CN110443314A - Scenic spot passenger flow forecast method and device based on machine learning - Google Patents
Scenic spot passenger flow forecast method and device based on machine learning Download PDFInfo
- Publication number
- CN110443314A CN110443314A CN201910728598.9A CN201910728598A CN110443314A CN 110443314 A CN110443314 A CN 110443314A CN 201910728598 A CN201910728598 A CN 201910728598A CN 110443314 A CN110443314 A CN 110443314A
- Authority
- CN
- China
- Prior art keywords
- passenger flow
- flow forecast
- scenic spot
- data
- machine learning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000010801 machine learning Methods 0.000 title claims abstract description 98
- 238000000034 method Methods 0.000 title claims abstract description 88
- 239000013598 vector Substances 0.000 claims abstract description 94
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 77
- 230000000694 effects Effects 0.000 claims abstract description 37
- 238000012549 training Methods 0.000 claims description 43
- 230000015654 memory Effects 0.000 claims description 23
- 238000004364 calculation method Methods 0.000 claims description 19
- 238000003860 storage Methods 0.000 claims description 19
- 238000009826 distribution Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 13
- 238000012216 screening Methods 0.000 claims description 12
- 238000011156 evaluation Methods 0.000 claims description 11
- 230000008859 change Effects 0.000 claims description 8
- 230000004907 flux Effects 0.000 claims description 6
- 230000007812 deficiency Effects 0.000 abstract description 2
- 238000005259 measurement Methods 0.000 description 12
- 238000007726 management method Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 238000012545 processing Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 4
- 230000001052 transient effect Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000007418 data mining Methods 0.000 description 3
- 238000003066 decision tree Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012417 linear regression Methods 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 238000010200 validation analysis Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013501 data transformation Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Marketing (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The scenic spot passenger flow forecast method and device based on machine learning that the embodiment of the invention discloses a kind of, this method comprises: generating the history passenger flow characteristic at scenic spot according to the scenic spot history pipelined data of acquisition and scenic spot characteristic data；The corresponding passenger flow forecast model of each machine learning algorithm is respectively trained out according to the history passenger flow characteristic and preset a variety of machine learning algorithms；The prediction effect of the corresponding passenger flow forecast model of each machine learning algorithm is assessed, and the preferable passenger flow forecast model of multiple prediction effects is selected from the corresponding passenger flow forecast model of each machine learning algorithm according to assessment result；It determines the weight vectors for the passenger flow forecast model selected, and carries out scenic spot passenger flow forecast according to the weight vectors and the passenger flow forecast model selected.The present invention solves the technical problem of the scenic spot passenger flow estimation accuracy deficiency of the prior art.
Description
Technical field
The present invention relates to artificial intelligence fields, in particular to a kind of scenic spot passenger flow forecast based on machine learning
Method and device.
Background technique
The product orientation of scenic spot project and integrated planning have a far-reaching influence to the development of tourism industry, and tourist flows amount
Prediction work is to formulate the important link of Large Tourism Developmant Strategy planning, and the person that is scenic spot management is in policies and daily management mission
In the significant data that wants to know about.Scenic spot management person by the analysis to tourism demand tendency, using certain method and because
Element estimates scenic spot tourist's quantity, is scenic spot flow control, sale marketing, personnel placement, traffic dispersion, safety management etc.
The arrangement of work provides decision-making foundation.Traditional passenger flow estimation, the person's artificial experience that usually relies on scenic spot management estimate anticipation visitor
Peak ranges are flowed, since practical volume of the flow of passengers influence factor is more and association complexity is high, by scenic spot traffic location, periphery thing
The interference of many factors such as part, season, festivals or holidays, weather, precisely judges that the volume of the flow of passengers is big compared with difficult and error by artificial experience.
The present invention at least one of to solve the above-mentioned problems, proposes a kind of scenic spot volume of the flow of passengers based on machine learning
Prediction technique and device.
Summary of the invention
The scenic spot passenger flow forecast method and device based on machine learning that the main purpose of the present invention is to provide a kind of, with
Solve the technical problem of the scenic spot passenger flow estimation accuracy deficiency of the prior art.
To achieve the goals above, according to an aspect of the invention, there is provided a kind of scenic spot visitor based on machine learning
Method for predicting, this method comprises:
The history passenger flow characteristic at scenic spot is generated according to the scenic spot history pipelined data of acquisition and scenic spot characteristic data；
The scenic spot characteristic data include: festivals or holidays data, scenic spot position data, scenic spot weather data, scenic spot dull and rush season data, scenic spot
At least one of ticket price data；
Each engineering is respectively trained out according to the history passenger flow characteristic and preset a variety of machine learning algorithms
Practise the corresponding passenger flow forecast model of algorithm；
The prediction effect of the corresponding passenger flow forecast model of each machine learning algorithm is assessed, and according to assessment result
The preferable passenger flow forecast model of multiple prediction effects is selected from the corresponding passenger flow forecast model of each machine learning algorithm；
Determine the weight vectors for the passenger flow forecast model selected, and according to the weight vectors and the visitor selected
Flux prediction model carries out scenic spot passenger flow forecast.
Optionally, the prediction effect to the corresponding passenger flow forecast model of each machine learning algorithm is assessed, tool
Body includes:
Verify data set is generated according to the history passenger flow characteristic；
Each verify data in the verify data set is separately input to the corresponding passenger flow of each machine learning algorithm
Prediction model is measured, prediction result is obtained；
Each passenger flow forecast mould is calculated according to the corresponding prediction result of each passenger flow forecast model and RMSE calculation method
The corresponding RMSE value of type.
Optionally, it is described selected from the corresponding passenger flow forecast model of each machine learning algorithm according to assessment result it is more
A preferable passenger flow forecast model of prediction effect, specifically includes:
The selection of passenger flow forecast model is carried out according to the corresponding RMSE value of each passenger flow forecast model.
Optionally, the weight vectors for the passenger flow forecast model that the determination is selected, specifically include:
Weight vectors set is generated according to the quantity for the passenger flow forecast model selected；
By the prediction knot with the passenger flow forecast model selected respectively of each weight vectors in the weight vectors set
Fruit is weighted and averaged calculating, obtains the corresponding result of weighted average of each weight vectors；
The RMSE value of each result of weighted average is calculated using RMSE calculation method, and determines the corresponding power of the smallest RMSE value
Weight vector.
Optionally, the quantity for the passenger flow forecast model that the basis is selected generates weight vectors set, specifically includes:
Determine weight change granularity 1/H, and according to weight change granularity generate set M, wherein M=1,1,1,1 ...,
1 }, altogether comprising H 1 in set M；
The use of plug hole method by H 1 point in set M is n group, sharesThe kind method of salary distribution, respectively every kind of distribution side
The 1 of every group adds up in formula, and divided by H, is sharedThe weight vectors set of a equally distributed weight vectors,
In, n is the quantity for the passenger flow forecast model selected.
Optionally, described to be respectively trained according to the history passenger flow characteristic and preset a variety of machine learning algorithms
The corresponding passenger flow forecast model of each machine learning algorithm out, specifically includes:
Machine learning algorithm pair is used according to the parameter value range of the machine learning algorithm of setting and value change step
The hyper parameter searching algorithm answered dynamically finds the hyper parameter for keeping forecast result of model optimal, so that the passenger flow forecast mould trained
The prediction effect global optimum of type or local optimum.
Optionally, described pre- according to the weight vectors and the passenger flow forecast model the selected progress scenic spot volume of the flow of passengers
It surveys, comprising:
Obtain current passenger flow characteristic；
The current passenger flow characteristic is separately input to each passenger flow forecast model selected, obtains each volume of the flow of passengers
The passenger flow forecast data of prediction model output；
Go out comprehensive passenger flow forecast data according to the passenger flow forecast data and the weight vector computation.
Optionally, this method further include:
Periodically calculate the comprehensive passenger flow forecast data and reality corresponding with the comprehensive passenger flow forecast data
Deviation amplitude between the volume of the flow of passengers data of border；
Re -training model signals are sent when the deviation amplitude is greater than preset value.
To achieve the goals above, according to another aspect of the present invention, a kind of scenic spot visitor based on machine learning is provided
Volume forecasting device, the device include:
History passenger flow characteristic acquiring unit, for the scenic spot history pipelined data and scenic spot characteristic number according to acquisition
According to the history passenger flow characteristic for generating scenic spot；The scenic spot characteristic data include: festivals or holidays data, scenic spot position data, scape
At least one of area's weather data, scenic spot dull and rush season data, entrance ticket price data；
Passenger flow forecast model training unit, for according to the history passenger flow characteristic and preset a variety of machines
The corresponding passenger flow forecast model of each machine learning algorithm is respectively trained out in learning algorithm；
Model evaluation screening unit, for the prediction effect to the corresponding passenger flow forecast model of each machine learning algorithm into
Row assessment, and multiple prediction effects are selected from the corresponding passenger flow forecast model of each machine learning algorithm according to assessment result
Preferable passenger flow forecast model；
Passenger flow forecast unit, for determining the weight vectors for the passenger flow forecast model selected, and according to the power
Weight vector and the passenger flow forecast model selected carry out scenic spot passenger flow forecast.
Optionally, the model evaluation screening unit, comprising:
Verify data generation module, for generating verify data set according to the history passenger flow characteristic；
Model prediction result obtains module, for each verify data in the verify data set to be separately input to
The corresponding passenger flow forecast model of each machine learning algorithm, obtains prediction result；
RMSE value computing module, for according to the corresponding prediction result of each passenger flow forecast model and RMSE calculation method
Calculate the corresponding RMSE value of each passenger flow forecast model.
Optionally, the model evaluation screening unit is specifically used for according to the corresponding RMSE value of each passenger flow forecast model
Carry out the selection of passenger flow forecast model.
Optionally, the passenger flow forecast unit, comprising:
Weight vectors set generation module, for generating weight vectors according to the quantity for the passenger flow forecast model selected
Set；
Weighted calculation module, for by each weight vectors in the weight vectors set respectively with the volume of the flow of passengers selected
The prediction result of prediction model is weighted and averaged calculating, obtains the corresponding result of weighted average of each weight vectors；
Optimal weights vector selecting module, for calculating the RMSE value of each result of weighted average using RMSE calculation method,
And determine the corresponding weight vectors of the smallest RMSE value.
Optionally, the weight vectors set generation module, comprising:
Weight variation granularity determines submodule, for determining that weight changes granularity 1/H, and changes granularity according to weight and generates
Set M, wherein altogether comprising H 1 in M={ 1,1,1,1 ..., 1 }, set M；
Plug hole method handles submodule, by H 1 point in set M is n group for using plug hole method, sharedKind distribution
Mode is added up the 1 of every group in every kind of method of salary distribution respectively, and divided by H, is sharedA equally distributed weight to
The weight vectors set of amount, wherein n is the quantity for the passenger flow forecast model selected.
Optionally, the passenger flow forecast model training unit, comprising:
Hyper parameter tuning module, for the parameter value range and value change step according to the machine learning algorithm of setting
The hyper parameter for keeping forecast result of model optimal is dynamically found using the corresponding hyper parameter searching algorithm of machine learning algorithm, so that instruction
The prediction effect global optimum for the passenger flow forecast model practised or local optimum.
Optionally, the passenger flow forecast unit, comprising:
Current passenger flow characteristic obtains module, for obtaining current passenger flow characteristic；
Model prediction module, for the current passenger flow characteristic to be separately input to each passenger flow forecast selected
Model obtains the passenger flow forecast data of each passenger flow forecast model output；
Integrated forecasting data computation module, for being gone out according to the passenger flow forecast data and the weight vector computation
Comprehensive passenger flow forecast data.
Optionally, the device further include:
Deviation amplitude computing unit, for periodically calculate the comprehensive passenger flow forecast data and with the comprehensive visitor
Deviation amplitude between the corresponding practical volume of the flow of passengers data of volume forecasting data；
Model re -training unit, for sending re -training model signals when the deviation amplitude is greater than preset value.
To achieve the goals above, according to another aspect of the present invention, a kind of computer equipment, including storage are additionally provided
Device, processor and storage on a memory and the computer program that can run on a processor, the processor execution meter
The step in the above-mentioned scenic spot passenger flow forecast method based on machine learning is realized when calculation machine program.
To achieve the goals above, according to another aspect of the present invention, a kind of computer readable storage medium is additionally provided,
The computer-readable recording medium storage has computer program, real when the computer program executes in the computer processor
Step in the existing above-mentioned scenic spot passenger flow forecast method based on machine learning.
The invention has the benefit that the embodiment of the present invention passes through to scenic spot history pipelined data and scenic spot characteristic data
Etc. data be processed and model training, train scenic spot passenger flow forecast model, and then accurately to the scenic spot volume of the flow of passengers
It is predicted, effectively raises the accuracy of scenic spot passenger flow forecast.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is the present invention
Some embodiments for those of ordinary skill in the art without creative efforts, can also basis
These attached drawings obtain other attached drawings.In the accompanying drawings:
Fig. 1 is the flow chart of scenic spot passenger flow forecast method of the embodiment of the present invention based on machine learning；
Fig. 2 is the flow chart that the embodiment of the present invention is assessed Passenger flow forecast model and screened；
Fig. 3 is the flow chart that the embodiment of the present invention determines optimal weights vector；
Fig. 4 is the flow chart that the embodiment of the present invention calculates comprehensive passenger flow forecast data；
Fig. 5 is the flow chart of model re -training of embodiment of the present invention judgement；
Fig. 6 is the first structure block diagram of scenic spot passenger flow forecast device of the embodiment of the present invention based on machine learning；
Fig. 7 is the structure chart of model evaluation screening unit of the embodiment of the present invention；
Fig. 8 is the first structure figure of passenger flow forecast unit of the embodiment of the present invention；
Fig. 9 is the structure chart of weight vectors set generation module of the embodiment of the present invention；
Figure 10 is the structure chart of passenger flow forecast model training unit of the embodiment of the present invention；
Figure 11 is the second structure chart of passenger flow forecast unit of the embodiment of the present invention；
Figure 12 is the second structural block diagram of scenic spot passenger flow forecast device of the embodiment of the present invention based on machine learning；
Figure 13 is computer equipment schematic diagram of the embodiment of the present invention.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
The model that the present invention protects all should belong in member's every other embodiment obtained without making creative work
It encloses.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the present invention, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
It should be noted that term " includes " and " tool in description and claims of this specification and above-mentioned attached drawing
Have " and their any deformation, it is intended that cover it is non-exclusive include, for example, containing a series of steps or units
Process, method, system, product or equipment those of are not necessarily limited to be clearly listed step or unit, but may include without clear
Other step or units listing to Chu or intrinsic for these process, methods, product or equipment.
It should be noted that in the absence of conflict, the feature in embodiment and embodiment in the present invention can phase
Mutually combination.The present invention will be described in detail below with reference to the accompanying drawings and embodiments.
The present invention provides a kind of scenic spot passenger flow forecasting based on machine learning.Acquire the historical trading flowing water at sight spot
Data, the characteristics data such as acquisition legal festivals and holidays, weather conditions, locating geographical location, dull and rush season period, integrated use engineering
The technologies such as data prediction, Feature Engineering, machine learning algorithm carry out processing and line modeling to collected data in habit, real
The tourist's quantity for having showed multiple measurement periods such as one day, one week, one month future of intelligent predicting scenic spot is supplied to business personnel's work
For the foundation for estimating scenic spot passenger flow, helps scenic spot science to realize that administrative staff work and distribute, promote scenic spot organization work management effect
Rate, while reducing business personnel's work load.
Fig. 1 is the first pass figure of scenic spot passenger flow forecast method of the embodiment of the present invention based on machine learning, such as Fig. 1 institute
Show, the scenic spot passenger flow forecast method based on machine learning of the present embodiment includes step S101 to step S104.
Step S101 generates the history passenger flow at scenic spot according to the scenic spot history pipelined data of acquisition and scenic spot characteristic data
Characteristic；The scenic spot characteristic data include: festivals or holidays data, scenic spot position data, scenic spot weather data, scenic spot dull and rush season
At least one of data, entrance ticket price data.
In embodiments of the present invention, the passenger flow business flowing water at corresponding client scenic spot can be obtained by operation system in bank
Data, every daily increment Flow Record are simultaneously stored in database.Further, it is also possible to pass through operation system or the acquisition of online crawler and scape
The relevant characteristic data in area specifically include that national legal festivals and holidays data, scenic spot geographic position data, scenic spot location day
Gas information (including wind scale, temperature, rank of raining etc.), tourism dull and rush season, ticket price etc..And then to the scenic spot of acquisition
The initial data such as history pipelined data and characteristic data are handled, processing mode specifically include that data cleansing, data integration,
Data regularization and data transformation.After data prediction, the high quality for having accuracy, integrality and consistency can be obtained
Data.
In turn, feature construction is carried out to the scenic spot history pipelined data and scenic spot characteristic data pre-processed and feature mentions
It takes, obtains the history passenger flow characteristic at scenic spot.The detailed stream of feature construction and feature extraction is carried out in embodiments of the present invention
Cheng Kejian step 11 is to step 44.
In embodiments of the present invention, the history passenger flow characteristic at scenic spot obtained above can for the scenic spot at one or
Passenger flow characteristic in multiple measurement periods, next, can be according to trained passenger flow forecast model to one following
The passenger flow number of measurement period is predicted.
Step S102 is respectively trained out according to the history passenger flow characteristic and preset a variety of machine learning algorithms
The corresponding passenger flow forecast model of each machine learning algorithm.
In embodiments of the present invention, machine learning algorithm candidate when model training includes but is not limited to: support vector machines
(Support Vector Machine, SVM) regression algorithm, k nearest neighbor (K-Nearest Neighbor, KNN) regression algorithm, with
Machine forest (RandomForest) regression algorithm, GBDT (Gradient Boosting Decision Tree) regression algorithm,
Xgboost regression algorithm, shot and long term memory network (Long Short-Term Memory, LSTM) algorithm.Implement in the present invention
In example, when carrying out model training, need for history passenger flow characteristic to be processed into the data of different machines learning algorithm requirement
Structure.
In embodiments of the present invention, the specific steps of training passenger flow forecast model may comprise steps of.
Firstly, history passenger flow characteristic to be processed into the input lattice of candidate a variety of machine learning algorithm requirements respectively
Formula is divided into training dataset and validation data set, and for the verifying of subsequent training and model, wherein training dataset accounts for characteristic
According to 90%, validation data set accounts for the 10% of characteristic, and training dataset and validation data set are not overlapped.
In addition, when carrying out the model training of each machine learning algorithm according to training sample set, according in systems to time
The preset parameter value range of machine learning algorithm and value change step are selected, dynamically to find keeps forecast result of model optimal
Hyper parameter during training pattern, will close until finding the model of prediction effect global optimum or local optimum
Whether note there is the case where over-fitting.The hyper parameter searching algorithm that candidate algorithm uses can be with are as follows: elastomeric network and decision tree return
Reduction method, linear regression use the mode of grid search (GridSearchCV), naive Bayesian, Light-GBM, xgboost
Regression algorithm uses the mode of random search (RandomizedSearchCV)；LSTM algorithm is neural network algorithm, by anti-
Carry out training pattern to the mode of propagation.The corresponding passenger flow forecast model of each machine learning algorithm is trained with this.
Step S103 assesses the prediction effect of the corresponding passenger flow forecast model of each machine learning algorithm, and root
The preferable passenger flow of multiple prediction effects is selected from the corresponding passenger flow forecast model of each machine learning algorithm according to assessment result
Measure prediction model.
In embodiments of the present invention, the prediction effect of passenger flow forecast model can be assessed using a variety of methods,
Such as calculating can be compared using the predicted value for exporting model and history actual value and is deviated to carry out prediction effect and comment
The methods of estimate.
Step S104, determines the weight vectors of passenger flow forecast model selected, and according to the weight vectors and
The passenger flow forecast model selected carries out scenic spot passenger flow forecast.
In embodiments of the present invention, after determining multiple passenger flow forecast models of practical application according to assessment result,
It include the weight of these passenger flow forecast models also it needs to be determined that the weight vectors of these passenger flow forecast models, in weight vectors
Than.When carrying out scenic spot passenger flow forecast, the predicted value that each passenger flow forecast model exports is weighted according to weight vectors
Average computation obtains comprehensive passenger flow forecast data, as final predicted value.
In embodiments of the present invention, determine the weight vectors of these passenger flow forecast models method can there are many, example
Weighted value is such as set according to above-mentioned model evaluation result, the weight ratio setting of the better flux prediction model of assessment result is more
It is high.
The embodiment of the present invention passes through to scenic spot history pipelined data and scenic spot characteristic data it can be seen from above description
Etc. data be processed and model training, train scenic spot passenger flow forecast model, effectively raise the scenic spot volume of the flow of passengers
The accuracy of prediction.
In embodiments of the present invention, above-mentioned steps S101 obtains the detailed process of the history passenger flow characteristic at scenic spot substantially
Two steps can be divided into: 1, scenic spot history pipelined data and characteristic data being pre-processed；2, pass through feature construction and spy
Sign is extracted and excavates history passenger flow characteristic.
Firstly, in embodiments of the present invention, carrying out pretreated specific side to scenic spot history pipelined data and characteristic data
Method may comprise steps of:
Step 1, by the nearest passenger flow business historical record in the corresponding client scenic spot of operation system acquisition in bank and daily
Increment Flow Record is stored in database；
Step 2, characteristic data relevant to scenic spot are obtained by banking system or online crawler, specifically included that
National legal festivals and holidays data, scenic spot geographic position data, scenic spot location Weather information (including wind scale, temperature, under
Rain rank etc.), tourism dull and rush season, ticket price etc.；
Step 3, obtained scenic spot history pipelined data and characteristic data are assessed；
Step 4, obtained scenic spot history pipelined data and characteristic data are cleared up, passes through filling missing values, smooth
The modes such as noise and identification outlier, are corrected inconsistent in data.When filling missing values, mean value, median is taken to fill out
It fills；When handling noise data and outlier, noise data is determined by the lower edges of box traction substation first, is calculated by cluster
Method detects outlier, then in conjunction with specific business experience marks noise data and outlier, finally by calculating front and back four
Zhou Tongzhou several average value repairs noise data and outlier；
Step 5, obtained historical data and characteristic data are integrated, since the data source that system obtains is various,
Might have different name or unit in different data sources in the presence of the attribute for representing identical concept, will lead to inconsistency and
Redundancy in the invention integrates data using the method for correlation analysis；
Step 6, reduction is carried out to data, device, which obtains simplifying for data by reduction techniques, to be indicated, simplified data
Occupied space can become smaller, but can generate almost it is identical analysis as a result, it is possible to increase whole system efficiency；
Step 7, data are converted, is converted by data so that data are more suitable for system and carry out data mining.Than
Such as to the transformation of geographical location information, geographical location information is classified, same category uses the same digital representation, in this way
Text data is just for conversion into discrete numeric data.
So far, system has been completed the pretreatment to a scenic spot passenger flow data, has been substantially achieved and has had accuracy, complete
Whole property, consistency, the quality data suitable for data mining.
In embodiments of the present invention, the specific side of history passenger flow characteristic is excavated by feature construction and feature extraction
Method may comprise steps of:
Step 11, by analyzing pretreated scenic spot history pipelined data in figures such as line chart, scatter plot, histograms
Characteristic distributions classify to scenic spot, specifically, scenic spot type is first divided into natural landscape class scape according to its tourism resource development value
Area and places of cultural interest class scenic spot, then the daily tourist's amplitude of variation in scenic spot is analyzed, and scenic spot is segmented are as follows: Wave type and stable type；
Step 22, data are converted for administrative staff's business experience by certain rule, for subsequent characteristics building and spy
Sign is extracted and excavates new feature；
Step 33, go out characteristic as much as possible in conjunction with the classification of above-mentioned scenic spot and administrative staff's business experience data mining
According in embodiments of the present invention, the characteristic excavated can be divided into four classes: public attribute feature, is spread out at characteristic attributive character
Raw data characteristics and scenic spot service feature；Wherein, public attribute feature may include: festivals or holidays (whether festivals or holidays, festivals or holidays day
Number, festivals or holidays distance, working day mark etc.), it is weather (same day rain grade, preceding N days rain grade, rear N days rain grade, same day wind-force etc.), all
Phase characteristic (beginning of the month/in/end, what day, current year in which week, date in month etc.)；Characteristic attributive character may include: Characteristic activities
Day (same day whether, first N days whether, latter N days whether etc.), specific type (special holidays, school are had a holiday or vacation day etc.)；Derivative data
Feature can specifically include: the same period volume of the flow of passengers (same period last week, same period last month, the same period last year), ring than passenger flow (heaven ring ratio, on
Chow ring ratio, last year ring ratio etc.), previous passenger flow (all day volumes of the flow of passengers, the day before yesterday volume of the flow of passengers, volume of the flow of passengers etc. before three days), the average volume of the flow of passengers
(last month, two months first, n days first etc.), statistical nature (maximum value, minimum value, mean value, codomain, variance, median, the degree of bias, kurtosis
Deng)；Scenic spot service feature may include: basic business feature (scenic spot customer quantity, scenic spot client's average age, scenic spot classification,
Scenic spot dull and rush season, scenic spot working day etc.), previous transaction feature (booking number of transaction under individual traveler's line, booking number of deals on individual traveler's line
Amount, OTA booking platform/travel community passenger flow number etc.)；
Step 44, history passenger flow characteristic is generated in above-mentioned steps 33 to each measurement period in each scenic spot and carries out dynamic
Selection is picked out and allows the preferable character subset of scenic spot passenger flow estimation effect, therefore corresponding to different scenic spots, different measurement periods
Character subset be possible to be different.The character subset picked out is that the subsequent history passenger flow for carrying out model training is special
Levy data.
So far, system, which has been completed, generates history passenger flow characteristic to a measurement period at a scenic spot, connects down
It will be using these characteristics come training pattern.
Fig. 2 is the flow chart that the embodiment of the present invention is assessed Passenger flow forecast model and screened, as shown in Fig. 2, at this
In inventive embodiments, the method that the prediction effect of passenger flow forecast model is assessed and screened of above-mentioned steps S103
Step S201 be can specifically include to step S204.
Step S201 generates verify data set according to the history passenger flow characteristic.
Each verify data in the verify data set is separately input to each machine learning algorithm pair by step S202
The passenger flow forecast model answered, obtains prediction result.
Step S203 calculates each passenger flow according to the corresponding prediction result of each passenger flow forecast model and RMSE calculation method
Measure the corresponding RMSE value of prediction model.
Step S204 carries out the selection of passenger flow forecast model according to the corresponding RMSE value of each passenger flow forecast model.
In embodiments of the present invention, root-mean-square error i.e. RMSE method can be used to assess modelling effect, count
The method for calculating the RMSE value of model can be divided into two steps, calculate it in institute to each sample data in verify data set first
There is the prediction result on model, and saves the prediction result in the database；Then, when the prediction result meter of all verify datas
After calculation, the RMSE value on each model, the prediction effect of the smaller representative model of RMSE value are asked according to the formula of RMSE respectively
It is better, more stable.
In embodiments of the present invention, one, scenic spot system can be selected according to the RMSE value of each passenger flow forecast model
Count the best several models of the prediction effect in period, if all model RMSE value difference are little, all conducts of all models
Available model；If model RMSE value difference it is larger, to RMSE value according to from small to large sort after, select top n mould
Type can be set according to actual scene by engineer as available model, the value of N.
Fig. 3 is the flow chart that the embodiment of the present invention determines optimal weights vector, as shown in figure 3, in embodiments of the present invention,
The weight vectors for the passenger flow forecast model that the determination of above-mentioned steps S104 is selected can specifically include step S301 to step
S303。
Step S301 generates weight vectors set according to the quantity for the passenger flow forecast model selected.
Step S302, by each weight vectors in the weight vectors set respectively with the passenger flow forecast model selected
Prediction result be weighted and averaged calculating, obtain the corresponding result of weighted average of each weight vectors.
Step S303 is calculated the RMSE value of each result of weighted average using RMSE calculation method, and determines the smallest RMSE
It is worth corresponding weight vectors.
In embodiments of the present invention, each of weight vectors set weight vectors are pre- to the history of each model respectively
Result is surveyed to be weighted and averaged；Finally, using RMSE method under each weight vectors history be weighted and averaged prediction result into
Row evaluation, the smallest RMSE value is required optimal weights vector.For example, at the large-scale 4A grades of scenic spots that certain is saved, process is above-mentioned
Step selects the preferable decision tree recurrence of effect, Light-GBM, linear regression model (LRM), via the optimal weights being calculated to
Amount is { 0.35,0.4,0.25 }, and using the optimal vector value as the weighing vector parameter of subsequent combination model prediction.
In an embodiment of the present invention, the specific method of the generation weight vectors set of above-mentioned steps S301 can be with are as follows: really
Determine weight variation granularity 1/H, and granularity is changed according to weight and generates set M, wherein in M={ 1,1,1,1 ..., 1 }, set M
Altogether comprising H 1；The use of plug hole method by H 1 point in set M is n group, sharesThe kind method of salary distribution, respectively every kind point
1 with every group in mode adds up, and divided by H, is sharedThe weight vectors set of a equally distributed weight vectors,
Wherein, n is the quantity for the passenger flow forecast model selected.
Be illustrated below: the first step gives n=2, H=3, i.e. Number of Models is 2, and it is 1/3 that weight, which changes granularity,；The
Two steps, according to the plug hole method in arrangement thought, obtain in 3 the method for salary distribution be { zero 1, three 1 }, { 11, two 1 }, { two
A 1, one 1 }, { three 1, zero 1 } }；Third step, first by every kind of method of salary distribution 1 be added obtain { 0,3 }, { 1,2 },
{ 2,1 }, { 3,0 } }, the weight vectors set obtained respectively divided by H=3 are as follows: { 0,1 }, { 1/3,2/3 }, { 2/3,1/3 }, 1,
0}}。
Fig. 4 is the flow chart that the embodiment of the present invention calculates comprehensive passenger flow forecast data, as shown in figure 4, of the invention real
It applies in example, above-mentioned steps S104's carries out the scenic spot volume of the flow of passengers according to the weight vectors and the passenger flow forecast model selected
Prediction, can specifically include step S401 value step S403.
Step S401 obtains current passenger flow characteristic.
The current passenger flow characteristic is separately input to each passenger flow forecast model selected, obtained by step S402
The passenger flow forecast data exported to each passenger flow forecast model.
Step S403 goes out comprehensive passenger flow forecast number according to the passenger flow forecast data and the weight vector computation
According to.
In embodiments of the present invention, the history passenger flow characteristic at the above-mentioned scenic spot for model training can be the scenic spot
Passenger flow characteristic in one or more measurement periods, next, can be according to trained passenger flow forecast model pair
The passenger flow number of the following measurement period is predicted.In embodiments of the present invention, above-mentioned measurement period may include day,
Week, the moon, season, year etc..
In embodiments of the present invention, in the next measurement period of passenger flow forecast model prediction trained using the present invention
When the volume of the flow of passengers, need first to obtain current passenger flow characteristic, and current passenger flow characteristic is input to trained visitor
In flux prediction model, and then obtain the synthesis passenger flow forecast data of next measurement period.
Fig. 5 is the flow chart of model re -training of embodiment of the present invention judgement, as shown in figure 5, in embodiments of the present invention,
It includes step S501 and step S502 that judgment models, which need the process of re -training,.
Step S501, periodically calculate the comprehensive passenger flow forecast data and with the comprehensive passenger flow forecast data
Deviation amplitude between corresponding practical volume of the flow of passengers data.
Step S502 sends re -training model signals when the deviation amplitude is greater than preset value.
It in an embodiment of the present invention, can be to one statistics week an of scenic spot after training a period of time distance model last time
The forecast result of model of phase is evaluated and is fed back, evaluate the process flow of feedback the following steps are included:
Obtain all prediction data from last time model training till now；Modelling effect is commented using RMSE method
Estimate, using the true volume of the flow of passengers data of prediction data and the scenic spot time interval as inputting parameter, calculated value and predetermined deviation
(such as deviation amplitude is set as 20%) is compared in threshold values；If calculated value is more than the preset threshold, then decision-feedback result is poor.
Conversely, then feedback result is preferable；Obtain business personnel's feedback data；When assessment result and feedback result are all difference, can feed back
Give model training apparatus one signal, in second day system flow start by set date, model training apparatus was needed to certain scenic spot
A measurement period carries out re -training.
As can be seen from the above embodiments, the embodiment of the invention provides a kind of intelligent scenic spot passenger flow forecast method, bases
In data analysis, machine learning techniques, the analysis of integrated use data and cleaning, web crawlers, algorithm study, dynamic weighting combination
Etc. means, the data such as scenic spot history passenger flow data, scenic spot business feature, external environmental factor are processed and model instruction
Practice, realizes the work such as wisdom operation and intelligent management for scenic spot, precisely provide scenic spot following one day, one week and one month tourist
Quantity estimation results effectively promote the precision and working efficiency of scenic spot customer account management.Scenic spot passenger flow forecast side of the invention
Method at least realize it is following the utility model has the advantages that
1, the aspect of model selection based on business feature: being analyzed based on business experience, business feature analysis and data characteristics,
The characteristic data set of effectively reflection scenic spot business feature, and the therefrom character subset at dynamic select scenic spot are extracted, prediction effect is made
It is optimal.
2, the multi-model weighted array calculated based on dynamic: being calculated based on changeable weight vector, realizes many algorithms model
The dynamic select and weighted array of prediction result, so that the scenic spot with different business characteristic is combined using optimal prediction algorithm
It is finally predicted, to keep model prediction stability, precision and the generalization ability at each scenic spot optimal.
It should be noted that step shown in the flowchart of the accompanying drawings can be in such as a group of computer-executable instructions
It is executed in computer system, although also, logical order is shown in flow charts, and it in some cases, can be with not
The sequence being same as herein executes shown or described step.
Based on the same inventive concept, the embodiment of the invention also provides a kind of scenic spot passenger flow forecast based on machine learning
Device can be used to implement the scenic spot passenger flow forecast method based on machine learning described in above-described embodiment, such as following
Described in embodiment.The principle solved the problems, such as due to the scenic spot passenger flow forecast device based on machine learning with based on machine learning
Scenic spot passenger flow forecast method is similar, therefore the embodiment of the scenic spot passenger flow forecast device based on machine learning may refer to base
In the embodiment of the scenic spot passenger flow forecast method of machine learning, overlaps will not be repeated.Used below, term is " single
The combination of the software and/or hardware of predetermined function may be implemented in member " or " module ".Although device described in following embodiment
It is preferably realized with software, but the realization of the combination of hardware or software and hardware is also that may and be contemplated.
Fig. 6 is the first structure block diagram of scenic spot passenger flow forecast device of the embodiment of the present invention based on machine learning, such as Fig. 6
Shown, scenic spot passenger flow forecast device of the embodiment of the present invention based on machine learning includes: that history passenger flow characteristic obtains list
Member 1, passenger flow forecast model training unit 2, model evaluation screening unit 3 and passenger flow forecast unit 4.
History passenger flow characteristic acquiring unit 1, for the scenic spot history pipelined data and scenic spot characteristic according to acquisition
The history passenger flow characteristic at data generation scenic spot；The scenic spot characteristic data include: festivals or holidays data, scenic spot position data,
At least one of scenic spot weather data, scenic spot dull and rush season data, entrance ticket price data.
Passenger flow forecast model training unit 2, for according to the history passenger flow characteristic and preset a variety of machines
The corresponding passenger flow forecast model of each machine learning algorithm is respectively trained out in device learning algorithm.
Model evaluation screening unit 3, for the prediction effect to the corresponding passenger flow forecast model of each machine learning algorithm
It is assessed, and selects multiple prediction effects from the corresponding passenger flow forecast model of each machine learning algorithm according to assessment result
The preferable passenger flow forecast model of fruit.
Passenger flow forecast unit 4, for determining the weight vectors for the passenger flow forecast model selected, and according to the power
Weight vector and the passenger flow forecast model selected carry out scenic spot passenger flow forecast.
Fig. 7 is the structure chart of model evaluation screening unit of the embodiment of the present invention, as shown in fig. 7, the mould of the embodiment of the present invention
Type assessment screening unit 3 includes: verify data generation module 301, model prediction result obtains module 302 and RMSE value calculates mould
Block 303.
Verify data generation module 301, for generating verify data set according to the history passenger flow characteristic.
Model prediction result obtains module 302, for each verify data difference in the verify data set is defeated
Enter passenger flow forecast model corresponding to each machine learning algorithm, obtains prediction result.
RMSE value computing module 303, for being calculated according to the corresponding prediction result of each passenger flow forecast model and RMSE
Method calculates the corresponding RMSE value of each passenger flow forecast model.
In embodiments of the present invention, model evaluation screening unit 3 can be according to the corresponding RMSE of each passenger flow forecast model
Value carries out the selection of passenger flow forecast model.
Fig. 8 is the first structure figure of passenger flow forecast unit of the embodiment of the present invention, as shown in figure 8, in implementation of the invention
In example, passenger flow forecast unit 4 include: weight vectors set generation module 401, weighted calculation module 402 and optimal weights to
Measure selecting module 403.
Weight vectors set generation module 401, for generating weight according to the quantity for the passenger flow forecast model selected
Vector set.
Weighted calculation module 402, for by each weight vectors in the weight vectors set respectively with the visitor that selects
The prediction result of flux prediction model is weighted and averaged calculating, obtains the corresponding result of weighted average of each weight vectors.
Optimal weights vector selecting module 403, for calculating the RMSE of each result of weighted average using RMSE calculation method
Value, and determine the corresponding weight vectors of the smallest RMSE value.
Fig. 9 is the structure chart of weight vectors set generation module of the embodiment of the present invention, as shown in figure 9, in reality of the invention
It applies in example, weight vectors set generation module 401 includes: that weight variation granularity determines submodule 4011 and plug hole method processing submodule
Block 4012.
Weight variation granularity determines submodule 4011, for determining that weight changes granularity 1/H, and changes granularity according to weight
Generate set M, wherein altogether comprising H 1 in M={ 1,1,1,1 ..., 1 }, set M.
Plug hole method handles submodule 4012, by H 1 point in set M is n group for using plug hole method, sharedKind
The method of salary distribution is added up the 1 of every group in every kind of method of salary distribution respectively, and divided by H, is sharedA equally distributed power
The weight vectors set of weight vector, wherein n is the quantity for the passenger flow forecast model selected.
Figure 10 is the structure chart of passenger flow forecast model training unit of the embodiment of the present invention, in embodiments of the present invention, visitor
Flux prediction model training unit 2 includes: hyper parameter tuning module 201.
Hyper parameter tuning module 201, for being changed according to the parameter value range and value of the machine learning algorithm of setting
Step-length dynamically finds the hyper parameter for keeping forecast result of model optimal using the corresponding hyper parameter searching algorithm of machine learning algorithm, with
Make prediction effect global optimum or the local optimum of the passenger flow forecast model trained.
Figure 11 is the second structure chart of passenger flow forecast unit of the embodiment of the present invention, as shown in figure 11, in reality of the invention
It applies in example, passenger flow forecast unit 4 further include: current passenger flow characteristic obtains module 404, model prediction module 405 and comprehensive
Close prediction data computing module 406.
Current passenger flow characteristic obtains module 404, for obtaining current passenger flow characteristic.
Model prediction module 405, for the current passenger flow characteristic to be separately input to each volume of the flow of passengers selected
Prediction model obtains the passenger flow forecast data of each passenger flow forecast model output.
Integrated forecasting data computation module 406, based on according to the passenger flow forecast data and the weight vectors
Calculate comprehensive passenger flow forecast data.
Figure 12 is the second structural block diagram of scenic spot passenger flow forecast device of the embodiment of the present invention based on machine learning, is such as schemed
Shown in 12, the scenic spot passenger flow forecast device based on machine learning of the embodiment of the present invention further include: deviation amplitude computing unit 5
With model re -training unit 6, wherein deviation amplitude computing unit 5 is connect with passenger flow forecast unit 4.
Deviation amplitude computing unit 5, for periodically calculate the comprehensive passenger flow forecast data and with the synthesis
Deviation amplitude between the corresponding practical volume of the flow of passengers data of passenger flow forecast data.
Model re -training unit 6, for sending re -training model signals when the deviation amplitude is greater than preset value.
To achieve the goals above, according to the another aspect of the application, a kind of computer equipment is additionally provided.Such as Figure 13 institute
Show, which includes memory, processor, communication interface and communication bus, and being stored with can locate on a memory
The computer program run on reason device, the processor realize the step in above-described embodiment method when executing the computer program
Suddenly.
Processor can be central processing unit (Central Processing Unit, CPU).Processor can also be it
His general processor, digital signal processor (Digital Signal Processor, DSP), specific integrated circuit
(Application Specific Integrated Circuit, ASIC), field programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
The combination of the chips such as discrete hardware components or above-mentioned all kinds of chips.
Memory as a kind of non-transient computer readable storage medium, can be used for storing non-transient software program, it is non-temporarily
State computer executable program and unit, such as corresponding program unit in above method embodiment of the present invention.Processor passes through
Non-transient software program, instruction and module stored in memory are run, thereby executing the various function application of processor
And work data processing, that is, realize the method in above method embodiment.
Memory may include storing program area and storage data area, wherein storing program area can storage program area, extremely
Application program required for a few function；It storage data area can the data etc. that are created of storage processor.In addition, memory can
It can also include non-transient memory, for example, at least disk memory, a flash memory to include high-speed random access memory
Device or other non-transient solid-state memories.In some embodiments, it includes remotely setting relative to processor that memory is optional
The memory set, these remote memories can pass through network connection to processor.The example of above-mentioned network includes but is not limited to
Internet, intranet, local area network, mobile radio communication and combinations thereof.
One or more of unit storages in the memory, when being executed by the processor, execute above-mentioned
Method in embodiment.
Above-mentioned computer equipment detail can correspond to refering to associated description corresponding in above-described embodiment and effect into
Row understands that details are not described herein again.
To achieve the goals above, according to the another aspect of the application, a kind of computer readable storage medium is additionally provided,
The computer-readable recording medium storage has computer program, real when the computer program executes in the computer processor
Step in the existing above-mentioned scenic spot passenger flow forecast method based on machine learning.It will be understood by those skilled in the art that on realizing
The all or part of the process in embodiment method is stated, is that relevant hardware can be instructed to complete by computer program, institute
The program stated can be stored in a computer-readable storage medium, and the program is when being executed, it may include such as above-mentioned each method
The process of embodiment.Wherein, the storage medium can for magnetic disk, CD, read-only memory (Read-Only Memory,
ROM), random access memory (RandomAccessMemory, RAM), flash memory (Flash Memory), hard disk
(Hard Disk Drive, abbreviation: HDD) or solid state hard disk (Solid-State Drive, SSD) etc.；The storage medium is also
It may include the combination of the memory of mentioned kind.
Obviously, those skilled in the art should be understood that each module of the above invention or each step can be with general
Computing device realize that they can be concentrated on a single computing device, or be distributed in multiple computing devices and formed
Network on, optionally, they can be realized with the program code that computing device can perform, it is thus possible to which they are stored
Be performed by computing device in the storage device, perhaps they are fabricated to each integrated circuit modules or by they
In multiple modules or step be fabricated to single integrated circuit module to realize.In this way, the present invention is not limited to any specific
Hardware and software combines.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field
For art personnel, the invention may be variously modified and varied.All within the spirits and principles of the present invention, made any to repair
Change, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.
Claims (18)
1. a kind of scenic spot passenger flow forecast method based on machine learning characterized by comprising
The history passenger flow characteristic at scenic spot is generated according to the scenic spot history pipelined data of acquisition and scenic spot characteristic data；It is described
Scenic spot characteristic data include: festivals or holidays data, scenic spot position data, scenic spot weather data, scenic spot dull and rush season data, entrance ticket
At least one of price data；
Each machine learning is respectively trained out according to the history passenger flow characteristic and preset a variety of machine learning algorithms to calculate
The corresponding passenger flow forecast model of method；
The prediction effect of the corresponding passenger flow forecast model of each machine learning algorithm is assessed, and according to assessment result from each
The preferable passenger flow forecast model of multiple prediction effects is selected in the corresponding passenger flow forecast model of machine learning algorithm；
Determine the weight vectors for the passenger flow forecast model selected, and according to the weight vectors and the volume of the flow of passengers selected
Prediction model carries out scenic spot passenger flow forecast.
2. the scenic spot passenger flow forecast method according to claim 1 based on machine learning, which is characterized in that described to each
The prediction effect of the corresponding passenger flow forecast model of machine learning algorithm is assessed, and is specifically included:
Verify data set is generated according to the history passenger flow characteristic；
It is pre- that each verify data in the verify data set is separately input to the corresponding volume of the flow of passengers of each machine learning algorithm
Model is surveyed, prediction result is obtained；
Each passenger flow forecast model pair is calculated according to the corresponding prediction result of each passenger flow forecast model and RMSE calculation method
The RMSE value answered.
3. the scenic spot passenger flow forecast method according to claim 2 based on machine learning, which is characterized in that the basis
Assessment result selects the preferable volume of the flow of passengers of multiple prediction effects from the corresponding passenger flow forecast model of each machine learning algorithm
Prediction model specifically includes:
The selection of passenger flow forecast model is carried out according to the corresponding RMSE value of each passenger flow forecast model.
4. the scenic spot passenger flow forecast method according to claim 1 based on machine learning, which is characterized in that the determination
The weight vectors for the passenger flow forecast model selected, specifically include:
Weight vectors set is generated according to the quantity for the passenger flow forecast model selected；
By each weight vectors in the weight vectors set respectively with the prediction result for the passenger flow forecast model selected into
Row weighted average calculation obtains the corresponding result of weighted average of each weight vectors；
Calculate the RMSE value of each result of weighted average using RMSE calculation method, and determine the corresponding weight of the smallest RMSE value to
Amount.
5. the scenic spot passenger flow forecast method according to claim 4 based on machine learning, which is characterized in that the basis
The quantity for the passenger flow forecast model selected generates weight vectors set, specifically includes:
It determines that weight changes granularity 1/H, and granularity is changed according to weight and generates set M, wherein M={ 1,1,1,1 ..., 1 },
Altogether comprising H 1 in set M；
The use of plug hole method by H 1 point in set M is n group, sharesThe kind method of salary distribution, respectively in every kind of method of salary distribution
The 1 of every group adds up, and divided by H, is sharedThe weight vectors set of a equally distributed weight vectors, wherein n is
The quantity for the passenger flow forecast model selected.
6. the scenic spot passenger flow forecast method according to claim 1 based on machine learning, which is characterized in that the basis
It is corresponding that each machine learning algorithm is respectively trained out in the history passenger flow characteristic and preset a variety of machine learning algorithms
Passenger flow forecast model, specifically includes:
It is corresponding using machine learning algorithm according to the parameter value range of the machine learning algorithm of setting and value change step
Hyper parameter searching algorithm dynamically finds the hyper parameter for keeping forecast result of model optimal, so that the passenger flow forecast model trained
Prediction effect global optimum or local optimum.
7. the scenic spot passenger flow forecast method according to claim 1 based on machine learning, which is characterized in that the basis
The weight vectors and the passenger flow forecast model selected carry out scenic spot passenger flow forecast, comprising:
Obtain current passenger flow characteristic；
The current passenger flow characteristic is separately input to each passenger flow forecast model selected, obtains each passenger flow forecast
The passenger flow forecast data of model output；
Go out comprehensive passenger flow forecast data according to the passenger flow forecast data and the weight vector computation.
8. the scenic spot passenger flow forecast method according to claim 7 based on machine learning, which is characterized in that further include:
Periodically calculate the comprehensive passenger flow forecast data and practical visitor corresponding with the comprehensive passenger flow forecast data
Deviation amplitude between data on flows；
Re -training model signals are sent when the deviation amplitude is greater than preset value.
9. a kind of scenic spot passenger flow forecast device based on machine learning characterized by comprising
History passenger flow characteristic acquiring unit, for raw according to the scenic spot history pipelined data and scenic spot characteristic data of acquisition
At the history passenger flow characteristic at scenic spot；The scenic spot characteristic data include: festivals or holidays data, scenic spot position data, scenic spot day
At least one of destiny evidence, scenic spot dull and rush season data, entrance ticket price data；
Passenger flow forecast model training unit, for according to the history passenger flow characteristic and preset a variety of machine learning
The corresponding passenger flow forecast model of each machine learning algorithm is respectively trained out in algorithm；
Model evaluation screening unit is commented for the prediction effect to the corresponding passenger flow forecast model of each machine learning algorithm
Estimate, and it is preferable according to assessment result to select multiple prediction effects from the corresponding passenger flow forecast model of each machine learning algorithm
Passenger flow forecast model；
Passenger flow forecast unit, for determining the weight vectors of passenger flow forecast model selected, and according to the weight to
The passenger flow forecast model measured and selected carries out scenic spot passenger flow forecast.
10. the scenic spot passenger flow forecast device according to claim 9 based on machine learning, which is characterized in that the mould
Type assesses screening unit, comprising:
Verify data generation module, for generating verify data set according to the history passenger flow characteristic；
Model prediction result obtains module, for each verify data in the verify data set to be separately input to each machine
The corresponding passenger flow forecast model of device learning algorithm, obtains prediction result；
RMSE value computing module, for being calculated according to the corresponding prediction result of each passenger flow forecast model and RMSE calculation method
The corresponding RMSE value of each passenger flow forecast model.
11. the scenic spot passenger flow forecast device according to claim 10 based on machine learning, which is characterized in that the mould
Type assesses screening unit, specifically for carrying out the choosing of passenger flow forecast model according to the corresponding RMSE value of each passenger flow forecast model
It selects.
12. the scenic spot passenger flow forecast device according to claim 9 based on machine learning, which is characterized in that the visitor
Volume forecasting unit, comprising:
Weight vectors set generation module, for generating weight vectors collection according to the quantity for the passenger flow forecast model selected
It closes；
Weighted calculation module, for by each weight vectors in the weight vectors set respectively with the passenger flow forecast selected
The prediction result of model is weighted and averaged calculating, obtains the corresponding result of weighted average of each weight vectors；
Optimal weights vector selecting module, for calculating the RMSE value of each result of weighted average using RMSE calculation method, and really
Determine the corresponding weight vectors of the smallest RMSE value.
13. the scenic spot passenger flow forecast device according to claim 12 based on machine learning, which is characterized in that the power
Weight vector set closes generation module, comprising:
Weight variation granularity determines submodule, for determining that weight changes granularity 1/H, and changes granularity according to weight and generates set
M, wherein altogether comprising H 1 in M={ 1,1,1,1 ..., 1 }, set M；
Plug hole method handles submodule, by H 1 point in set M is n group for using plug hole method, sharedThe kind method of salary distribution,
It adds up the 1 of every group in every kind of method of salary distribution, and divided by H, is shared respectivelyA equally distributed weight vectors
Weight vectors set, wherein n is the quantity for the passenger flow forecast model selected.
14. the scenic spot passenger flow forecast device according to claim 9 based on machine learning, which is characterized in that the visitor
Flux prediction model training unit, comprising:
Hyper parameter tuning module, for being used according to the parameter value range and value change step of the machine learning algorithm of setting
The corresponding hyper parameter searching algorithm of machine learning algorithm dynamically finds the hyper parameter for keeping forecast result of model optimal, so as to train
Passenger flow forecast model prediction effect global optimum or local optimum.
15. the scenic spot passenger flow forecast device according to claim 9 based on machine learning, which is characterized in that the visitor
Volume forecasting unit, comprising:
Current passenger flow characteristic obtains module, for obtaining current passenger flow characteristic；
Model prediction module, each passenger flow forecast mould for the current passenger flow characteristic to be separately input to select
Type obtains the passenger flow forecast data of each passenger flow forecast model output；
Integrated forecasting data computation module, for going out to integrate according to the passenger flow forecast data and the weight vector computation
Passenger flow forecast data.
16. the scenic spot passenger flow forecast device according to claim 15 based on machine learning, which is characterized in that also wrap
It includes:
Deviation amplitude computing unit, for periodically calculate the comprehensive passenger flow forecast data and with the comprehensive volume of the flow of passengers
Deviation amplitude between the corresponding practical volume of the flow of passengers data of prediction data；
Model re -training unit, for sending re -training model signals when the deviation amplitude is greater than preset value.
17. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor
Calculation machine program, which is characterized in that the processor realizes any one of claim 1 to 8 method when executing the computer program
In step.
18. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In realization such as the step in claim 1 to 8 any one method when the computer program executes in the computer processor
Suddenly.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910728598.9A CN110443314A (en) | 2019-08-08 | 2019-08-08 | Scenic spot passenger flow forecast method and device based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910728598.9A CN110443314A (en) | 2019-08-08 | 2019-08-08 | Scenic spot passenger flow forecast method and device based on machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110443314A true CN110443314A (en) | 2019-11-12 |
Family
ID=68433747
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910728598.9A Pending CN110443314A (en) | 2019-08-08 | 2019-08-08 | Scenic spot passenger flow forecast method and device based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110443314A (en) |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110929926A (en) * | 2019-11-18 | 2020-03-27 | 西北工业大学 | Short-term explosion passenger flow prediction method based on long and short-term memory network and random forest |
| CN111027771A (en) * | 2019-12-10 | 2020-04-17 | 浙江力石科技股份有限公司 | Scenic spot passenger flow volume estimation method, system and device and storable medium |
| CN111026626A (en) * | 2019-11-29 | 2020-04-17 | 中国建设银行股份有限公司 | CPU consumption estimation and estimation model training method and device |
| CN111126624A (en) * | 2019-12-20 | 2020-05-08 | 北京首汽智行科技有限公司 | Method for judging validity of model prediction result |
| CN111144652A (en) * | 2019-12-26 | 2020-05-12 | 浙江力石科技股份有限公司 | Tour comfort degree algorithm and trend prediction method, system and device |
| CN111144660A (en) * | 2019-12-30 | 2020-05-12 | 北京百度网讯科技有限公司 | People flow prediction method and device, electronic equipment and storage medium |
| CN111178633A (en) * | 2019-12-31 | 2020-05-19 | 上饶市中科院云计算中心大数据研究院 | Method and device for predicting scenic spot passenger flow based on random forest algorithm |
| CN111353828A (en) * | 2020-03-30 | 2020-06-30 | 中国工商银行股份有限公司 | Method and device for predicting number of people arriving at store from network |
| CN111428911A (en) * | 2020-02-29 | 2020-07-17 | 中国平安人寿保险股份有限公司 | Method and device for predicting number of people |
| CN111556445A (en) * | 2020-05-14 | 2020-08-18 | 中原工学院 | Scenic spot management system based on real-time dynamic position information feedback of tourists |
| CN111651444A (en) * | 2020-05-25 | 2020-09-11 | 成都千嘉科技有限公司 | Self-adaptive time series data prediction method |
| CN111754336A (en) * | 2020-06-28 | 2020-10-09 | 中国银行股份有限公司 | Network scheduling method and device |
| CN111754326A (en) * | 2020-06-24 | 2020-10-09 | 中国银行股份有限公司 | Risk early warning method and device |
| CN111768093A (en) * | 2020-06-23 | 2020-10-13 | 中国工商银行股份有限公司 | Network point configuration method and device based on time sequence |
| CN111796995A (en) * | 2020-06-30 | 2020-10-20 | 中国工商银行股份有限公司 | Cyclic serial number usage early warning method and system based on ensemble learning |
| CN111898322A (en) * | 2020-08-11 | 2020-11-06 | 腾讯科技（深圳）有限公司 | Data processing method and related equipment |
| CN112235152A (en) * | 2020-09-04 | 2021-01-15 | 北京邮电大学 | Flow size estimation method and device |
| CN112465183A (en) * | 2020-10-15 | 2021-03-09 | 上海齐崂科技发展有限公司 | Sewage treatment data acquisition method and application thereof |
| CN112561153A (en) * | 2020-12-08 | 2021-03-26 | 南京莱斯信息技术股份有限公司 | Scenic spot crowd gathering prediction method based on model integration |
| CN112819258A (en) * | 2021-03-24 | 2021-05-18 | 中国工商银行股份有限公司 | Bank branch to store customer quantity prediction method and device |
| CN113159364A (en) * | 2020-12-30 | 2021-07-23 | ***通信集团广东有限公司珠海分公司 | Passenger flow prediction method and system for large-scale traffic station |
| CN113205884A (en) * | 2021-05-12 | 2021-08-03 | 中国科学院深圳先进技术研究院 | Blood sugar prediction method, system and application thereof |
| CN113313318A (en) * | 2021-06-15 | 2021-08-27 | 京东方科技集团股份有限公司 | Scenic spot passenger flow detection method and related equipment |
| CN113450557A (en) * | 2020-03-24 | 2021-09-28 | 支付宝(杭州)信息技术有限公司 | Method and device for updating prediction model for vehicle passenger flow |
| CN113537656A (en) * | 2020-04-14 | 2021-10-22 | 华晨宝马汽车有限公司 | Method, system, device and medium for assessing quality of service of a provider |
| CN113590962A (en) * | 2021-08-03 | 2021-11-02 | 合肥泰瑞数创科技有限公司 | Flow data prediction system, method, computer equipment and medium |
| CN113743971A (en) * | 2020-06-17 | 2021-12-03 | 北京沃东天骏信息技术有限公司 | Data processing method and device |
| CN113837581A (en) * | 2021-09-16 | 2021-12-24 | 江苏泰扬金属制品有限公司 | Big data application dispatching system and method |
| TWI793412B (en) * | 2020-03-31 | 2023-02-21 | 廣達電腦股份有限公司 | Consumption prediction system and consumption prediction method |
| CN117455079A (en) * | 2023-12-22 | 2024-01-26 | 烟台云朵软件有限公司 | Macroscopic passenger flow volume prediction method and macroscopic passenger flow volume prediction system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106960250A (en) * | 2017-03-03 | 2017-07-18 | 陕西师范大学 | A kind of method of dynamic prediction tourist attractions passenger flow |
| CN107145962A (en) * | 2017-04-01 | 2017-09-08 | 上海诺悦智能科技有限公司 | A kind of sight spot domestic visitors forecasting system |
| CN107590569A (en) * | 2017-09-25 | 2018-01-16 | 山东浪潮云服务信息科技有限公司 | A kind of data predication method and device |
| CN109034469A (en) * | 2018-07-20 | 2018-12-18 | 成都中科大旗软件有限公司 | A kind of tourist flow prediction technique based on machine learning |
-
2019
- 2019-08-08 CN CN201910728598.9A patent/CN110443314A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106960250A (en) * | 2017-03-03 | 2017-07-18 | 陕西师范大学 | A kind of method of dynamic prediction tourist attractions passenger flow |
| CN107145962A (en) * | 2017-04-01 | 2017-09-08 | 上海诺悦智能科技有限公司 | A kind of sight spot domestic visitors forecasting system |
| CN107590569A (en) * | 2017-09-25 | 2018-01-16 | 山东浪潮云服务信息科技有限公司 | A kind of data predication method and device |
| CN109034469A (en) * | 2018-07-20 | 2018-12-18 | 成都中科大旗软件有限公司 | A kind of tourist flow prediction technique based on machine learning |
Non-Patent Citations (1)
| Title |
|---|
| 宋宇辰: "《基于***动力学的能源 经济 环境 人口可持续发展建模研究》", 30 November 2016, 冶金工业出版社 * |
Cited By (42)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110929926A (en) * | 2019-11-18 | 2020-03-27 | 西北工业大学 | Short-term explosion passenger flow prediction method based on long and short-term memory network and random forest |
| CN111026626A (en) * | 2019-11-29 | 2020-04-17 | 中国建设银行股份有限公司 | CPU consumption estimation and estimation model training method and device |
| CN111027771A (en) * | 2019-12-10 | 2020-04-17 | 浙江力石科技股份有限公司 | Scenic spot passenger flow volume estimation method, system and device and storable medium |
| CN111126624A (en) * | 2019-12-20 | 2020-05-08 | 北京首汽智行科技有限公司 | Method for judging validity of model prediction result |
| CN111144652B (en) * | 2019-12-26 | 2023-08-08 | 浙江力石科技股份有限公司 | Tour comfort algorithm and trend prediction based method, system and device |
| CN111144652A (en) * | 2019-12-26 | 2020-05-12 | 浙江力石科技股份有限公司 | Tour comfort degree algorithm and trend prediction method, system and device |
| CN111144660A (en) * | 2019-12-30 | 2020-05-12 | 北京百度网讯科技有限公司 | People flow prediction method and device, electronic equipment and storage medium |
| CN111178633A (en) * | 2019-12-31 | 2020-05-19 | 上饶市中科院云计算中心大数据研究院 | Method and device for predicting scenic spot passenger flow based on random forest algorithm |
| CN111428911A (en) * | 2020-02-29 | 2020-07-17 | 中国平安人寿保险股份有限公司 | Method and device for predicting number of people |
| CN113450557B (en) * | 2020-03-24 | 2022-06-28 | 支付宝(杭州)信息技术有限公司 | Method and device for updating prediction model for passenger flow of vehicle |
| CN113450557A (en) * | 2020-03-24 | 2021-09-28 | 支付宝(杭州)信息技术有限公司 | Method and device for updating prediction model for vehicle passenger flow |
| CN111353828B (en) * | 2020-03-30 | 2023-09-12 | 中国工商银行股份有限公司 | Method and device for predicting number of people coming to store at website |
| CN111353828A (en) * | 2020-03-30 | 2020-06-30 | 中国工商银行股份有限公司 | Method and device for predicting number of people arriving at store from network |
| TWI793412B (en) * | 2020-03-31 | 2023-02-21 | 廣達電腦股份有限公司 | Consumption prediction system and consumption prediction method |
| US11983726B2 (en) | 2020-03-31 | 2024-05-14 | Quanta Computer Inc. | Consumption prediction system and consumption prediction method |
| CN113537656A (en) * | 2020-04-14 | 2021-10-22 | 华晨宝马汽车有限公司 | Method, system, device and medium for assessing quality of service of a provider |
| CN111556445B (en) * | 2020-05-14 | 2023-05-30 | 中原工学院 | Scenic spot management system based on tourist real-time dynamic position information feedback |
| CN111556445A (en) * | 2020-05-14 | 2020-08-18 | 中原工学院 | Scenic spot management system based on real-time dynamic position information feedback of tourists |
| CN111651444B (en) * | 2020-05-25 | 2023-04-18 | 成都千嘉科技股份有限公司 | Self-adaptive time series data prediction method |
| CN111651444A (en) * | 2020-05-25 | 2020-09-11 | 成都千嘉科技有限公司 | Self-adaptive time series data prediction method |
| CN113743971A (en) * | 2020-06-17 | 2021-12-03 | 北京沃东天骏信息技术有限公司 | Data processing method and device |
| CN111768093A (en) * | 2020-06-23 | 2020-10-13 | 中国工商银行股份有限公司 | Network point configuration method and device based on time sequence |
| CN111754326A (en) * | 2020-06-24 | 2020-10-09 | 中国银行股份有限公司 | Risk early warning method and device |
| CN111754326B (en) * | 2020-06-24 | 2024-04-05 | 中国银行股份有限公司 | Risk early warning method and device |
| CN111754336A (en) * | 2020-06-28 | 2020-10-09 | 中国银行股份有限公司 | Network scheduling method and device |
| CN111754336B (en) * | 2020-06-28 | 2024-05-24 | 中国银行股份有限公司 | Method and device for scheduling network points |
| CN111796995A (en) * | 2020-06-30 | 2020-10-20 | 中国工商银行股份有限公司 | Cyclic serial number usage early warning method and system based on ensemble learning |
| CN111796995B (en) * | 2020-06-30 | 2024-02-09 | 中国工商银行股份有限公司 | Integrated learning-based cyclic serial number usage early warning method and system |
| CN111898322B (en) * | 2020-08-11 | 2024-03-01 | 腾讯科技（深圳）有限公司 | Data processing method and related equipment |
| CN111898322A (en) * | 2020-08-11 | 2020-11-06 | 腾讯科技（深圳）有限公司 | Data processing method and related equipment |
| CN112235152A (en) * | 2020-09-04 | 2021-01-15 | 北京邮电大学 | Flow size estimation method and device |
| CN112465183A (en) * | 2020-10-15 | 2021-03-09 | 上海齐崂科技发展有限公司 | Sewage treatment data acquisition method and application thereof |
| CN112561153A (en) * | 2020-12-08 | 2021-03-26 | 南京莱斯信息技术股份有限公司 | Scenic spot crowd gathering prediction method based on model integration |
| CN113159364A (en) * | 2020-12-30 | 2021-07-23 | ***通信集团广东有限公司珠海分公司 | Passenger flow prediction method and system for large-scale traffic station |
| CN112819258A (en) * | 2021-03-24 | 2021-05-18 | 中国工商银行股份有限公司 | Bank branch to store customer quantity prediction method and device |
| CN113205884B (en) * | 2021-05-12 | 2023-02-24 | 中国科学院深圳先进技术研究院 | Blood sugar prediction method, system and application thereof |
| CN113205884A (en) * | 2021-05-12 | 2021-08-03 | 中国科学院深圳先进技术研究院 | Blood sugar prediction method, system and application thereof |
| CN113313318A (en) * | 2021-06-15 | 2021-08-27 | 京东方科技集团股份有限公司 | Scenic spot passenger flow detection method and related equipment |
| CN113590962B (en) * | 2021-08-03 | 2023-07-25 | 合肥泰瑞数创科技有限公司 | Flow data prediction system, method, computer equipment and medium |
| CN113590962A (en) * | 2021-08-03 | 2021-11-02 | 合肥泰瑞数创科技有限公司 | Flow data prediction system, method, computer equipment and medium |
| CN113837581A (en) * | 2021-09-16 | 2021-12-24 | 江苏泰扬金属制品有限公司 | Big data application dispatching system and method |
| CN117455079A (en) * | 2023-12-22 | 2024-01-26 | 烟台云朵软件有限公司 | Macroscopic passenger flow volume prediction method and macroscopic passenger flow volume prediction system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110443314A (en) | Scenic spot passenger flow forecast method and device based on machine learning | |
| He et al. | Mining transition rules of cellular automata for simulating urban expansion by using the deep learning techniques | |
| Liu et al. | A multi-type ant colony optimization (MACO) method for optimal land use allocation in large areas | |
| CN110400022A (en) | Self-help teller machine cash dosage prediction technique and device | |
| CN110400021A (en) | Bank outlets' cash dosage prediction technique and device | |
| Bununu | Integration of Markov chain analysis and similarity-weighted instance-based machine learning algorithm (SimWeight) to simulate urban expansion | |
| CN107194605A (en) | Energy consumption estimation system and method | |
| Akay et al. | Using optimization techniques in designing forest roads and road networks | |
| Rosendorff et al. | Artificial Intelligence for last-mile logistics-Procedures and architecture | |
| Islam et al. | Mineral import demand and wind energy deployment in the USA: Co-integration and counterfactual analysis approaches | |
| CN112767126A (en) | Collateral grading method and device based on big data | |
| CN116959249A (en) | City information management platform and method based on CIM | |
| CN115691140B (en) | Analysis and prediction method for space-time distribution of automobile charging demand | |
| Vukeya | The impact of infrastructure investment on economic growth in South Africa | |
| Arnaiz et al. | Facilitating universal energy access for developing countries with micro-hydropower: Insights from Nepal, Bolivia, Cambodia and the Philippines | |
| Patil et al. | GIS-based multi-criteria decision-making for ranking potential sites for centralized rainwater harvesting | |
| Mukhamediev et al. | Multi-criteria decision support system for RES evaluation | |
| Zou et al. | An empirical ensemble rainfall nowcasting model using multi-scaled analogues | |
| JP2021197085A (en) | Server, road network planning system, road network planning metho, and program | |
| Du et al. | Mining multicity urban data for sustainable population relocation | |
| CN104778636A (en) | Ship equipment maintenance support information service recommendation method | |
| Tang et al. | A data-driven framework for natural feature profile of public transport ridership: Insights from Suzhou and Lianyungang, China | |
| Peng et al. | Development of a prototype land use model for statewide transportation planning activities. | |
| Rohrbach et al. | Parameter-free aggregation of value functions from multiple experts and uncertainty assessment in multi-criteria evaluation | |
| Pan | Advancing precipitation prediction using a composite of models and data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191112 |