CN105955981A - Personalized travel package recommendation method based on demand classification and subject analysis - Google Patents
Personalized travel package recommendation method based on demand classification and subject analysis Download PDFInfo
- Publication number
- CN105955981A CN105955981A CN201610237787.2A CN201610237787A CN105955981A CN 105955981 A CN105955981 A CN 105955981A CN 201610237787 A CN201610237787 A CN 201610237787A CN 105955981 A CN105955981 A CN 105955981A
- Authority
- CN
- China
- Prior art keywords
- service
- user
- demand
- traveling bag
- list
- 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
- 238000000034 method Methods 0.000 title claims abstract description 62
- 238000004458 analytical method Methods 0.000 title claims abstract description 26
- 230000008569 process Effects 0.000 claims abstract description 29
- 230000003203 everyday effect Effects 0.000 claims abstract description 11
- 238000013461 design Methods 0.000 claims abstract description 9
- 230000011218 segmentation Effects 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 4
- 235000013305 food Nutrition 0.000 claims description 20
- 235000013361 beverage Nutrition 0.000 claims description 13
- 238000009826 distribution Methods 0.000 claims description 11
- 230000008878 coupling Effects 0.000 claims description 8
- 238000010168 coupling process Methods 0.000 claims description 8
- 238000005859 coupling reaction Methods 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 4
- 230000004069 differentiation Effects 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 239000013589 supplement Substances 0.000 claims description 3
- 239000011717 all-trans-retinol Substances 0.000 claims description 2
- 230000013011 mating Effects 0.000 claims 2
- 230000006870 function Effects 0.000 abstract description 6
- 238000003058 natural language processing Methods 0.000 abstract 1
- 238000005457 optimization Methods 0.000 description 12
- 239000000047 product Substances 0.000 description 8
- 235000012054 meals Nutrition 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000005304 joining Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 241000283074 Equus asinus Species 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 235000021152 breakfast Nutrition 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000002620 method output Methods 0.000 description 1
- 238000000465 moulding Methods 0.000 description 1
- 238000012856 packing Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/31—Indexing; Data structures therefor; Storage structures
- G06F16/313—Selection or weighting of terms for indexing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
-
- 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
- G06Q50/14—Travel agencies
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Human Computer Interaction (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a personalized travel package recommendation method based on demand classification and subject analysis. The method includes: analyzing natural language form demands input by a user, utilizing word segmentation, demand classification and other natural language processing techniques to process and classify the user demands so as to obtain rigid demand, flexible demand and negative demand of the user; then utilizing an LDA (latent dirichlet allocation) document theme generation model, making travel service individuals effectively cluster into different service fields by theme similarity, and then conducting similarity matching with the user demands so as to obtain a service list best matching user expectation; finally carrying out travel package design recommendation by means of travel package optimized recommendation algorithm: firstly acquiring a travel package scenic spot set according to user time demand and service priority information; then combining location information, preference information and the like to determine travel package hotel service; and then calculating the optimal journey of every day according to the distance function L, and ranking the travel package according to travel package recommendation index; and finally, selecting travel package catering service according to scenic spot location and user preference. By integrating the processing, the purpose of designing and recommending personalized travel package best meeting the user demand can be realized.
Description
Technical field
The present invention relates to a kind of personalized traveling bag based on demand classification and subject analysis and recommend method.The method
From the input of user, extract demand, and desired content is classified；In conjunction with subject analysis, generate alternative
Set of service；On this basis, traveling bag optimization is utilized to recommend method to carry out traveling bag design and recommendation for user.
The method belongs to computer system modeling and data analysis field.
Background technology
In recent years, along with the fast development of Internet technology and improving constantly of living standards of the people, online tourism
Industry Quick Development, also with service-oriented computing (Service Oriented Computing is called for short SOC)
Rise, people get used to service concept to treat each single travelling products (such as sight spot, hotel etc.),
And treat traveling bag with the visual angle of Services Composition, service recommendation and recommend problem.Currently, people can facilitate
Ground in online browsing, inquire about the complete travel solutions that product of service in tourism interested or travel agency issue.But,
The most personalized, detail along with the constantly aggravation of internet information overload phenomenon and people's demand, traditional
Online tourism service shows following drawback:
1) information overload is serious.Although user can scan on tradition online tourism platform, search sense
The tourist service of interest, but owing to internet information amount is excessive, and it is full of a large amount of garbage so that user
Search efficiency declines, and is difficult to quickly find the service meeting oneself demand；
2) lacking individuality customization service.In addition to single product of service in tourism, a large amount of online tourism platforms are also
Set of travelling solution is provided to select for user.But these solutions be all changeless (or
The fixing product that travel agency provides, or " donkey friend " fixing attack strategy of issuing), and the demand of user is often
Changeable, be difficult to find the molding scheme complying fully with requirement, this just require online tourism platform can according to
Family demand, automatically, independently generates new personalized solution, namely traveling bag, it is recommended that to user；
3) user's natural language demand can not be processed.The many patterns with fixed options of current each online tourism platform
Require that user selects Reiseziel, goes sight-seeing the information such as natural law, travelling budget, and the fixed options that platform provides
Typically can not contain institute's likely demand of user.For a user, most convenient, be also to express the most flexibly
The mode of demand and hobby is exactly the form one section of demand of input with plain text natural language, and current each online trip
Trip platform all can not use effective technology to process this type of demand.
In order to solve the above technical problem faced in traveling bag recommendation process, the present invention proposes a kind of based on needing
Method recommended by the personalized traveling bag asking classification and subject analysis: the method can effectively analyze, process user from
So language needs, extracts user's request and classifies, in conjunction with subject analysis technology, generating alternative travelling
Set of service, finally utilizes traveling bag optimization to recommend method, carries out the personalized designs of traveling bag for user and push away
Recommend.
In the present invention, travel service refers to single travel products, the most single sight spot, hotel, restaurant
Etc. being considered single travel service；The concept of traveling bag (travel package/tour package) is
Being different from single product of service in tourism, it is a kind of set of travelling products, and it is by multiple single outing dresses
Business product packing, recommends with an entirety, contains the food and drink in travelling, lodging, traffic, sight spot etc.
Multiple elements in four broad aspect, provide the user a whole set of travelling solution.
Summary of the invention
Present invention firstly provides a kind of personalized traveling bag based on demand classification and subject analysis and recommend method.This
The method of invention comprises three processes: 1) user's request categorizing process based on key word dictionary；2) based on master
The alternative services Set-search process that topic is analyzed；3) based on distance optimized traveling bag recommendation process:
1) user's request categorizing process based on key word dictionary comprises the following two stage:
(1) demand of the natural language form of user's input is carried out the pretreatment work such as participle；
(2) according to key word dictionary, the key word extracted is carried out demand differentiation, respectively obtain the rigid of user
Demand, soft demand and negative demand.
2) alternative services Set-search process based on subject analysis comprises the three below stage:
(1) according to the service description document of service each in data base, LDA (Latent Dirichlet is utilized
Allocation) document subject matter generates model extraction and respectively services the probability distribution of theme, and deposits
Storage is in " service theme dictionary " storehouse；
(2) according to the user's hard requirement obtained and soft demand, based on LDA subject analysis, and by
" service theme dictionary " storehouse coupling draws the alternative services list similar to these demands；
(3) in alternative services list, supplement the service in hard requirement, then reject the service in negative demand,
Hard requirement service priority is labeled as height, and other service priority are labeled as low, final
The alternative services set recommended to traveling bag, for generation and the recommendation of traveling bag.
3) the following four stage is comprised based on distance optimized traveling bag recommendation process:
(1) alternative sight spot combination is generated.Natural law according to user requires and alternative services obtained above
Set, the preliminary sight spot combination generated in traveling bag, obtain alternative a series of traveling bags.?
During the generation of traveling bag, the hard requirement of user can obtain preferential meeting；
(2) selected lodging hotel.In conjunction with geographical location information and the preference of user at sight spot, selected traveling bag
In lodging hotel；
(3) concrete stroke sequence traveling bag side by side is formulated according to distance function.Calculate every according to distance function L
It optimum stroke, obtains the expection distance of each traveling bag, and combines precedence information, calculate
Index recommended by traveling bag, in this, as according to sequence traveling bag.Expection distance is the shortest, priority
The highest traveling bag can obtain preferential recommendation；
(4) selected traveling bag food and drink on the way.According to position, sight spot and the preference of user, for the selected travelling of user
Food and beverage sevice in Tu.
The present invention uses personalized traveling bag based on demand classification and subject analysis to recommend method, has following two
The advantage of aspect:
(1) by participle and the use of key word dictionary, the demand type of user is made a distinction, refines
The natural language form demand of user, to obtain recommendation results more accurately；
(2) present invention can use subject analysis, generates alternative Services Composition, and based on distance optimization
Principle automatically for user generate and recommend traveling bag.
Accompanying drawing explanation
Recommendation method is had been described in detail by the application by accompanying drawing, and these descriptions are merely to illustrate the present invention's
Content, is not intended to limit the present invention.
Fig. 1 is that in the present invention, the overall frame of method recommended by personalized traveling bag based on demand classification and subject analysis
Figure.
Fig. 2 is that in the present invention, traveling bag optimizes recommendation process detailed step block diagram.
Detailed description of the invention
Symbolic significance involved in the present invention and concrete steps are described.
Fig. 1 illustrates personalized traveling bag based on demand classification and subject analysis and recommends method the general frame.As
Shown in Fig. 1, user inputs natural language form demand, comprises Reiseziel, it is desirable to travelling natural law and possibility
Individual requirement etc., the present invention first to user input carry out word segmentation processing and demand classification, analyze respectively
Hard requirement, soft demand and negative demand, obtain alternative set of service then in conjunction with subject analysis technology,
The recommendation traveling bag after method output sequence is recommended afterwards by traveling bag optimization.The method of the present invention comprises three mistakes
Journey: 1) user's request categorizing process based on key word dictionary；2) alternative services set based on subject analysis
Search procedure；3) based on distance optimized traveling bag recommendation process.
Fig. 2 illustrates traveling bag and optimizes the detailed step block diagram of recommendation process.As in figure 2 it is shown, method input letter
Breath requires N, alternative services set S for travelling natural lawal' (comprise: alternative sight spot set of service S, alternative meal
Drink set of service R, alternative hotel service set H), first according to the subset of N screening S, generate user and exist
Whole sight spots set S of visit in N daysF, it is every further according to the priority of H and " center of gravity " of sight spot distance
One SFDetermine lodging hotel, then utilize distance function L solving-optimizing problem, obtain the optimum row of every day
Journey arranges, and then according to recommending index I sequence traveling bag, the most preferential finally according to R and routing every day
Level and range information determine food and drink on the way, the final optimization recommendation results exporting traveling bag.Specifically, process
Comprise the following four stage: generate alternative sight spot combination.Natural law according to user requires and obtained above
Alternative services set, the preliminary sight spot combination generated in traveling bag, obtain alternative a series of traveling bags.In trip
During the generation of row bag, the hard requirement of user can obtain preferential meeting；Selected lodging hotel.In conjunction with scape
The geographical location information of point and the preference of user, the lodging hotel in selected traveling bag；Formulate according to distance function
Concrete stroke sequence traveling bag side by side.Calculate optimum stroke every day according to distance function L, obtain each traveling bag
Expection distance, and combine precedence information, calculate traveling bag and recommend index, in this, as according to sequence travelling
Bag.The expection traveling bag that distance is the shortest, priority is the highest can obtain preferential recommendation；Selected traveling bag is eaten on the way
Drink.According to position, sight spot and the preference of user, select the food and beverage sevice in travelling way for user.
It is embodied as step and is divided into three parts:
Part I (step 1～step 4), user's request based on key word dictionary is classified；Part II (step
Rapid 5～step 8), alternative services Set-search based on subject analysis；Part III (step 9～step 13),
Recommend based on the optimized traveling bag of distance.
User's request based on key word dictionary is divided into two stages
Stage 1 corresponding step 1
Step 1: the demand of user's natural language form input is carried out word segmentation processing, carries out Chinese word segmentation (example
As, call wvtool participle instrument), utilize stop words dictionary remove unrelated word (such as ", we,
" etc.), user is inputted the vocabulary vector that requirement representation is a string order, is denoted as Q={q1,q2,...,qp}.Example
As, when user's input " I thinks that Beijing is played three days, eats Quanjude, likes staying in center, city, be not desired to the Temple of Heaven "
Time, word segmentation result for " think, Beijing, object for appreciation, three days, eat, Quanjude, like, stay in, center, city,
Be not desired to, the Temple of Heaven ".
Stage 2 corresponding step 2, step 3, step 4
Step 2: word segmentation result Q in step 1 mated with negative demand key word dictionary, negative demand is closed
Keyword dictionary stores common negative demand key word, such as " be not desired to, do not go, do not like " etc..Work as generation
After coupling, negative demand list will be listed in by negative demand key word noun followed by, be denoted as
NQ={nq1,nq2,...,nqnp, represent that user does not expect that this services.Such as example in step 1, " is not desired to
Go " it is negative demand key word, noun " the Temple of Heaven " the most followed by is i.e. put into negative demand list, rear
In continuous recommendation process, we will not recommend " the Temple of Heaven " this sight spot to service.After step 2 terminates, at Q
Middle rejecting negative demand key word and the vocabulary being put into negative demand list, more newly obtained Q'.
Step 3: mated with service dictionaries by the Q' obtained in step 2, stores data in service dictionaries
The title of each service and each common first names in storehouse, such as " Summer Palace, Quanjude, the four seasons " etc..When generation
After joining, the service standard title that the noun mated will be occurred corresponding lists hard requirement list in, is denoted as
MQ={mq1,mq2,...,mqmp, represent that these demands of user need preferentially to be met.Such as in step 1
Example, " Quanjude " is i.e. put into hard requirement list, and in follow-up recommendation process, we will be preferential
Recommend " Quanjude " this food and beverage sevice.Additionally, we are also by the destination information (D) of specific aim identification user
With travel time information (N), process as special hard requirement, also will be updated to by MQ
MQ={D, N, mq1,mq2,...,mqmp, such as " Beijing " of example, " three days " in step 1.Step 3
After end, Q' rejects hard requirement key word, more newly obtained Q ".
Step 4: the Q that will obtain in step 3 " mate with soft demand dictionary, soft demand dictionary is deposited
Store up each conventional Preference demand key word, such as " peppery, western-style food, peace and quiet, history inside information " etc..When generation
After joining, the vocabulary of coupling will be occurred to list soft list of requirements in, be denoted as SQ={sq1,sq2,...,sqsp, represent and use
Family intentionally gets this type of service.Such as example in step 1, " city's " center " is i.e. put into soft list of requirements,
In follow-up recommendation process, we meet " the hotel service of city's " center " condition by recommending.
So far, user's natural language demand is successfully processed and has been classified by we, respectively obtains user rigid
List of requirements MQ={D, N, mq1,mq2,...,mqmp, soft list of requirements SQ={sq1,sq2,...,sqspAnd negative demand
List NQ={nq1,nq2,...,nqnp}。
Alternative services Set-search based on subject analysis is divided into three phases
Stage 1 corresponding step 5
Step 5: text based on service each in service dictionaries and service database describes document, utilizes LDA
Theme probabilistic model obtains the theme distribution of each service, carrys out formalized description with matrix ST.ST is Ns×Ks
Dimension matrix, NsRepresent the sum of serve individual, KsRepresent the theme sum that serve individual relates to.ST matrix
(i j) represents T to element S T of the i-th row jth rowjAt S'iIn distribution of weights, wherein TjRepresent in serve individual
Theme j, S'iRepresent serve individual i.The theme distribution of one serve individual of each behavior of matrix.This ST
Matrix is i.e. stored in " service theme dictionary " storehouse.
Stage 2 corresponding step 6, step 7
Step 6: first, according to the user's hard requirement list obtained in step 3
MQ={D, N, mq1,mq2,...,mqmpCustomer objective ground information (D) in }, in service database, screening is available
Service total list S ".(when such as D represents Beijing, S " represent whole sight spot, incity, Beijing, hotel, meal
Drink and other services) then, the user's hard requirement list that will obtain in step 3
MQ={D, N, mq1,mq2,...,mqmpAnd step 4 in soft list of requirements SQ={sq that obtains1,sq2,...,sqspEnter
Row merges, and obtains user's request total list QMS={ mq1,mq2,...,mqmp,sq1,sq2,...,sqsp}。
Step 7: the distribution probability of theme is arrived in each service utilizing LDA subject analysis in step 5 to draw
P (T | S') (each ST (i, j) item) of being i.e. stored in the ST matrix in " service theme dictionary " storehouse and
Each theme is to the distribution probability p (Q of vocabularyMS| T), calculate each service and user's request total list QMSSimilarity,
Specifically it is calculated as follows:And then according to similarity
p(QMS|S”i) size from high to low to S " in service be ranked up, obtain alternative services list Sal, its
In the priority of each service i.e. represent by its normalized similarity.The service that similarity is the biggest will be in recommendation
It is prioritized.
Stage 3 corresponding step 8
Step 8: the user hard requirement list MQ={D that will obtain in step 3, N, mq1,mq2,...,mqmpIn }
Service { mq1,mq2,...,mqmpAdd to alternative services list SalIn, and by { mq1,mq2,...,mqmpWhole in }
Service priority is labeled as 1.The negative demand list NQ={nq that will obtain in step 2 again1,nq2,...,nqnpIn }
Service is from alternative services list SalMiddle rejecting, finally gives the alternative services list S after renewalal'.Each service
All comprising the attribute of a priority, value is 0～1, and this service of the biggest expression priority when combination is the highest.
The service priority obtained by hard requirement is 1, and remaining is the value of the similarity calculated in step 7.To alternative
Service list SalService in ' is divided into following three classes:
1) alternative sight spot set of service S.Each sight spot services package contains two attributes, priority SvWith required trip
Object for appreciation time St.Priority SvSpan be 0～1, needed for play time StSpan be also 0～1.
StNeeded for expression, the time of playing accounts for whole day and may utilize the ratio of the time of playing, such as StValue is this scape of 0.5 expression
Point needs time half a day to go sight-seeing；
2) alternative food and beverage sevice set R.Each food and beverage sevice comprises attribute priority Rv, its span is
0～1；
3) alternative hotel service set H.Each hotel service comprises attribute priority Hv, its span is
0～1.
Recommend to be divided into four-stage based on the optimized traveling bag of distance
Stage 1 corresponding step 9
Step 9: alternative sight spot set of service S obtained according to travelling natural law requirement N and the step 8 of user
Solving-optimizing problem, generates whole sight spots set S that user went sight-seeing in N daysF.First, alternative sight spot clothes
Business set S medium priority SvSet S is the most all included at the sight spot of=1 inFIn, hereafter solve following optimization problem:
Wherein a is arithmetic number, and optimization aim is that the sight spot service priority making recommendation adds and maximizes, and constraint representation pushes away
The time of always playing recommending sight spot requires N equal to the travelling natural law of user, introduces error parameter ξ (desirable herein
Meaning real number), it is allowed to there is certain error in natural law of always playing, but by the adjustment to coefficient a, it is ensured that error
Within the acceptable range.Obtain the S meeting the problems referred to aboveFAll feasible solutions after, i.e. obtain alternative trip
Row bag set SFall={ SF1,SF2,...,SFn}。
Stage 2 corresponding step 10
Step 10: respectively to the alternative traveling bag set S tried to achieve in step 9Fall={ SF1,SF2,...,SFnEvery in }
One SFiThe corresponding hotel service of calculating sifting:
If alternative hotel service set H having and only priority HvThe hotel of=1, the most all SFiAll
Arrange to recommend this hotel；
If alternative hotel service set H has multiple priority HvThe hotel of=1, might as well remember that these hotels gather
For H', now solve following optimization problem and obtain each SFiThe corresponding hotel H recommendedi:
s.t.Hij∈H'
WhereinRepresent sight spot S to hotel HijTraffic distance；
If alternative hotel service set H does not has priority HvThe hotel of=1, then ask by solving following optimization
Topic obtains each SFiThe corresponding hotel H recommendedi:
s.t.Hij∈H
Wherein a is arithmetic number.
To sum up, we have obtained alternative traveling bag set SFall={ SF1,SF2,...,SFnEach S in }FiCorresponding
The hotel service H recommendedi。
Stage 3 corresponding step 11, step 12
Step 11: to alternative traveling bag set SFall={ SF1,SF2,...,SFnEach S in }Fi, calculate corresponding
Optimum path planning, define distance functionRepresent traveling bag SFiThe
The shortest whole traffic distances needed for routing in j days, wherein njRepresent the sight spot number of jth sky visit.Enter
And to each SFiSolve following optimization problem, be desirably to obtain optimum traveling bag path planning:
S∈SFi
Wherein a is arithmetic number, [SFi] it is traveling bag SFiMiddle sight spot quantity, optimization aim is the distance making pass by every day
Adding up the total distance of travelling obtained to minimize, all sight spots of Section 1 constraint representation are both needed to be arranged stroke, and second
Item constraint represents that the sight spot of every day time of playing adds up and should be equal to " 1 ", namely one day, introduces error herein and joins
Number ξj(desirable any real number), it is allowed to the time of playing every day exists certain error, but by the tune to coefficient a
Whole, it is ensured that error is within the acceptable range.The difference at sight spot is gone sight-seeing all possible every day by traversal
Integrated mode, finds the optimal solution of the problems referred to above, namely the path design that total visit traffic distance is the shortest,
And we try one's best service arrangement high for priority in stroke a few days ago.So far, we are each SFi
Have found the optimum routing of its correspondence.
Step 12: calculate the recommendation index of each traveling bagWherein b, c
For arithmetic number, and then according to recommending index I to sort from high to low each traveling bag, obtain traveling bag suggested design
SFall'={ SF1',SF2',...,SFn', when finally carrying out traveling bag recommendation for user, i.e. recommend according to this clooating sequence.
Stage 4 corresponding step 13
Step 13: design food and drink on the way for user according to the alternative food and beverage sevice set R that step 8 obtains, its
In often arrange morning, noon, late breakfast, lunch and dinner day by day.First, alternative food and beverage sevice set R medium priority RvThe meal of=1
Drink service priority is ensured, search and R in stroke planningvThe sight spot that the food and beverage sevice of=1 is close together,
Then the time for eating meals near this sight spot i.e. arranges this food and beverage sevice.Meeting RvAfter the food and beverage sevice of=1 is guaranteed,
Sight spot near the time for eating meals not yet arranging food and beverage sevice scans for, the R of comprehensive food and beverage sevicevScoring
(this weight is desirable relatively big, more to meet the hobby of user) and the distance at distance sight spot, recommend meal for user
Drink-service is engaged in.
Finally, we have obtained the traveling bag set S recommended for userFall'={ SF1',SF2',...,SFn', the most each
Traveling bag includes that lodging hotel, every day go sight-seeing sight spot and the information such as order and food and drink on the way, and according to step 12
The traveling bag clooating sequence calculated is that user recommends.
By the present invention, method recommended by a kind of personalized traveling bag based on demand classification and subject analysis: can
The demand utilizing the means such as participle, demand classification extraction to input user's natural language processes and classifies, and carries
Take the rigid, soft of family and negative demand；And then utilize LDA document subject matter to generate model, by theme phase
Effectively cluster individual for travel service in different service fields like degree, and then carry out similarity coupling, with
Go out to meet the most the desired service list of user；Traveling bag optimization is finally utilized to recommend method to carry out the row of traveling bag
Journey design, food and drink and hotel service design etc., recommend the traveling bag of the complete demand that best suits for user.
Claims (9)
1. method recommended by a personalized traveling bag based on demand classification and subject analysis, it is characterised in that described
Method comprises three processes: 1) user's request categorizing process based on key word dictionary, and 2) based on subject analysis
Alternative services Set-search process, 3) based on distance optimized traveling bag recommendation process；Wherein
1) user's request categorizing process based on key word dictionary includes the following two stage:
(1) demand of the natural language form of user's input is carried out the pretreatment work such as participle；
(2) according to key word dictionary, the key word extracted is carried out demand differentiation, respectively obtain the rigid of user
Demand, soft demand and negative demand；
2) alternative services Set-search process based on subject analysis includes the three below stage:
(1) according to the service description document of service each in data base, LDA (Latent Dirichlet is utilized
Allocation) document subject matter generates model extraction and respectively services the probability distribution of theme, and is stored in " service
Theme dictionary " in storehouse；
(2) according to the user's hard requirement obtained and soft demand, based on LDA subject analysis, and by
" service theme dictionary " storehouse coupling draws the alternative services list similar to these demands；
(3) in alternative services list, supplement the service in hard requirement, then reject the clothes in negative demand
Business, is labeled as height by hard requirement service priority, and other service priority are labeled as low, finally give travelling
The alternative services set that bag is recommended, for generation and the recommendation of traveling bag；
3) the following four stage is included based on distance optimized traveling bag recommendation process:
(1) require and alternative services set obtained above according to the natural law of user, generate alternative sight spot
Combination, obtains alternative a series of traveling bags, and during the generation of traveling bag, the hard requirement of user can obtain
Meet to preferential；
(2) geographical location information at sight spot and the preference of user are combined, the lodging hotel in selected traveling bag；
(3) determine optimum stroke every day according to distance, obtain the expection distance of each traveling bag, and combine excellent
First level information, calculates traveling bag and recommends index, and sort traveling bag；
(4) according to position, sight spot and the preference of user, the food and beverage sevice in travelling way is selected for user.
Method recommended by personalized traveling bag the most according to claim 1, it is characterised in that described (1) is right
The demand of the natural language form of user's input carries out the pretreatment work such as participle and includes:
Step 1: the demand of user's natural language form input is carried out word segmentation processing, carries out Chinese word segmentation, profit
Remove unrelated word with stop words dictionary, user is inputted the vocabulary vector that requirement representation is a string order, is denoted as
Q={q1,q2,...,qp}。
Method recommended by personalized traveling bag the most according to any one of claim 1 to 2, and its feature exists
In, described (2) carry out demand differentiation according to key word dictionary to the key word extracted, and respectively obtain user's
Hard requirement, soft demand and negative demand include:
Step 2: word segmentation result Q in step 1 mated with negative demand key word dictionary, negative demand is closed
Keyword dictionary stores common negative demand key word, will list in by negative demand key word noun followed by
Negative demand list, is denoted as NQ={nq1,nq2,...,nqnp, represent that user does not expect that this services, reject in Q
Negative demand key word and the vocabulary being put into negative demand list, more newly obtained Q'；
Step 3: mated with service dictionaries by the Q' obtained in step 2, stores data in service dictionaries
Each title serviced and each common first names in storehouse, after mating, the noun that coupling will occur is corresponding
Service standard title lists hard requirement list in, is denoted as MQ={mq1,mq2,...,mqmp, represent user these
Demand needs preferentially to be met, by destination information (D) and the travel time information of specific aim identification user
(N), process as special hard requirement, also will be updated to by MQ
MQ={D, N, mq1,mq2,...,mqmp, Q' rejects hard requirement key word, more newly obtained Q "；
Step 4: the Q that will obtain in step 3 " mate with soft demand dictionary, soft demand dictionary is deposited
Store up each conventional Preference demand key word, after mating, the vocabulary of coupling will have been occurred to list soft need in
Ask list, be denoted as SQ={sq1,sq2,...,sqsp, represent that user intentionally gets this type of service.
Method recommended by personalized traveling bag the most according to any one of claim 1 to 3, it is characterised in that
(1) according to the service description document of service each in data base, LDA (Latent Dirichlet Allocation) is utilized
Document subject matter generates model extraction and respectively services the probability distribution of theme, and is stored in " service theme dictionary "
Include in storehouse:
Step 5: text based on service each in service dictionaries and service database describes document, utilizes LDA
Theme probabilistic model obtains the theme distribution of each service, carrys out formalized description with matrix ST, this ST matrix
I.e. it is stored in " service theme dictionary " storehouse.
Method recommended by personalized traveling bag the most according to any one of claim 1 to 4, it is characterised in that
(2) according to the user's hard requirement obtained and soft demand, based on LDA subject analysis, and by " service
Theme dictionary " storehouse coupling show that the alternative services list similar to these demands includes:
Step 6: first, according to the user's hard requirement list obtained in step 3
MQ={D, N, mq1,mq2,...,mqmpCustomer objective ground information (D) in }, in service database, screening is available
Service total list S "；Then, the user's hard requirement list that will obtain in step 3
MQ={D, N, mq1,mq2,...,mqmpAnd step 4 in soft list of requirements SQ={sq that obtains1,sq2,...,sqspEnter
Row merges, and obtains user's request total list QMS={ mq1,mq2,...,mqmp,sq1,sq2,...,sqsp}；
Step 7: the distribution probability of theme is arrived in each service utilizing LDA subject analysis in step 5 to draw
P (T | S') and each theme are to the distribution probability p (Q of vocabularyMS| T), calculate each service and the total list of user's request
QMSSimilarity；And then according to similarity p (QMS|S”i) size from high to low to S " in service arrange
Sequence, obtains alternative services list Sal, wherein the priority of each service i.e. represents by its normalized similarity；
The service that similarity is the biggest will be prioritized in recommendation.
Method recommended by personalized traveling bag the most according to any one of claim 1 to 5, it is characterised in that
(3) in alternative services list, supplement the service in hard requirement, then reject the service in negative demand, will
Hard requirement service priority is labeled as height, and other service priority are labeled as low, finally gives traveling bag and recommends
Alternative services set, generation and recommendation for traveling bag include:
Step 8: the user hard requirement list MQ={D that will obtain in step 3, N, mq1,mq2,...,mqmpIn }
Service { mq1,mq2,...,mqmpAdd to alternative services list SalIn, and by { mq1,mq2,...,mqmpWhole in }
Service priority is labeled as 1；The negative demand list NQ={nq that will obtain in step 2 again1,nq2,...,nqnpIn }
Service is from alternative services list SalMiddle rejecting, finally gives the alternative services list S after renewalal'；Each service
All comprising the attribute of a priority, value is 0～1, and this service of the biggest expression priority when combination is the highest；
The service priority obtained by hard requirement is 1, and remaining is the value of the similarity calculated in step 7.
Method recommended by personalized traveling bag the most according to any one of claim 1 to 6, it is characterised in that
(1) require and alternative services set obtained above according to the natural law of user, generate alternative sight spot combination,
Obtaining alternative a series of traveling bags, during the generation of traveling bag, the hard requirement of user can obtain preferentially
Satisfied include:
Step 9: alternative sight spot set of service S obtained according to travelling natural law requirement N and the step 8 of user
Solving-optimizing problem, generates whole sight spots set S that user went sight-seeing in N daysF, obtain alternative traveling bag collection
Close SFall={ SF1,SF2,...,SFn}。
Method recommended by personalized traveling bag the most according to any one of claim 1 to 7, it is characterised in that
(2) combining the geographical location information at sight spot and the preference of user, the lodging hotel in selected traveling bag includes:
Step 10: respectively to the alternative traveling bag set S tried to achieve in step 9Fall={ SF1,SF2,...,SFnEvery in }
One SFiThe corresponding hotel service of calculating sifting, has obtained alternative traveling bag set SFall={ SF1,SF2,...,SFnIn }
Each SFiThe corresponding hotel service H recommendedi。
Method recommended by personalized traveling bag the most according to any one of claim 1 to 8, it is characterised in that
(3) determine optimum stroke every day according to distance, obtain the expection distance of each traveling bag, and combine priority
Information, calculates traveling bag and recommends index, sequence traveling bag to include:
Step 11: to alternative traveling bag set SFall={ SF1,SF2,...,SFnEach S in }Fi, calculate corresponding
Optimum path planning, define distance functionRepresent traveling bag SFiThe
The shortest whole traffic distances needed for routing in j days, wherein njRepresent the sight spot number of jth sky visit；Enter
And to each SFiSolving-optimizing problem, to obtain the traveling bag path planning of optimum；
Step 12: calculate the recommendation index I of each traveling bag:Wherein b, c
For arithmetic number, and then according to recommending index I to sort from high to low each traveling bag, obtain traveling bag suggested design
SFall'={ SF1',SF2',...,SFn'}。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610237787.2A CN105955981B (en) | 2016-04-15 | 2016-04-15 | A kind of personalized traveling bag recommended method based on demand classification and subject analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610237787.2A CN105955981B (en) | 2016-04-15 | 2016-04-15 | A kind of personalized traveling bag recommended method based on demand classification and subject analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105955981A true CN105955981A (en) | 2016-09-21 |
| CN105955981B CN105955981B (en) | 2019-04-23 |
Family
ID=56917570
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610237787.2A Active CN105955981B (en) | 2016-04-15 | 2016-04-15 | A kind of personalized traveling bag recommended method based on demand classification and subject analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105955981B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106775883A (en) * | 2016-12-26 | 2017-05-31 | 上海携程国际旅行社有限公司 | For the dynamic packaging method and packaging system of tourist resources |
| CN107145955A (en) * | 2017-04-11 | 2017-09-08 | 浙江大学 | A kind of stroke planning method based on condition and user preference |
| CN107977883A (en) * | 2017-11-24 | 2018-05-01 | 清华大学 | The recommendation method, apparatus and computer equipment of traveling bag |
| CN108829852A (en) * | 2018-06-21 | 2018-11-16 | 桂林电子科技大学 | A kind of individualized travel route recommended method |
| CN109146324A (en) * | 2018-09-13 | 2019-01-04 | 北京三快在线科技有限公司 | Recommended method, device and electronic equipment |
| CN109839120A (en) * | 2017-11-24 | 2019-06-04 | 北京三快在线科技有限公司 | Stroke planning method, device, medium and electronic equipment |
| CN109992245A (en) * | 2019-04-11 | 2019-07-09 | 河南师范大学 | A kind of method and system carrying out the modeling of science and technology in enterprise demand for services based on topic model |
| CN110147490A (en) * | 2017-08-07 | 2019-08-20 | 声音猎手公司 | Natural language recommends feedback |
| WO2019191868A1 (en) * | 2018-04-02 | 2019-10-10 | 徐璇炫 | Hotel-oriented conference planning method and conference management system |
| CN110457466A (en) * | 2019-06-28 | 2019-11-15 | 谭浩 | Generate method, computer readable storage medium and the terminal device of interview report |
| CN110895554A (en) * | 2018-09-12 | 2020-03-20 | 清华大学 | Destination recommendation method and device, computer equipment and storage medium |
| CN111143680A (en) * | 2019-12-27 | 2020-05-12 | 上海携程商务有限公司 | Method and system for recommending route, electronic device and computer storage medium |
| CN111210307A (en) * | 2020-01-03 | 2020-05-29 | 武汉理工大学 | Scientific and technological service chain intelligent recommendation system and method with response user preference as core |
| CN113435975A (en) * | 2021-06-29 | 2021-09-24 | 平安科技（深圳）有限公司 | Wheelchair leasing processing method and device and related equipment |
| CN115471188A (en) * | 2022-09-05 | 2022-12-13 | 漱玉平民大药房连锁股份有限公司 | Identification method and equipment for supply requirements of participant service ticket |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080319974A1 (en) * | 2007-06-21 | 2008-12-25 | Microsoft Corporation | Mining geographic knowledge using a location aware topic model |
| CN104463346A (en) * | 2014-11-04 | 2015-03-25 | 黄超超 | Travel information organizing device and method with hierarchical route map as framework |
| CN104679802A (en) * | 2013-11-28 | 2015-06-03 | 财团法人资讯工业策进会 | Travel planning device and travel planning method |
| CN105023205A (en) * | 2014-04-16 | 2015-11-04 | 珀斯特传媒有限公司 | Tour service system for providing recommended tour course |
-
2016
- 2016-04-15 CN CN201610237787.2A patent/CN105955981B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080319974A1 (en) * | 2007-06-21 | 2008-12-25 | Microsoft Corporation | Mining geographic knowledge using a location aware topic model |
| CN104679802A (en) * | 2013-11-28 | 2015-06-03 | 财团法人资讯工业策进会 | Travel planning device and travel planning method |
| CN105023205A (en) * | 2014-04-16 | 2015-11-04 | 珀斯特传媒有限公司 | Tour service system for providing recommended tour course |
| CN104463346A (en) * | 2014-11-04 | 2015-03-25 | 黄超超 | Travel information organizing device and method with hierarchical route map as framework |
Non-Patent Citations (1)
| Title |
|---|
| 刘淇: "基于用户兴趣建模的推荐方法及应用研究", 《中国博士学位论文全文数据库信息科技辑》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106775883A (en) * | 2016-12-26 | 2017-05-31 | 上海携程国际旅行社有限公司 | For the dynamic packaging method and packaging system of tourist resources |
| CN107145955A (en) * | 2017-04-11 | 2017-09-08 | 浙江大学 | A kind of stroke planning method based on condition and user preference |
| US11276398B1 (en) | 2017-08-07 | 2022-03-15 | Soundhound, Inc. | System and method for providing natural language recommendations |
| CN110147490A (en) * | 2017-08-07 | 2019-08-20 | 声音猎手公司 | Natural language recommends feedback |
| CN107977883A (en) * | 2017-11-24 | 2018-05-01 | 清华大学 | The recommendation method, apparatus and computer equipment of traveling bag |
| CN109839120A (en) * | 2017-11-24 | 2019-06-04 | 北京三快在线科技有限公司 | Stroke planning method, device, medium and electronic equipment |
| WO2019191868A1 (en) * | 2018-04-02 | 2019-10-10 | 徐璇炫 | Hotel-oriented conference planning method and conference management system |
| CN108829852A (en) * | 2018-06-21 | 2018-11-16 | 桂林电子科技大学 | A kind of individualized travel route recommended method |
| CN110895554A (en) * | 2018-09-12 | 2020-03-20 | 清华大学 | Destination recommendation method and device, computer equipment and storage medium |
| CN109146324A (en) * | 2018-09-13 | 2019-01-04 | 北京三快在线科技有限公司 | Recommended method, device and electronic equipment |
| CN109146324B (en) * | 2018-09-13 | 2021-02-02 | 北京三快在线科技有限公司 | Recommendation method and device and electronic equipment |
| CN109992245A (en) * | 2019-04-11 | 2019-07-09 | 河南师范大学 | A kind of method and system carrying out the modeling of science and technology in enterprise demand for services based on topic model |
| CN110457466A (en) * | 2019-06-28 | 2019-11-15 | 谭浩 | Generate method, computer readable storage medium and the terminal device of interview report |
| CN111143680A (en) * | 2019-12-27 | 2020-05-12 | 上海携程商务有限公司 | Method and system for recommending route, electronic device and computer storage medium |
| CN111143680B (en) * | 2019-12-27 | 2024-03-26 | 上海携程商务有限公司 | Route recommendation method, system, electronic equipment and computer storage medium |
| CN111210307A (en) * | 2020-01-03 | 2020-05-29 | 武汉理工大学 | Scientific and technological service chain intelligent recommendation system and method with response user preference as core |
| CN113435975A (en) * | 2021-06-29 | 2021-09-24 | 平安科技（深圳）有限公司 | Wheelchair leasing processing method and device and related equipment |
| CN113435975B (en) * | 2021-06-29 | 2023-09-26 | 平安科技（深圳）有限公司 | Wheelchair leasing processing method and device and related equipment |
| CN115471188A (en) * | 2022-09-05 | 2022-12-13 | 漱玉平民大药房连锁股份有限公司 | Identification method and equipment for supply requirements of participant service ticket |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105955981B (en) | 2019-04-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105955981B (en) | A kind of personalized traveling bag recommended method based on demand classification and subject analysis | |
| Zhang et al. | An online reviews-driven method for the prioritization of improvements in hotel services | |
| Chang et al. | Predicting aspect-based sentiment using deep learning and information visualization: The impact of COVID-19 on the airline industry | |
| CN108073569A (en) | A kind of law cognitive approach, device and medium based on multi-layer various dimensions semantic understanding | |
| CN110457442A (en) | The knowledge mapping construction method of smart grid-oriented customer service question and answer | |
| CN108710663A (en) | A kind of data matching method and system based on ontology model | |
| CN109447266A (en) | A kind of agricultural science and technology service intelligent sorting method based on big data | |
| Prameswari et al. | Mining online reviews in Indonesia's priority tourist destinations using sentiment analysis and text summarization approach | |
| Xu et al. | Where to go and what to play: Towards summarizing popular information from massive tourism blogs | |
| CN103838857A (en) | Automatic service combination system and method based on semantics | |
| CN111914157A (en) | Travel scheme generation method and system based on user preference | |
| CN114997288A (en) | Design resource association method | |
| Jalan et al. | Context-aware hotel recommendation system based on hybrid approach to mitigate cold-start-problem | |
| CN109063171A (en) | Semantic-based reso urce matching method | |
| CN110781300A (en) | Tourism resource culture characteristic scoring algorithm based on Baidu encyclopedia knowledge graph | |
| Gim et al. | A trend analysis method for IoT technologies using patent dataset with goal and approach concepts | |
| Yoo et al. | An ontology-based co-creation enhancing system for idea recommendation in an online community | |
| Renganathan et al. | Dubai restaurants: A sentiment analysis of tourist reviews | |
| Wang et al. | Sentiment Classification of Educational Tourism Reviews Based on Parallel CNN and LSTM with Attention Mechanism | |
| Kumar et al. | Combined sentiment score and star rating analysis of travel destination prediction based on user preference using morphological linear neural network model with correlated topic modelling approach | |
| Sun et al. | Urban region function mining service based on social media text analysis | |
| Nayak et al. | Application of machine learning to improve tourism industry | |
| CN113268607A (en) | Knowledge graph construction method and device | |
| Khruahong et al. | Ontology design for Thailand travel industry | |
| Cunningham et al. | Foodgpt: Amachine learning approach to ingredient substitution and recipe recommendation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |