 2018-07-06 | 平安好房(上海)电子商务有限公司 | Calculate method, apparatus, equipment and the storage medium of source of houses similarity |
CN108471455A (en) * | 2018-06-14 | 2018-08-31 | 万翼科技有限公司 | Method for pushing, server and the storage medium of house property information |
CN108510321A (en) * | 2018-03-23 | 2018-09-07 | 北京焦点新干线信息技术有限公司 | A kind of construction method and device of house property user portrait |
-
2018
- 2018-12-20 CN CN201811564897.5A patent/CN109658192A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106384246A (en) * | 2016-09-09 | 2017-02-08 | 珠海幸福家网络科技有限公司 | House purchase intention determination device and method |
CN108256067A (en) * | 2018-01-16 | 2018-07-06 | 平安好房(上海)电子商务有限公司 | Calculate method, apparatus, equipment and the storage medium of source of houses similarity |
CN108510321A (en) * | 2018-03-23 | 2018-09-07 | 北京焦点新干线信息技术有限公司 | A kind of construction method and device of house property user portrait |
CN108471455A (en) * | 2018-06-14 | 2018-08-31 | 万翼科技有限公司 | Method for pushing, server and the storage medium of house property information |
Cited By (31)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110210564B (en) * | 2019-05-30 | 2020-07-24 | 贝壳找房(北京)科技有限公司 | Similar house type detection method and device |
CN110222621A (en) * | 2019-05-30 | 2019-09-10 | 贝壳技术有限公司 | Similar house type detection method, electronic equipment and storage medium |
CN110210564A (en) * | 2019-05-30 | 2019-09-06 | 贝壳技术有限公司 | Similar house type detection method and device |
CN110222621B (en) * | 2019-05-30 | 2021-06-11 | 贝壳找房(北京)科技有限公司 | Similar house type detection method, electronic equipment and storage medium |
CN110245160A (en) * | 2019-06-03 | 2019-09-17 | 贝壳技术有限公司 | A kind of method and system of determining house decoration scheme |
CN113316778A (en) * | 2019-06-24 | 2021-08-27 | 深圳市欢太科技有限公司 | Equipment recommendation method and related product |
CN110764665A (en) * | 2019-09-19 | 2020-02-07 | 深圳思为科技有限公司 | Information processing method, information processing device and terminal equipment |
CN111028044A (en) * | 2019-10-22 | 2020-04-17 | 贝壳技术有限公司 | Method and device for sharing lease, electronic equipment and storage medium |
CN111028044B (en) * | 2019-10-22 | 2023-10-27 | 贝壳技术有限公司 | Renting method and device, electronic equipment and storage medium |
CN110930217A (en) * | 2019-11-05 | 2020-03-27 | 贝壳技术有限公司 | Decoration recommendation method and device, readable storage medium and processor |
CN111047394A (en) * | 2019-11-20 | 2020-04-21 | 泰康保险集团股份有限公司 | Commodity recommendation method and device, electronic equipment and computer readable medium |
CN111080407A (en) * | 2019-12-04 | 2020-04-28 | 青梧桐有限责任公司 | House information recommendation method and device, electronic equipment and readable storage medium |
CN111143689A (en) * | 2019-12-31 | 2020-05-12 | 青梧桐有限责任公司 | Method for constructing recommendation engine according to user requirements and user portrait |
CN111159561A (en) * | 2019-12-31 | 2020-05-15 | 青梧桐有限责任公司 | Method for constructing recommendation engine according to user behaviors and user portrait |
CN111222960A (en) * | 2020-01-17 | 2020-06-02 | 青梧桐有限责任公司 | Room source recommendation method and system based on public traffic zone |
CN111292170A (en) * | 2020-02-18 | 2020-06-16 | 重庆锐云科技有限公司 | Method, device and storage medium for recommending intention customers for appointed building |
CN111639988A (en) * | 2020-05-22 | 2020-09-08 | 贝壳技术有限公司 | Broker recommendation method, device, electronic equipment and storage medium |
CN111639988B (en) * | 2020-05-22 | 2024-01-12 | 贝壳技术有限公司 | Broker recommendation method, device, electronic equipment and storage medium |
CN112330387A (en) * | 2020-09-29 | 2021-02-05 | 重庆锐云科技有限公司 | Virtual broker applied to house-watching software |
CN112330387B (en) * | 2020-09-29 | 2023-07-18 | 重庆锐云科技有限公司 | Virtual broker applied to house watching software |
CN112365031A (en) * | 2020-10-22 | 2021-02-12 | 汪永强 | House renting system and method based on block chain |
CN112365031B (en) * | 2020-10-22 | 2021-11-16 | 江苏三六五网络股份有限公司 | House renting system and method based on block chain |
CN112381453A (en) * | 2020-12-04 | 2021-02-19 | 五八同城信息技术有限公司 | Information processing method, information processing apparatus, electronic device, and medium |
CN113159997A (en) * | 2021-03-03 | 2021-07-23 | 杭州萝卜智能技术有限公司 | Online house-selection transaction system and method |
CN113034224A (en) * | 2021-03-16 | 2021-06-25 | 重庆锐云科技有限公司 | Similarity-based house source recommendation method, system, equipment and storage medium |
CN113205383A (en) * | 2021-05-07 | 2021-08-03 | 中山市云经纪网络科技有限公司 | Intelligent house property recommendation method and system |
CN113763029A (en) * | 2021-07-20 | 2021-12-07 | 北京房江湖科技有限公司 | Client information recommendation method and device |
CN114494742A (en) * | 2021-12-20 | 2022-05-13 | 上海品览数据科技有限公司 | Building professional house type comparison method based on deep learning |
CN114298796A (en) * | 2021-12-30 | 2022-04-08 | 江苏冲浪软件科技有限公司 | Housing estate service management method based on block chain |
CN115239400A (en) * | 2022-09-21 | 2022-10-25 | 广州越创智数信息科技有限公司 | House purchasing user intention degree calculation method and system |
CN117495515B (en) * | 2023-12-29 | 2024-04-05 | 优材优建(青岛)供应链科技有限公司 | Bid intelligent matching method, system, computer equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109658192A (en) | A kind of source of houses recommended method and server | |
CN109670113A (en) | A kind of source of houses recommended method, device and server | |
JP7084970B2 (en) | Systems and methods for marketing mobile ad supply | |
Zhang et al. | Geosoca: Exploiting geographical, social and categorical correlations for point-of-interest recommendations | |
KR102189065B1 (en) | Interest profile of a user of a mobile application | |
JP4750198B2 (en) | A system to change the price paid by the advertiser | |
US20160027055A1 (en) | Combining attributes from multiple sources | |
US20130198281A1 (en) | System for Creating Anonymous Social Gatherings | |
WO2020186819A1 (en) | Product data processing | |
JP6937776B2 (en) | Methods and devices for identifying geographic location attribute information | |
CN102591911A (en) | Real-time personalized recommendation of location-related entities | |
CN107436893A (en) | A kind of webpage recommending method and device | |
EP1395926A2 (en) | Domain place registration system and method for registering for geographic based services | |
CN108369586A (en) | System and method for detecting interaction and influence in network | |
CN106471499A (en) | Determine the aspect of physical location and/or the text summaries with regard to physical location are generated based on the calculate interaction relevant with physical location | |
Cao et al. | Revisitation in urban space vs. online: A comparison across pois, websites, and smartphone apps | |
US11341109B2 (en) | Method and system for detecting and using locations of electronic devices of users in a specific space to analyze social relationships between the users | |
CN108280748A (en) | Assemble the method for pushing and device of store information | |
CN109903086A (en) | A kind of similar crowd's extended method, device and electronic equipment | |
JP2020537252A (en) | Systems and methods for predicting similar mobile devices | |
CN107220308A (en) | POI rational detection method and device, equipment and computer-readable recording medium | |
CN109658165A (en) | Advertisement bid method, apparatus, electronic equipment and computer-readable medium | |
CN106030627A (en) | Location based ranking of real world locations | |
KR20120135790A (en) | Method for managing meta information server | |
CN110019180A (en) | Multi-source data account relating and device |
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 |
Application publication date: 20190419 |
|
RJ01 | Rejection of invention patent application after publication |