CN109740072A - Hotel's sort method and system under OTA platform based on POI - Google Patents
Hotel's sort method and system under OTA platform based on POI Download PDFInfo
- Publication number
- CN109740072A CN109740072A CN201811638493.6A CN201811638493A CN109740072A CN 109740072 A CN109740072 A CN 109740072A CN 201811638493 A CN201811638493 A CN 201811638493A CN 109740072 A CN109740072 A CN 109740072A
- Authority
- CN
- China
- Prior art keywords
- hotel
- user
- recommended
- sequence
- poi
- 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.)
- Granted
Links
Abstract
The invention discloses the hotel's sort method and system under a kind of OTA platform based on POI, method includes: the region obtained where POI；Obtain all selectable hotels；Each selectable hotel is calculated at a distance from POI, the selectable hotel that setting distance is less than distance threshold is the first hotel；Obtain user's dimensional information of user to be recommended；User to be recommended is calculated to the preference in each hotel under OTA platform；Obtain hotel's dimensional information in every one first hotel；The probability that user to be recommended places an order to every one first hotel is calculated with prediction model；Recommended user is treated to be ranked up probability single under every one first hotel；Result after output sequence.The present invention is by machine learning algorithm, the information based on user and hotel, predicts the probability that user places an order to hotel in real time, to realize the intelligent sorting in online hotel, personalized ordering can not only be realized for different user, can also promote the sales volume in hotel, improve the brand image of OTA.
Description
Technical field
The present invention relates to OTA (online travel agency) field, in particular to POI (point of interest) is based under a kind of OTA platform
Hotel's sort method and system.
Background technique
At present in OTA industry, for hotel's sort method after the POI information of user's selection, it is mainly based upon hotel
The method to sort from small to large to the distance of POI, this method filter out for user from user's selection using distance as judgment criteria
The hotel that POI is closer.Also there is the sort method based on hotel's sales volume, this method considers the feelings of the popularity in hotel
The high hotel of sales volume is come front by condition, and the low hotel of sales volume comes below to facilitate user to make a choice, in practice, according to
There are also other kinds of sortords for different OTA platforms, such as arrange according to the sequence of hotel's price, according to user's collection
Sequence etc..
Summary of the invention
The technical problem to be solved by the present invention is in order to overcome in the prior art the hotel OTA sortord it is single, do not meet
The defect of user's actual need provides hotel's sort method and system based on POI under a kind of OTA platform.
The present invention is to solve above-mentioned technical problem by following technical proposals:
The present invention provides hotel's sort methods based on POI under a kind of OTA platform, include the following steps:
The region where the POI is obtained according to the POI that user to be recommended inputs；
Obtain all selectable hotels in the region；
Each selectable hotel is calculated at a distance from the POI, the institute that the distance is less than distance threshold is set
Stating selectable hotel is the first hotel；
Obtain user's dimensional information of the user to be recommended；
The user to be recommended is calculated to the preference in each hotel under the OTA platform；
Obtain hotel's dimensional information in each first hotel；
By user's dimensional information of the user to be recommended, hotel's dimensional information in each first hotel and it is described to
Recommended user is input to a prediction model to the preference in each hotel, using the prediction model be calculated it is described to
The probability that recommended user places an order to each first hotel；
The probability to place an order to the user to be recommended to each first hotel is ranked up；
Result after output sequence.
Preferably, user's dimensional information includes the essential information of the user, user's dimensional information further includes
The information in the hotel that the information of the order in the hotel of the user ordered in the past, the user browsed, the user collect
At least one of the information in hotel；
Hotel's dimensional information includes the essential information in the hotel, and hotel's dimensional information further includes the hotel
Pricing information, the hotel favorable comment degree information and the hotel by least one of collection number information.
Preferably, the step of preference for calculating the user to be recommended to each hotel under the OTA platform
Include:
The institute in the hidden variable and the OTA platform of all users in the OTA platform is obtained using hidden semantic model
There is the hidden variable in hotel；
The hidden variable of the user to be recommended is obtained according to the hidden variable of all users；
The user to be recommended is calculated to the preference in each hotel using following formula:
Wherein, lfmijIndicate the user i to be recommended to the preference of each hotel j, uikIndicate it is described to
The kth dimension values of the vector of the hidden variable of recommended user i, hjkIndicate the kth dimension of the vector of the hidden variable of each hotel j
Value, m indicate the length of the vector of the hidden variable in the user and the hotel.
Preferably, the hidden variable for obtaining all users in the OTA platform using hidden semantic model and the OTA
The step of hidden variable in all hotels in platform are as follows:
Obtain all users；
Obtain the score data in all hotels；
The institute is calculated using the score data of all users and all hotels in the hidden semantic model
There are the hidden variable of user and the hidden variable in all hotels.
Preferably, the generation step of the prediction model are as follows:
Obtain History Order data, every History Order data include corresponding historical user and history hotel and
The result whether to place an order；
History Order data described in every are handled, to obtain the corresponding user's dimension letter of the historical user
Breath, the corresponding hotel's dimensional information in the history hotel and the user to be recommended form the preference in each hotel
History feature vector；
History Order described in every is marked according to the result whether to place an order special to obtain each history
Levy the corresponding label of vector；
It is carried out using the history feature vector and the corresponding label using a kind of XGBOOST (Integrated Algorithm) algorithm
The building of disaggregated model, to obtain the prediction model.
Wherein it is possible to be adjusted by AUC (the lower area of song) index to the prediction model, the specific step of the adjustment
Suddenly are as follows: judge whether the prediction model reaches the AUC index, if it is not, after then adjusting the parameter in the prediction model again
According to model adjusted to obtain the prediction model for meeting the AUC index.
Preferably, the probability to place an order to the user to be recommended to each first hotel is ranked up are as follows:
The probability to place an order to the user to be recommended to each first hotel according to being ranked up from high to low, with
To First ray；
Hotel's sort method is further comprising the steps of:
Hotel in the First ray is ranked up according to scoring, to obtain the second sequence；
Result after the output sequence is to export second sequence.
Preferably, hotel's sort method, further includes:
The selectable hotel that the distance is arranged more than or equal to the distance threshold is the second hotel, to described
Second hotel is ranked up to obtain third sequence；
By the third sequence permutation to second sequence later to obtain the 4th sequence；
Result after the output sequence is to export the 4th sequence.
Preferably, the 5th sequence is set by all selectable hotels that are no picture and/or can not determining, by institute
The 5th sequence permutation is stated to the 4th sequence later to obtain the 6th sequence.Result after the output sequence is described in output
6th sequence.
The present invention also provides hotel's ordering systems based on POI under a kind of OTA platform, comprising:
Region obtains module, and the POI for being inputted according to user to be recommended obtains the region where the POI；
Hotel may be selected and obtain module, for obtaining all selectable hotels in the region；
First hotel's setup module, for calculating each selectable hotel at a distance from the POI, described in setting
The selectable hotel that distance is less than distance threshold is the first hotel；
User's dimension obtains module, for obtaining user's dimensional information of the user to be recommended；
Preference computing module, for calculating the user to be recommended to the preference in each hotel under the OTA platform
Degree；
Hotel's dimension obtains module, for obtaining hotel's dimensional information in each first hotel；
Prediction module, for by the wine of user's dimensional information of the user to be recommended, each first hotel
Shop dimensional information and the user to be recommended are input to a prediction model to the preference in each hotel, use the prediction model
The probability that the user to be recommended places an order to each first hotel is calculated；
Sorting module, the probability for placing an order to the user to be recommended to each first hotel are ranked up；
Output module, for exporting the result after sorting.
Preferably, user's dimensional information includes the essential information of the user, user's dimensional information further includes
The information in the hotel that the information of the order in the hotel of the user ordered in the past, the user browsed, the user collect
At least one of the information in hotel；
Hotel's dimensional information includes the essential information in the hotel, and hotel's dimensional information further includes the hotel
Pricing information, the hotel favorable comment degree information and the hotel by least one of collection number information.
Preferably, the preference computing module includes:
First hidden variable acquiring unit, for obtaining the hidden of all users in the OTA platform using hidden semantic model
The hidden variable in variable and all hotels in the OTA platform；
Second hidden variable acquiring unit, for showing that the user's to be recommended is hidden according to the hidden variable of all users
Variable；
Preference computing unit, for the user to be recommended to be calculated to each hotel by following formula
Preference:
Wherein, lfmijIndicate the user i to be recommended to the preference of each hotel j, uikIndicate it is described to
The kth dimension values of the vector of the hidden variable of recommended user i, hjkIndicate the kth dimension of the vector of the hidden variable of each hotel j
Value, m indicate the length of the vector of the hidden variable in the user and the hotel.
Preferably, the first hidden variable acquiring unit includes:
First obtains subelement, for obtaining all users；
Second obtains subelement, for obtaining the score data in all hotels；
Hidden variable computation subunit, for using all users and all hotels using the hidden semantic model
Score data the hidden variable of all users and the hidden variable in all hotels is calculated.
Preferably, hotel's ordering system further includes model generation unit, the model generation unit includes:
Order acquiring unit, for obtaining History Order data, every History Order data include corresponding history
User and history hotel and the result whether to place an order；
Data processing unit, for handling History Order data described in every, to obtain the historical user couple
User's dimensional information, the corresponding hotel's dimensional information in the history hotel and the user to be recommended answered is to each
The history feature vector of the preference composition in hotel；
Marking unit, it is every to obtain for History Order described in every to be marked according to the result whether to place an order
The corresponding label of the one history feature vector；
Model construction unit, for using XGBOOST algorithm using the history feature vector and the corresponding label
The building of disaggregated model is carried out, to obtain the prediction model.
Wherein it is possible to be adjusted by AUC index to the prediction model, the specific steps of the adjustment are as follows: judgement
Whether the prediction model reaches the AUC index, if it is not, after then adjusting the parameter in the prediction model further according to adjustment after
Model to obtain the prediction model for meeting the AUC index.
Preferably, the sorting module includes: the first sequencing unit and the second sequencing unit；
The probability that first sequencing unit is used to place an order to each first hotel to the user to be recommended according to
It is ranked up from high to low, to obtain First ray；
Second sequencing unit is for being ranked up the hotel in the First ray according to scoring, to obtain second
Sequence；
Result in the output module after output sequence is to export second sequence.
Preferably, the sorting module further include: third sequencing unit and the 4th sequencing unit；
The third sequencing unit, it is described selectable more than or equal to the distance threshold for the distance to be arranged
Hotel is the second hotel, is ranked up second hotel to obtain third sequence；
4th sequencing unit is used for the third sequence permutation to second sequence later to obtain the 4th sequence
Column；
Result in the output module after output sequence is to export the 4th sequence.
Preferably, the 5th sequence is set by all selectable hotels that are no picture and/or can not determining, by institute
The 5th sequence permutation is stated to the 4th sequence later to obtain the 6th sequence.Result after the output sequence is described in output
6th sequence.
The positive effect of the present invention is that:
Hotel's sort method and system under OTA platform provided by the invention based on POI, can be according to the history number of user
According to, the historical data in hotel, user to the preference in hotel, placed an order by machine learning model real-time judge user to hotel
Probability, to realize the intelligent sorting in online hotel, can not only quicksort go out to meet user consumption habit itself hotel,
The retrieval dynamics of user is reduced, while also promoting the sales volume in hotel, improves the brand image of OTA.
Detailed description of the invention
Fig. 1 be embodiment 1 OTA platform under hotel's sort method based on POI flow chart.
Fig. 2 is the flow chart of step S15 in Fig. 1.
The step of Fig. 3 obtains for the hidden semantic model of hotel's sort method based on POI under the OTA platform of embodiment 1
Flow chart.
The stream for the step of Fig. 4 generates for the prediction model of hotel's sort method based on POI under the OTA platform of embodiment 1
Cheng Tu.
Fig. 5 is the flow chart of step S18 in Fig. 1.
Fig. 6 be embodiment 2 OTA platform under hotel's ordering system based on POI module diagram.
Fig. 7 is the module diagram of preference computing module in Fig. 6.
Fig. 8 is the module diagram of the first hidden variable acquiring unit in Fig. 7.
Fig. 9 illustrates for the module of the model generation unit of hotel's ordering system based on POI under the OTA platform of embodiment 2
Figure.
Figure 10 is the module diagram of sorting module in Fig. 6.
Specific embodiment
The present invention is further illustrated below by the mode of embodiment, but does not therefore limit the present invention to the reality
It applies among a range.
Embodiment 1
Hotel's sort method based on POI under a kind of OTA platform is present embodiments provided, detailed process is as shown in Figure 1, packet
It includes:
Step S11, the region where the POI is obtained according to the POI that user to be recommended inputs；
In the present embodiment, the region can be some city, can be some region, can be some street, specifically
The range of the region can voluntarily be selected according to the demand of the user to be recommended.
Step S12, all selectable hotels in the region are obtained；
The selectable hotel is that currently can be specifically included traditional hotel with scheduled hotel, also be included
People place, inn, Youth Hotel etc..
Step S13, each selectable hotel is calculated at a distance from the POI, and the distance is set less than distance
The selectable hotel of threshold value is the first hotel；
Wherein, distance threshold is that such as can be 1 kilometer according to needing preset value.
Step S14, user's dimensional information of the user to be recommended is obtained；
Wherein, user's dimensional information includes the essential information of the user, and user's dimensional information further includes institute
State the information of the order in the hotel of user ordered in the past, the information in the hotel that the user browsed, the user collection wine
At least one of the information in shop；The essential information of the user may include the letter such as age, gender, educational background, local of user
Breath.
Wherein, if user does not have essential information, the essential information of the user is set for sky；If user does not order
Hotel is crossed, then the information of the order in the hotel of the user ordered in the past is set for sky；If user did not browse hotel,
The information in the hotel that the user browsed is set for sky；If user did not collect hotel, user's collection is set
Hotel information be sky.
Step S15, the user to be recommended is calculated to the preference in each hotel under the OTA platform；
Step S16, hotel's dimensional information in each first hotel is obtained；
Wherein, hotel's dimensional information includes the essential information in the hotel, and hotel's dimensional information further includes institute
The pricing information in hotel, the favorable comment degree information in the hotel and the hotel are stated by least one of collection number information；Institute
The essential information for stating hotel may include the information such as the star in hotel, house type.
Wherein, if hotel does not have essential information, the essential information in the hotel is set for sky；If hotel does not have valence
The pricing information in the hotel is arranged then as sky in lattice information；If hotel does not have favorable comment degree information, the good of the hotel is set
Degree of commenting information is sky；If hotel is not by collection number information, it is sky that the hotel, which is arranged, by collection number information.
Step S17, the user to be recommended is calculated using the prediction model to place an order to each first hotel
Probability；
Wherein, the specific implementation of step S17 is by user's dimensional information of the user to be recommended, each described the
Hotel's dimensional information in one hotel and the user to be recommended are input to a prediction model to the preference in each hotel, then
The probability that the user to be recommended places an order to each first hotel is calculated using the prediction model.
Step S18, the probability to place an order to the user to be recommended to each first hotel is ranked up；
Step S19, the result after output sequence.
In the present embodiment, by the customized POI of user to be recommended, reasonably sorted to the hotel near POI, needle
To the different consumer behaviors habit of different users to be recommended, in conjunction with the quality in hotel, according to user to lower single probability in hotel
It is ranked up, the hotel for being best suitable for user demand is come into foremost, exports ranking results, so that reducing user screens hotel
Process promotes the sales volume in hotel while reduction user's operation is arduously spent.
As shown in Fig. 2, in the present embodiment, step S15 the following steps are included:
Step S151, hidden variable and the OTA that all users in the OTA platform are obtained using hidden semantic model are flat
The hidden variable in all hotels in platform；
Step S152, the hidden variable of the user to be recommended is obtained according to the hidden variable of all users；
Step S153, the user to be recommended is calculated to the preference in each hotel, institute using the first formula
State the first formula are as follows:
Wherein, lfmijIndicate the user i to be recommended to the preference of each hotel j, uikIndicate it is described to
The kth dimension values of the vector of the hidden variable of recommended user i, hjkIndicate the kth dimension of the vector of the hidden variable of each hotel j
Value, m indicate the length of the vector of the hidden variable in the user and the hotel.
In the present embodiment, the flow chart for the step of hidden semantic model in the step S151 obtains is as shown in figure 3, packet
It includes:
Step S1511, all users are obtained；
Step S1512, the score data in all hotels is obtained；
Step S1513, the described hidden semantic model is calculated using the score data in all users and all hotels
Obtain the hidden variable of all users and the hidden variable in all hotels.
Wherein, the data that the hidden semantic model uses both had included the data of scored user, and had also included not scoring
The data of the user crossed.It should be appreciated that if a certain user does not score to a certain hotel, scoring of the user to the hotel
Data are sky.
In the present embodiment, the hidden variable of all users and the hidden variable in all hotels, root are calculated by hidden semantic model
According to all users hidden variable and all hotels hidden variable to obtain user to be recommended hidden variable and each hotel it is hidden
Variable, to predict user to be recommended to the preference in each hotel, wherein it is primary to can be set daily system update, i.e., often
It calculates the primary hidden variable in all hotels and the hidden variable of all users by hidden semantic model when specific, is needing in advance
It, only need to be from the hidden variable of all users calculated and all hotels when surveying preference of the user to be recommended to each hotel
The corresponding data of user to be recommended are obtained in hidden variable, to not only meet the demand of different user, but also to calculate
Journey is rapider.
As shown in figure 4, the generation step of the prediction model in the present embodiment in step 17 includes:
Step S171, History Order data are obtained, every History Order data include corresponding historical user and Li
History hotel and the result whether to place an order；
Step S172, History Order data described in every are handled to obtain history feature vector；
Wherein, the specific implementation of step S172 is to handle History Order data described in every, to obtain
State the corresponding user's dimensional information of historical user, the corresponding hotel's dimensional information in the history hotel and described wait push away
User is recommended to the history feature vector of the preference composition in each hotel.
Step S173, History Order described in every is marked according to the result whether to place an order to obtain each institute
State the corresponding label of history feature vector；
In the present embodiment, 1 is labeled as to the hotel to place an order, the hotel only browsed without placing an order is labeled as 0.
Step S174, classified using the history feature vector and the corresponding label using XGBOOST algorithm
The building of model, to obtain the prediction model.
In the present embodiment, during obtaining the prediction model, certain evaluation index can be used to described pre-
The effect for surveying model is verified, and the present embodiment uses AUC index to be verified, and the AUC index is higher, the prediction
Model is better, and other indexs also can be used in practice and verify to the prediction model, such as accuracy rate, recall rate,
Which kind of specifically used index needs difference according to the actual situation to be selected.
Wherein, if the prediction model does not reach the evaluation index, according to the evaluation index to described pre-
Survey model parameter be adjusted and re -training described in prediction model, until the prediction model meet it is described evaluation refer to
Mark.
In the present embodiment, using the integrated learning approach in machine learning, it is able to ascend the generalization ability of model.To user
After selecting POI information, carry out according to the distance of hotel distance POI apart from screening and filtering.Excavate the shadow that user places an order to hotel
The factor of sound (user's dimensional information, hotel's dimensional information, user to be recommended are to the preference in each first hotel) carries out machine
The building of device learning model.The prediction for carrying out single probability under user to hotel using machine learning model, according to placing an order for prediction
Probability from high in the end, is ranked up hotel, to carry out auto-sequencing automatically to hotel to for different user, meets not
With the demand of user.
In the present embodiment, prediction model is trained by the way that certain evaluation index is arranged, to play the optimization prediction
The effect of model, thus it is more close needed for the ranking come and user to be recommended that the prediction model can be made to predict.
As shown in figure 5, step S18 is specifically included in the present embodiment:
Step S181, the probability to place an order to the user to be recommended to each first hotel according to carrying out from high to low
Sequence, to obtain First ray；
Step S182, the hotel in the First ray is ranked up according to scoring, to obtain the second sequence；
In the present embodiment, it is according to the specific implementation method that scoring is ranked up to the hotel in the First ray, if
A scoring threshold value is set, the hotel for being unsatisfactory for the point scoring threshold value in the First ray is placed on and meets the point scoring threshold
Behind the hotel of value, to obtain the second sequence.
Step S183, it is the second wine that the distance, which is arranged, to be greater than or equal to the selectable hotel of the distance threshold
Shop is ranked up to obtain third sequence second hotel；
Step S184, by the third sequence permutation to second sequence later to obtain the 4th sequence；
Step S185, the 5th sequence is set by all selectable hotels that are no picture and/or can not determining；
Wherein, all selectable hotels that can not be determined sort to the hotel of the no picture, both without picture
All selectable hotels that can not be determined again come finally.
Step S186, by the 5th sequence permutation to the 4th sequence later to obtain the 6th sequence；
Result after the output sequence is to export the 6th sequence.
It wherein, can also include to the hotel for being unsatisfactory for required distance and being filtered, to described in the 6th sequence
The hotel filtered out sorts from small to large according to distance POI distance, replenishes behind low hotel of scoring, gives the enough choosings of user
Select the chance in hotel.
In the present embodiment, by using hidden semantic model and prediction model, user's need to be recommended can be met with automatic Prediction
The sequence in the hotel asked, and the knot for meeting user demand can be predicted in real time on line according to the different situations of different user
Fruit, so that the unicity for overcoming hotel to predict reduces user and manually select the cumbersome of hotel while rationally sorting to hotel
Property, make hotel's sequencing production faster, user experience more preferably.
The method that the present embodiment utilizes machine learning, according to the POI information that user to be recommended selects, sieve series certain distance model
The hotel enclosed predicts that user to be recommended can correct again the probability that the hotel places an order to hotel's quality, supplement distance range it
Outer hotel more selects space to user.Different hotel's sequencing schemes are targetedly carried out to different user, thus
It is inconvenient when selecting hotel to reduce user to be recommended, improves user experience, promotes hotel's order sales volume, optimizes brand image.
Embodiment 2
As shown in fig. 6, present embodiments providing hotel's ordering system based on POI under a kind of OTA platform, comprising: region
It obtains module 11, hotel's acquisition module 12, first hotel's setup module 13, user's dimension acquisition module 14, preference may be selected
Computing module 15, hotel's dimension obtain module 16, prediction module 17, sorting module 18 and output module 19.
Region obtains module 11 and is used to obtain the region where the POI according to the POI that user to be recommended inputs.
In the present embodiment, the region can be some city, can be some region, can be some street, specifically
The range of the region can voluntarily be selected according to the demand of the user to be recommended.
Hotel may be selected and obtain module 12 for obtaining all selectable hotels in the region.
The selectable hotel both may include traditional hotel, also may include people place, inn, young trip
Society etc..
Institute is arranged for calculating each selectable hotel at a distance from the POI in first hotel's setup module 13
Stating distance to be less than the selectable hotel of distance threshold is the first hotel.
User's dimension obtains module 14, for obtaining user's dimensional information of the user to be recommended.
Wherein, user's dimensional information includes the essential information of the user, and user's dimensional information further includes institute
State the information of the order in the hotel of user ordered in the past, the information in the hotel that the user browsed, the user collection wine
At least one of the information in shop；The essential information of the user includes the information such as age, gender, educational background, the local of user.
Wherein, if user does not have essential information, the essential information of the user is set for sky；If user does not order
Hotel is crossed, then the information of the order in the hotel of the user ordered in the past is set for sky；If user did not browse hotel,
The information in the hotel that the user browsed is set for sky；If user did not collect hotel, user's collection is set
Hotel information be sky.Preference computing module 15, for calculating the user to be recommended to each under the OTA platform
The preference in hotel.
Hotel's dimension obtains module 16, for taking hotel's dimensional information in each first hotel.
Wherein, hotel's dimensional information includes the essential information in the hotel, and hotel's dimensional information further includes institute
The pricing information in hotel, the favorable comment degree information in the hotel and the hotel are stated by least one of collection number information；Institute
The essential information for stating hotel includes the information such as the star in hotel, house type.
Wherein, if hotel does not have essential information, the essential information in the hotel is set for sky；If hotel does not have valence
The pricing information in the hotel is arranged then as sky in lattice information；If hotel does not have favorable comment degree information, the good of the hotel is set
Degree of commenting information is sky；If hotel is not by collection number information, it is sky that the hotel, which is arranged, by collection number information.Prediction
Module 17 be used for by user's dimensional information of the user to be recommended, hotel's dimensional information in each first hotel and
The user to be recommended is input to a prediction model to the preference in each hotel, is calculated using the prediction model described
The probability that user to be recommended places an order to each first hotel.
The probability that sorting module 18 is used to place an order to each first hotel to the user to be recommended is ranked up.
Output module 19 is after sorting as a result, in the present embodiment in exporting, and the result after the sequence is the 6th sequence.
As shown in fig. 7, preference computing module 15 specifically includes in the present embodiment: the first hidden variable acquiring unit 151,
Second hidden variable acquiring unit 152 and preference computing unit 153.
First hidden variable acquiring unit 151 is used to obtain all users' in the OTA platform using hidden semantic model
Hidden variable and the hidden variable in all hotels in the OTA platform.
Second hidden variable acquiring unit 152 is used to obtain the user's to be recommended according to the hidden variable of all users
Hidden variable.
Preference computing unit 153 is used to that the user to be recommended to be calculated to each hotel using the first formula
Preference, first formula are as follows:
Wherein, lfmijIndicate the user i to be recommended to the preference of each hotel j, uikIndicate it is described to
The kth dimension values of the vector of the hidden variable of recommended user i, hjkIndicate the kth dimension of the vector of the hidden variable of each hotel j
Value, m indicate the length of the vector of the hidden variable in the user and the hotel.
As shown in figure 8, the first hidden variable acquiring unit 151 specifically includes in the present embodiment: first obtains subelement
1511, second subelement 1512 and hidden variable computation subunit 1513 are obtained.
First acquisition subelement 1511 is for obtaining all users.
Second acquisition subelement 1512 is used to obtain the score data in all hotels.
Hidden variable computation subunit 1513 is used to use all users and described all using the hidden semantic model
The hidden variable of all users and the hidden variable in all hotels is calculated in the score data in hotel.
Wherein, the data that the hidden semantic model uses both had included the data of scored user, and had also included not scoring
The data of the user crossed.It should be appreciated that if a certain user does not score to a certain hotel, scoring of the user to the hotel
Data are sky.
In the present embodiment, the hidden variable of all users and the hidden variable in all hotels, root are calculated by hidden semantic model
According to all users hidden variable and all hotels hidden variable to obtain user to be recommended hidden variable and each hotel it is hidden
Variable, to predict user to be recommended to the preference in each hotel, wherein it is primary to can be set daily system update, i.e., often
It calculates the primary hidden variable in all hotels and the hidden variable of all users by hidden semantic model when specific, is needing in advance
It, only need to be from the hidden variable of all users calculated and all hotels when surveying preference of the user to be recommended to each hotel
The corresponding data of user to be recommended are obtained in hidden variable, to not only meet the demand of different user, but also to calculate
Journey is rapider.
As shown in figure 9, hotel's ordering system further includes model generation unit 177 in the present embodiment, the model is raw
It is specifically included at unit: order acquiring unit 171, data processing unit 172, marking unit 173 and model construction unit 174.
For order acquiring unit 171 for obtaining History Order data, every History Order data include corresponding go through
History user and history hotel and the result whether to place an order.
Data processing unit 172 is for handling History Order data described in every, to obtain the historical user
Corresponding user's dimensional information, the corresponding hotel's dimensional information in the history hotel and the user to be recommended are to every
The history feature vector of the preference composition in one hotel.
Marking unit 173 is used to that History Order described in every to be marked to obtain according to the result whether to place an order
The corresponding label of each history feature vector.
In the present embodiment, 1 is labeled as to the hotel to place an order, the hotel only browsed without placing an order is labeled as 0.
Model construction unit 174 using the history feature vector and the corresponding label using XGBOOST for being calculated
Method carries out the building of disaggregated model, to obtain the prediction model.
In the present embodiment, during obtaining the prediction model, certain evaluation index can be used to described pre-
The effect for surveying model is verified, and the present embodiment uses AUC index to be verified, and the AUC index is higher, the prediction
Model is better, and other indexs also can be used in practice and verify to the prediction model, such as accuracy rate, recall rate,
Which kind of specifically used index needs difference according to the actual situation to be selected.
Wherein, if the prediction model does not reach the evaluation index, according to the evaluation index to described pre-
Survey model parameter be adjusted and re -training described in prediction model, until the prediction model meet it is described evaluation refer to
Mark.
In the present embodiment, prediction model is trained by the way that certain evaluation index is arranged, to play the optimization prediction
The effect of model, thus can be more close needed for the ranking and user to be recommended that the prediction model predicts.
As shown in Figure 10, in the present embodiment, sorting module 18 is specifically included: the first sequencing unit 181, the second sequencing unit
182, third sequencing unit 183, the 4th sequencing unit 184, the 5th sequencing unit 185, the 6th sequencing unit 186.
The probability that first sequencing unit 181 is used to place an order to each first hotel to the user to be recommended according to from
It is high to Low to be ranked up, to obtain First ray.
Second sequencing unit 182 is for being ranked up the hotel in the First ray according to scoring, to obtain second
Sequence.
In the present embodiment, it is according to the specific implementation method that scoring is ranked up to the hotel in the First ray, if
A scoring threshold value is set, the hotel for being unsatisfactory for the point scoring threshold value in the First ray is placed on and meets the point scoring threshold
Behind the hotel of value, to obtain the second sequence.
Third sequencing unit 183 is used to be arranged the selectable wine that the distance is greater than or equal to the distance threshold
Shop is the second hotel, is ranked up second hotel to obtain third sequence.
4th sequencing unit 184 is used for the third sequence permutation to second sequence later to obtain the 4th sequence
Column.
5th sequencing unit 185 is used for all selectable hotels that are by no picture and/or can not determining and is set as
5th sequence.
Wherein, all selectable hotels that can not be ordered sort to the hotel of the no picture, both without picture
All selectable hotels that can not be determined again come finally.
6th sequencing unit 186 is used for the 5th sequence permutation to the 4th sequence later to obtain the 6th sequence
Column.
Result after the output sequence is to export the 6th sequence.
It,, can be certainly by using hidden semantic model and prediction model by obtaining the POI of user's input in the present embodiment
The sequence in the dynamic hotel for predicting to meet user demand to be recommended, and can be according to the different situations of different user, on line in real time
Predict meet user demand as a result, to overcome hotel predict unicity, rationally to hotel sort while, reduce
User manually selects the triviality in hotel, make hotel's sequencing production faster, user experience more preferably.
Although specific embodiments of the present invention have been described above, it will be appreciated by those of skill in the art that this is only
For example, protection scope of the present invention is to be defined by the appended claims.Those skilled in the art without departing substantially from
Under the premise of the principle and substance of the present invention, many changes and modifications may be made, but these change and
Modification each falls within protection scope of the present invention.
Claims (14)
1. hotel's sort method under a kind of OTA platform based on POI, which is characterized in that include the following steps:
The region where the POI is obtained according to the POI that user to be recommended inputs；
Obtain all selectable hotels in the region；
Each selectable hotel is calculated at a distance from the POI, the distance is arranged can less than the described of distance threshold
The hotel selected is the first hotel；
Obtain user's dimensional information of the user to be recommended；
The user to be recommended is calculated to the preference in each hotel under the OTA platform；
Obtain hotel's dimensional information in each first hotel；
By user's dimensional information of the user to be recommended, hotel's dimensional information in each first hotel and described to be recommended
User is input to a prediction model to the preference in each hotel, is calculated using the prediction model described to be recommended
The probability that user places an order to each first hotel；
The probability to place an order to the user to be recommended to each first hotel is ranked up；
Result after output sequence.
2. hotel's sort method under OTA platform as described in claim 1 based on POI, which is characterized in that
User's dimensional information includes the essential information of the user, user's dimensional information further include the user with
Into the information in the hotel that the information in the hotel that the information of the order in the hotel ordered, the user browsed, the user collect
At least one；
Hotel's dimensional information includes the essential information in the hotel, and hotel's dimensional information further includes the valence in the hotel
Lattice information, the favorable comment degree information in the hotel and the hotel are by least one of collection number information.
3. hotel's sort method under OTA platform as described in claim 1 based on POI, which is characterized in that described in the calculating
User to be recommended includes: to the step of preference in each hotel under the OTA platform
All wine in the hidden variable and the OTA platform of all users in the OTA platform are obtained using hidden semantic model
The hidden variable in shop；
The hidden variable of the user to be recommended is obtained according to the hidden variable of all users；
The user to be recommended is calculated to the preference in each hotel using following formula:
Wherein, lfmijIndicate the user i to be recommended to the preference of each hotel j, uikIndicate described to be recommended
The kth dimension values of the vector of the hidden variable of user i, hjkIndicate the kth dimension values of the vector of the hidden variable of each hotel j, m
Indicate the length of the vector of the hidden variable in the user and each hotel.
4. hotel's sort method under OTA platform as claimed in claim 3 based on POI, which is characterized in that described to use enigmatic language
Adopted model obtains the hidden variable of all users in the OTA platform and the hidden variable in all hotels in the OTA platform
Step are as follows:
Obtain all users；
Obtain the score data in all hotels；
The hidden semantic model using the score data of all users and all hotels be calculated it is described institute it is useful
The hidden variable at family and the hidden variable in all hotels.
5. hotel's sort method under OTA platform as described in claim 1 based on POI, which is characterized in that the prediction model
Generation step are as follows:
Obtain History Order data, every History Order data include corresponding historical user and history hotel and whether
The result to place an order；
History Order data described in every are handled, with obtain the corresponding user's dimensional information of the historical user,
The corresponding hotel's dimensional information in history hotel and the user to be recommended go through to what the preference in each hotel formed
History feature vector；
History Order described in every is marked according to the result whether to place an order with obtain each history feature to
Measure corresponding label；
The building of disaggregated model is carried out using XGBOOST algorithm using the history feature vector and the corresponding label, with
Obtain the prediction model.
6. hotel's sort method under OTA platform as described in claim 1 based on POI, which is characterized in that
The probability to place an order to the user to be recommended to each first hotel is ranked up are as follows:
The probability to place an order to the user to be recommended to each first hotel according to being ranked up from high to low, to obtain
One sequence；
Hotel's sort method is further comprising the steps of:
Hotel in the First ray is ranked up according to scoring, to obtain the second sequence；
Result after the output sequence is to export second sequence.
7. hotel's sort method under OTA platform as claimed in claim 6 based on POI, which is characterized in that hotel's sequence
Method further include:
The selectable hotel that the distance is arranged more than or equal to the distance threshold is the second hotel, to described second
Hotel is ranked up to obtain third sequence；
By the third sequence permutation to second sequence later to obtain the 4th sequence；
Result after the output sequence is to export the 4th sequence.
8. hotel's ordering system under a kind of OTA platform based on POI characterized by comprising
Region obtains module, and the POI for being inputted according to user to be recommended obtains the region where the POI；
Hotel may be selected and obtain module, for obtaining all selectable hotels in the region；
The distance is arranged for calculating each selectable hotel at a distance from the POI in first hotel's setup module
The selectable hotel less than distance threshold is the first hotel；
User's dimension obtains module, for obtaining user's dimensional information of the user to be recommended；
Preference computing module, for calculating the user to be recommended to the preference in each hotel under the OTA platform；
Hotel's dimension obtains module, for obtaining hotel's dimensional information in each first hotel；
Prediction module, for tieing up the hotel of user's dimensional information of the user to be recommended, each first hotel
Degree information and the user to be recommended are input to a prediction model to the preference in each hotel, are calculated using the prediction model
Obtain the probability that the user to be recommended places an order to each first hotel；
Sorting module, the probability for placing an order to the user to be recommended to each first hotel are ranked up；
Output module, for exporting the result after sorting.
9. hotel's ordering system under OTA platform as claimed in claim 8 based on POI, which is characterized in that
User's dimensional information includes the essential information of the user, user's dimensional information further include the user with
Into the information in the hotel that the information in the hotel that the information of the order in the hotel ordered, the user browsed, the user collect
At least one；
Hotel's dimensional information includes the essential information in the hotel, and hotel's dimensional information further includes the valence in the hotel
Lattice information, the favorable comment degree information in the hotel and the hotel are by least one of collection number information.
10. hotel's ordering system under OTA platform as claimed in claim 8 based on POI, which is characterized in that the preference
Computing module includes:
First hidden variable acquiring unit, for obtaining the hidden variable of all users in the OTA platform using hidden semantic model
And the hidden variable in all hotels in the OTA platform；
Second hidden variable acquiring unit obtains the hidden variable of the user to be recommended according to the hidden variable of all users；
Preference computing unit, for the user to be recommended to be calculated to the preference in each hotel by following formula
Degree:
Wherein, lfmijIndicate the user i to be recommended to the preference of each hotel j, uikIndicate described to be recommended
The kth dimension values of the vector of the hidden variable of user i, hjkIndicate the kth dimension values of the vector of the hidden variable of each hotel j, m
Indicate the length of the vector of the hidden variable in the user and each hotel.
11. hotel's ordering system under OTA platform as claimed in claim 10 based on POI, which is characterized in that described first is hidden
Variable acquiring unit includes:
First obtains subelement, for obtaining all users；
Second obtains subelement, for obtaining the score data in all hotels；
Hidden variable computation subunit, for the commenting using all users and all hotels using the hidden semantic model
The hidden variable of all users and the hidden variable in all hotels is calculated in divided data.
12. hotel's ordering system under OTA platform as claimed in claim 8 based on POI, which is characterized in that the hotel row
Sequence system further includes model generation unit, and the model generation unit includes:
Order acquiring unit, for obtaining History Order data, every History Order data include corresponding historical user
With history hotel and the result whether to place an order；
Data processing unit, it is corresponding to obtain the historical user for handling History Order data described in every
User's dimensional information, the corresponding hotel's dimensional information in the history hotel and the user to be recommended are to each hotel
Preference composition history feature vector；
Marking unit, for History Order described in every being marked according to the result whether to place an order to obtain each institute
State the corresponding label of history feature vector；
Model construction unit, for being carried out using the history feature vector and the corresponding label using XGBOOST algorithm
The building of disaggregated model, to obtain the prediction model.
13. hotel's ordering system under OTA platform as claimed in claim 8 based on POI, which is characterized in that the sequence mould
Block includes the first sequencing unit and the second sequencing unit；
The probability that first sequencing unit is used to place an order to each first hotel to the user to be recommended is according to from height
It is ranked up to low, to obtain First ray；
Second sequencing unit is for being ranked up the hotel in the First ray according to scoring, to obtain the second sequence
Column；
Result in the output module after output sequence is to export second sequence.
14. hotel's ordering system under OTA platform as claimed in claim 13 based on POI, which is characterized in that the sequence mould
Block further include: third sequencing unit and the 4th sequencing unit；
The third sequencing unit is used to be arranged the selectable hotel that the distance is greater than or equal to the distance threshold
For the second hotel, second hotel is ranked up to obtain third sequence；
4th sequencing unit is used for the third sequence permutation to second sequence later to obtain the 4th sequence；
Result in the output module after output sequence is to export the 4th sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811638493.6A CN109740072B (en) | 2018-12-29 | 2018-12-29 | Hotel ordering method and system based on POI under OTA platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811638493.6A CN109740072B (en) | 2018-12-29 | 2018-12-29 | Hotel ordering method and system based on POI under OTA platform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109740072A true CN109740072A (en) | 2019-05-10 |
| CN109740072B CN109740072B (en) | 2023-07-04 |
Family
ID=66362360
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811638493.6A Active CN109740072B (en) | 2018-12-29 | 2018-12-29 | Hotel ordering method and system based on POI under OTA platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109740072B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110619558A (en) * | 2019-09-18 | 2019-12-27 | 北京三快在线科技有限公司 | Search sorting method, system and equipment for online lodging products |
| CN110633370A (en) * | 2019-09-19 | 2019-12-31 | 携程计算机技术（上海）有限公司 | Generation method, system, electronic device and medium of OTA hotel label |
| CN111445280A (en) * | 2020-03-10 | 2020-07-24 | 携程计算机技术（上海）有限公司 | Model generation method, restaurant ranking method, system, device and medium |
| CN113590937A (en) * | 2021-07-05 | 2021-11-02 | 深圳市天下房仓科技有限公司 | Hotel searching and information management method and device, electronic equipment and storage medium |
| CN113763112A (en) * | 2021-02-25 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Information pushing method and device |
Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5948040A (en) * | 1994-06-24 | 1999-09-07 | Delorme Publishing Co. | Travel reservation information and planning system |
| JP2004061503A (en) * | 2002-07-18 | 2004-02-26 | Alpine Electronics Inc | Point-of-interest (poi) information display method and navigation system |
| US20060287810A1 (en) * | 2005-06-16 | 2006-12-21 | Pasha Sadri | Systems and methods for determining a relevance rank for a point of interest |
| US20070032942A1 (en) * | 2005-08-04 | 2007-02-08 | Microsoft Corporation | Ranking landmarks in a geographical area |
| US20100305842A1 (en) * | 2009-05-27 | 2010-12-02 | Alpine Electronics, Inc. | METHOD AND APPARATUS TO FILTER AND DISPLAY ONLY POIs CLOSEST TO A ROUTE |
| US20120197696A1 (en) * | 2011-02-02 | 2012-08-02 | Mapquest, Inc. | Systems and methods for generating electronic map displays with points-of-interest information |
| US20120221363A1 (en) * | 2011-02-25 | 2012-08-30 | Hipmunk, Inc. | System and method for displaying hotel information |
| CN102867031A (en) * | 2012-08-27 | 2013-01-09 | 百度在线网络技术（北京）有限公司 | Method and system for optimizing point of interest (POI) searching results, mobile terminal and server |
| EP2581703A1 (en) * | 2011-10-12 | 2013-04-17 | Mapquest, Inc. | Systems and methods for ranking points of interest |
| CN103150309A (en) * | 2011-12-07 | 2013-06-12 | 清华大学 | Method and system for searching POI (Point of Interest) points of awareness map in space direction |
| US20130262479A1 (en) * | 2011-10-08 | 2013-10-03 | Alohar Mobile Inc. | Points of interest (poi) ranking based on mobile user related data |
| US20130321466A1 (en) * | 2012-06-05 | 2013-12-05 | Kenneth L. Kocienda | Determining to Display Designations of Points of Interest Within a Map View |
| US20150073941A1 (en) * | 2011-12-13 | 2015-03-12 | Emma Burrows | Hotel finder interface |
| CN104504064A (en) * | 2014-12-19 | 2015-04-08 | 百度在线网络技术（北京）有限公司 | Information recommendation method and device |
| CN105589899A (en) * | 2014-11-18 | 2016-05-18 | 北京四维图新科技股份有限公司 | Display method and display device of point of interest in electronic map |
| CN105653736A (en) * | 2016-03-01 | 2016-06-08 | 北京师范大学 | Interest point group recommendation method based on geographical locations |
| CN106709767A (en) * | 2017-03-07 | 2017-05-24 | 携程计算机技术（上海）有限公司 | Personalized recommendation method and system of OTA (online travel website) hotels |
| CN107194774A (en) * | 2017-05-22 | 2017-09-22 | 携程旅游网络技术（上海）有限公司 | Personalized hotel's commending system and method in OTA websites |
| CN107507042A (en) * | 2017-09-15 | 2017-12-22 | 携程计算机技术（上海）有限公司 | Marketing method and system based on user's portrait |
| US9857177B1 (en) * | 2012-06-20 | 2018-01-02 | Amazon Technologies, Inc. | Personalized points of interest for mapping applications |
| CN107688662A (en) * | 2017-09-08 | 2018-02-13 | 携程计算机技术（上海）有限公司 | The recommendation method and system in OTA hotels |
| CN108717640A (en) * | 2018-01-24 | 2018-10-30 | 北京穷游天下科技发展有限公司 | The data processing method and electronic equipment of travel information |
-
2018
- 2018-12-29 CN CN201811638493.6A patent/CN109740072B/en active Active
Patent Citations (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5948040A (en) * | 1994-06-24 | 1999-09-07 | Delorme Publishing Co. | Travel reservation information and planning system |
| JP2004061503A (en) * | 2002-07-18 | 2004-02-26 | Alpine Electronics Inc | Point-of-interest (poi) information display method and navigation system |
| US20040204821A1 (en) * | 2002-07-18 | 2004-10-14 | Tu Ihung S. | Navigation method and system for extracting, sorting and displaying POI information |
| US20060287810A1 (en) * | 2005-06-16 | 2006-12-21 | Pasha Sadri | Systems and methods for determining a relevance rank for a point of interest |
| US20070032942A1 (en) * | 2005-08-04 | 2007-02-08 | Microsoft Corporation | Ranking landmarks in a geographical area |
| US20100305842A1 (en) * | 2009-05-27 | 2010-12-02 | Alpine Electronics, Inc. | METHOD AND APPARATUS TO FILTER AND DISPLAY ONLY POIs CLOSEST TO A ROUTE |
| US20120197696A1 (en) * | 2011-02-02 | 2012-08-02 | Mapquest, Inc. | Systems and methods for generating electronic map displays with points-of-interest information |
| US20120221363A1 (en) * | 2011-02-25 | 2012-08-30 | Hipmunk, Inc. | System and method for displaying hotel information |
| US20130262479A1 (en) * | 2011-10-08 | 2013-10-03 | Alohar Mobile Inc. | Points of interest (poi) ranking based on mobile user related data |
| EP2581703A1 (en) * | 2011-10-12 | 2013-04-17 | Mapquest, Inc. | Systems and methods for ranking points of interest |
| CN103150309A (en) * | 2011-12-07 | 2013-06-12 | 清华大学 | Method and system for searching POI (Point of Interest) points of awareness map in space direction |
| US20150073941A1 (en) * | 2011-12-13 | 2015-03-12 | Emma Burrows | Hotel finder interface |
| US20130321466A1 (en) * | 2012-06-05 | 2013-12-05 | Kenneth L. Kocienda | Determining to Display Designations of Points of Interest Within a Map View |
| US9857177B1 (en) * | 2012-06-20 | 2018-01-02 | Amazon Technologies, Inc. | Personalized points of interest for mapping applications |
| CN102867031A (en) * | 2012-08-27 | 2013-01-09 | 百度在线网络技术（北京）有限公司 | Method and system for optimizing point of interest (POI) searching results, mobile terminal and server |
| CN105589899A (en) * | 2014-11-18 | 2016-05-18 | 北京四维图新科技股份有限公司 | Display method and display device of point of interest in electronic map |
| CN104504064A (en) * | 2014-12-19 | 2015-04-08 | 百度在线网络技术（北京）有限公司 | Information recommendation method and device |
| CN105653736A (en) * | 2016-03-01 | 2016-06-08 | 北京师范大学 | Interest point group recommendation method based on geographical locations |
| CN106709767A (en) * | 2017-03-07 | 2017-05-24 | 携程计算机技术（上海）有限公司 | Personalized recommendation method and system of OTA (online travel website) hotels |
| CN107194774A (en) * | 2017-05-22 | 2017-09-22 | 携程旅游网络技术（上海）有限公司 | Personalized hotel's commending system and method in OTA websites |
| CN107688662A (en) * | 2017-09-08 | 2018-02-13 | 携程计算机技术（上海）有限公司 | The recommendation method and system in OTA hotels |
| CN107507042A (en) * | 2017-09-15 | 2017-12-22 | 携程计算机技术（上海）有限公司 | Marketing method and system based on user's portrait |
| CN108717640A (en) * | 2018-01-24 | 2018-10-30 | 北京穷游天下科技发展有限公司 | The data processing method and electronic equipment of travel information |
Non-Patent Citations (2)
| Title |
|---|
| D. GAVALAS: "Web application for recommending personalised mobile tourist routes", 《 IET SOFTWARE》, vol. 6, no. 4, pages 313, XP006039547, DOI: 10.1049/iet-sen.2011.0156 * |
| 连德富: "基于位置社交网络的数据挖掘", 《中国博士学位论文全文数据库 信息科技》, pages 138 - 31 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110619558A (en) * | 2019-09-18 | 2019-12-27 | 北京三快在线科技有限公司 | Search sorting method, system and equipment for online lodging products |
| CN110633370A (en) * | 2019-09-19 | 2019-12-31 | 携程计算机技术（上海）有限公司 | Generation method, system, electronic device and medium of OTA hotel label |
| CN111445280A (en) * | 2020-03-10 | 2020-07-24 | 携程计算机技术（上海）有限公司 | Model generation method, restaurant ranking method, system, device and medium |
| CN113763112A (en) * | 2021-02-25 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Information pushing method and device |
| CN113590937A (en) * | 2021-07-05 | 2021-11-02 | 深圳市天下房仓科技有限公司 | Hotel searching and information management method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109740072B (en) | 2023-07-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109740072A (en) | Hotel's sort method and system under OTA platform based on POI | |
| CN105675002B (en) | A kind of more transit point navigation route planning method and systems | |
| CN105139301B (en) | A kind of guidance method based on BP neural network | |
| CN106951925A (en) | Data processing method, device, server and system | |
| CN108961245A (en) | Picture quality classification method based on binary channels depth parallel-convolution network | |
| CN107239203A (en) | A kind of image management method and device | |
| CN106844614A (en) | A kind of floor plan functional area system for rapidly identifying | |
| CN107194774A (en) | Personalized hotel's commending system and method in OTA websites | |
| CN106021363A (en) | Course recommendation method and device | |
| CN108492290A (en) | Image evaluation method and system | |
| CN108446351B (en) | Hotel screening method and system based on user preference of OTA platform | |
| CN106709767A (en) | Personalized recommendation method and system of OTA (online travel website) hotels | |
| CN110503154A (en) | Method, system, electronic equipment and the storage medium of image classification | |
| CN107688662A (en) | The recommendation method and system in OTA hotels | |
| CN106875028A (en) | A kind of method for aiding in selecting a school and system | |
| CN106484733A (en) | News clue personalized push method and system | |
| CN102053960B (en) | Method and system for constructing quick and accurate Internet of things and Internet search engine according to group requirement characteristics | |
| CN107895299A (en) | The exposure sort method and device of a kind of commodity | |
| CN110110663A (en) | A kind of age recognition methods and system based on face character | |
| CN106846082A (en) | Tourism cold start-up consumer products commending system and method based on hardware information | |
| CN110264284A (en) | New media marketing attributive analysis prediction technique, apparatus and system | |
| CN107577736A (en) | A kind of file recommendation method and system based on BP neural network | |
| CN107026755B (en) | A kind of service recommendation method based on sequence study | |
| CN110209916B (en) | Method and device for recommending point of interest images | |
| CN106980613A (en) | One kind search air navigation aid and equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |