CN112380425A - Community recommendation method, system, computer equipment and storage medium - Google Patents
Community recommendation method, system, computer equipment and storage medium Download PDFInfo
- Publication number
- CN112380425A CN112380425A CN202011142989.1A CN202011142989A CN112380425A CN 112380425 A CN112380425 A CN 112380425A CN 202011142989 A CN202011142989 A CN 202011142989A CN 112380425 A CN112380425 A CN 112380425A
- Authority
- CN
- China
- Prior art keywords
- community
- index
- layer
- indexes
- attribute data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000011156 evaluation Methods 0.000 claims abstract description 47
- 238000012549 training Methods 0.000 claims abstract description 38
- 238000003062 neural network model Methods 0.000 claims abstract description 33
- 239000011159 matrix material Substances 0.000 claims description 46
- 239000013598 vector Substances 0.000 claims description 27
- 238000010586 diagram Methods 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 7
- 235000019013 Viburnum opulus Nutrition 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 6
- 230000011664 signaling Effects 0.000 claims description 5
- 230000001186 cumulative effect Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 244000071378 Viburnum opulus Species 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 230000006399 behavior Effects 0.000 abstract description 6
- 238000010276 construction Methods 0.000 description 11
- 238000004590 computer program Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000011835 investigation Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 230000003203 everyday effect Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000011278 co-treatment Methods 0.000 description 1
- 238000005314 correlation function Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Development Economics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a community recommendation method, a community recommendation system, computer equipment and a storage medium, wherein the method comprises the following steps: acquiring resident socioeconomic attribute data and community attribute data; constructing a community evaluation index system; using the resident socioeconomic attribute data and the community attribute data as influencing factors, using the subentry index and the complete community index of each community as a predicted index, and training by using a back propagation neural network model to obtain a trained model; acquiring socioeconomic attribute data input by a user, and predicting by using the trained model to obtain a predicted value of each index of the community corresponding to the user; matching the predicted value of each index of the community corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result; and outputting a community recommendation list according to the matching result. The invention can provide guidance for decision and practice of participating in urban space and service reconstruction and provide reference for citizens' living selection behaviors.
Description
Technical Field
The invention relates to a community recommendation method, a community recommendation system, computer equipment and a storage medium, and belongs to the field of urban planning and real estate.
Background
In the 12-month national housing and urban and rural construction work meeting in 2019, the constructed complete community is classified into the working key points of the 2020 national housing and urban and rural construction department, the community infrastructure and public service are emphasized to be perfected, the housing community environment is created, local characteristic culture is built, and a co-construction co-treatment shared community treatment system is constructed. At present, the city planning industry has a certain discussion on the connotation of a complete community, namely the community has dual attributes of material space and social space, and various factors such as living room, education, sanitation and the like need to be considered, so that the governing attribute of the community is emphasized, and the care of people is reflected. In 2019, the urbanization rate of the national resident population is increased to 60.6%, the development strategy of urbanization enters the transformation period from scale expansion to quality improvement, and the key of community construction is to meet the increasingly beautiful life needs of people. At present, technical means about community and house source recommendation generally lack attention to community governance attributes, and social and economic characteristic attributes of community residents are not considered sufficiently.
Disclosure of Invention
In view of the above, the present invention provides a community recommendation method, system, computer device, and storage medium, which can provide a recommendation scheme for citizens 'living choices, help to provide research foundation and methodology test for planning and construction of complete communities, accurately grasp customer figures for government departments and enterprises, provide guidance for decision and practice of participating in urban space and service reconstruction, and provide reference for citizens' living choice behavior.
The first purpose of the invention is to provide a community recommendation method.
A second object of the present invention is to provide a community recommendation system.
It is a third object of the invention to provide a computer apparatus.
It is a fourth object of the present invention to provide a storage medium.
The first purpose of the invention can be achieved by adopting the following technical scheme:
a community recommendation method, the method comprising:
acquiring resident socioeconomic attribute data and community attribute data;
constructing a community evaluation index system; the community evaluation index system comprises a subentry index and a complete community index of each community;
using the resident socioeconomic attribute data and the community attribute data as influencing factors, using the subentry index and the complete community index of each community as a predicted index, and training by using a back propagation neural network model to obtain a trained model;
acquiring socioeconomic attribute data input by a user, and predicting by using the trained model to obtain a predicted value of each index of the community corresponding to the user;
matching the predicted value of each index of the community corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result;
and outputting a community recommendation list according to the matching result.
Further, the building of the community evaluation index system specifically includes:
for each community, education, medical treatment, commercial service, sports and leisure, public transportation, environment, economy, location and management are selected as criteria layer indexes of an evaluation system, and are subdivided downwards into fifty-three evaluation indexes;
calculating nine criteria layer indexes and weight values of evaluation indexes in the criteria layer indexes by using an analytic hierarchy process, wherein the calculation method specifically comprises the following steps: acquiring a certain amount of resident samples randomly through snowball type sampling, and requesting each resident to sort every two of nine criterion layer indexes and importance degrees of each evaluation index in each criterion layer index, integrating opinions of all visited residents, sorting every two of the nine criterion layer indexes and each evaluation index subdivided by the nine criterion layer indexes, assigning values according to priorities, listing the highest total score as the most preferred index, and calculating weight values of the nine criterion layer indexes and each evaluation index in each criterion layer index;
normalizing the data of each evaluation index in each criterion layer index, multiplying the data by the sum of weights to obtain the coefficients of nine criterion layer indexes, and taking the coefficients as the subentry indexes of nine communities;
and multiplying and adding the nine indexes and the corresponding weights, and performing normalization processing to obtain a complete community index serving as an index for comprehensively evaluating the community.
Further, the calculating of the weight values of the nine criterion layer indexes and the evaluation indexes inside the criterion layer indexes based on the analytic hierarchy process specifically includes:
dividing the decision into a highest layer, a middle layer and a lowest layer according to the decision target, the considered factors and the interrelation among decision objects, and drawing a hierarchical structure diagram to establish a hierarchical structure model;
constructing a judgment matrix by using a consistent matrix method;
the eigenvector corresponding to the maximum characteristic root of the judgment matrix is normalized and then recorded as W, and the element of W is a sequencing weight of the relative importance of the same-layer element to the previous-layer element;
introducing a random consistency index RI, defining a consistency ratio CR to be CI/RI, if the consistency ratio CR is less than 0.1, using a normalized feature vector of a judgment matrix as a weight vector through consistency test, otherwise, reconstructing the judgment matrix for adjustment;
and sequentially calculating the weight values of the relative importance of all factors of a certain level to the highest level from the highest level to the lowest level so as to obtain nine criteria level indexes and the weight values of all evaluation indexes in all criteria level indexes.
Further, the back propagation neural network model comprises an input layer, a hidden layer and an output layer, wherein a weight coefficient matrix from the input layer to the hidden layer is a first weight coefficient matrix, the bias of the hidden layer is a first bias, a weight coefficient matrix from the hidden layer to the output layer is a second weight coefficient matrix, and the bias of the output layer is a second bias;
the method comprises the following steps of training by using a back propagation neural network model by using resident socioeconomic attribute data and community attribute data as influencing factors and using subentry indexes and complete community indexes of various communities as predicted indexes to obtain a trained model, and specifically comprises the following steps:
the method comprises the following steps of (1) forming a training set by taking resident socioeconomic attribute data and community attribute data as influencing factors and taking the subentry index and the complete community index of each community as a predicted index;
setting a learning rate, a termination threshold, a maximum step number, a first weight coefficient matrix, a second weight coefficient matrix, a first bias and a second bias;
inputting the training set into a back propagation neural network model, calculating the input value and the output value of each layer, calculating the gradient/required partial derivative of the weight coefficient and the bias, and updating a first weight coefficient matrix, a second weight coefficient matrix, a first bias and a second bias;
and continuously iterating to correct the first weight coefficient matrix, the second weight coefficient matrix, the first bias and the second bias to obtain the trained model.
Further, the training set is denoted as D { (X)1,T1),(X2,T2),…,(XP,TP) }; wherein each input sample comprises n elements, corresponding to n nodes of the input layer,
each target output sample contains m elements, corresponding to m nodes of the output layer,
the hidden layer has l nodes, and the input of the jth node is set as
The first bias is thetajOutput is
The activation function is noted as f (x),
then
The calculation formula of (a) is as follows:
the output layer has m nodes, and the input of the kth node is set as
The second bias is phikOutput is
The actual value is
Then
The calculation formula of (a) is as follows:
the mean square error at sample point p is as follows:
the cumulative error is as follows:
the number of parameters to be determined in the whole network is as follows:
(n×l+l)+(l×m+m)＝(n+m+1)l+m
the updating formula of the parameters is as follows: e ← e + Δ e, where Δ e is updated according to a gradient descent strategy.
Further, the matching of the predicted value of each index of the community corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result specifically includes:
taking the predicted value of each community index corresponding to the user as a first sentence, and taking the subentry index and the complete community index of each community as a second sentence;
calculating the matching degree of the first sentence and the second sentence by a cosine similarity algorithm, which is as follows:
wherein x isiRepresenting individual word vectors, y, in a first sentenceiEach word vector in the second sentence is represented, and cos (theta) represents the cosine similarity of the included angle theta.
Further, the resident socioeconomic attribute data comprises gender, age, occupation, entertainment preference, telephone charge level and hobbies, and is acquired based on the mobile phone signaling data;
the community attribute data comprises facility attributes, environment attributes, location attributes, economic attributes and governance attributes, the facility attributes comprise education, medical treatment, commercial service, sports leisure and public transportation facilities, the environment attributes comprise greenbelts and water systems, the economic attributes comprise community housing prices and rent, the location attributes use the distance representation of the community from a central business center, and the governance attributes comprise government governance and resident participation.
The second purpose of the invention can be achieved by adopting the following technical scheme:
a recommendation system for community selection, the system comprising:
the acquisition module is used for acquiring the social and economic attribute data of residents and the community attribute data;
the building module is used for building a community evaluation index system; the community evaluation index system comprises a subentry index and a complete community index of each community;
the training module is used for training by using a back propagation neural network model by taking resident socioeconomic attribute data and community attribute data as influencing factors and taking the subentry index and the complete community index of each community as a predicted index to obtain a trained model;
the prediction module is used for acquiring socioeconomic attribute data input by a user, and predicting by using the trained model to obtain a predicted value of each index of the community corresponding to the user;
the matching module is used for matching the predicted value of each index of the community corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result;
and the recommending module is used for outputting a community recommending list according to the matching result.
The third purpose of the invention can be achieved by adopting the following technical scheme:
a computer device comprising a processor and a memory for storing a program executable by the processor, the processor implementing the proposed method when executing the program stored in the memory.
The fourth purpose of the invention can be achieved by adopting the following technical scheme:
a storage medium stores a program that, when executed by a processor, implements the recommendation method described above.
Compared with the prior art, the invention has the following beneficial effects:
the invention obtains the social economic attribute data and the community attribute data of residents and constructs a community evaluation index system which comprises the subentry index and the complete community index of each community, takes the social economic attribute data and the community attribute data of residents as influencing factors, takes the subentry index and the complete community index of each community as predicted indexes, uses a back propagation neural network model for training, predicts by the trained model, matches with the subentry index and the complete community index of each community, can provide a recommendation scheme for the community selection of the citizens according to the matching result, is beneficial to providing research foundation and methodology test for the planning and construction of the complete community, also provides an accurate grasp of client figures for government departments and enterprises, and provides an index for the decision and practice of participating in urban space and service reconstruction, and provides reference for community selection behavior of citizens.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the structures shown in the drawings without creative efforts.
Fig. 1 is a flowchart of a community recommendation method according to embodiment 1 of the present invention.
Fig. 2 is a structural diagram of a back propagation neural network model according to embodiment 1 of the present invention.
Fig. 3 is a structural diagram of a back propagation neural network model of the demonstration case of embodiment 2 of the present invention.
Fig. 4 is a diagram of an operation result of the back propagation neural network model using the training set in embodiment 2 of the present invention.
Fig. 5 is a diagram of an operation result of the back propagation neural network model using the validation set according to embodiment 2 of the present invention.
Fig. 6 is a diagram of an operation result of the back propagation neural network model using the test set according to embodiment 2 of the present invention.
Fig. 7 is a diagram illustrating an overall operation result of the back propagation neural network model according to embodiment 2 of the present invention.
Fig. 8 is a page diagram of the wechat applet acquiring the user information in embodiment 2 of the present invention.
Fig. 9 is a community recommendation list page diagram of the user a in embodiment 2 of the present invention.
Fig. 10 is a community detail page diagram of the user a in embodiment 2 of the present invention.
Fig. 11 is a community recommendation list page diagram of the user B in embodiment 2 of the present invention.
Fig. 12 is a community detail page diagram of the user B in embodiment 2 of the present invention.
Fig. 13 is a community recommendation list page diagram of the user C in embodiment 2 of the present invention.
Fig. 14 is a community detail page diagram of the user C in embodiment 2 of the present invention.
Fig. 15 is a block diagram showing a configuration of a community selection recommendation system according to embodiment 3 of the present invention.
Fig. 16 is a block diagram of a computer device according to embodiment 4 of the present invention.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer and more complete, the technical solutions in the embodiments of the present invention will be described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some but not all embodiments of the present invention, and all other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the protection scope of the present invention.
Example 1:
as shown in fig. 1, the present embodiment provides a community recommendation method, which is mainly implemented by a cloud server, and includes the following steps:
s101, acquiring resident socio-economic attribute data and community attribute data.
The resident socioeconomic attribute data comprise gender, age, occupation, entertainment preference, telephone charge level, interest and the like, and are mainly acquired based on mobile phone signaling data which records the behavior track of each resident at each moment.
The statistical mode of the resident residence is as follows: 1) setting an observation period to be 21:00 to 8:00 of the next day; 2) the method comprises the steps that the number of seconds observed by a user in an observation time period every day is accumulated monthly, ranking is carried out, and the place with the highest ranking is taken as the residence of the user; 3) and the condition that the number of days of working days in one month exceeds 10 days is met.
The statistical mode of the workplace is as follows: 1) setting an observation period to be 9:00 to 17: 00; 2) the number of seconds observed in the observation time period on the working day of the user is subjected to monthly accumulation and ranking is carried out; taking the place with the highest rank as the employment place of the user; 3) And the condition that the number of days of working days in one month exceeds 10 days is met.
The statistical mode of the entertainment place is as follows: all the sites which reside for more than 1 hour on weekends and are not in work and live, wherein the longest resident site is defined as the recreation ground.
The coordinates of the residents' residence, workplace and entertainment place are associated with the nearest community, company enterprise and Point of Interest (POI) for rest and entertainment, and the community, the company and the rest and entertainment place where the residents are located are determined.
The community attribute data comprises facility attributes, environment attributes, location attributes, economic attributes and governance attributes, and the specific descriptions of the facility attributes, the environment attributes, the location attributes, the economic attributes and the governance attributes are as follows:
1) the facility attributes comprise education, medical treatment, commercial service, sports and leisure and public transportation facilities, the data of the facilities are represented by POI data acquired by calling Baidu or Goodpasture map open platform API, and each row of data comprises information such as name, major class, middle class, minor class, address, belonging province and city area, longitude and latitude coordinates and the like.
2) The environmental attributes comprise a green space and a water system, a Landsat TM/ETM/OLI remote sensing image is used as a main data source, remote sensing map data are obtained, the remote sensing map data are imported into ArcGIS software for processing, grid images of the water system and the green space are respectively extracted, and the grid images are converted into vector graphic data.
3) The economic attributes comprise the housing price and rent of the community, and related data can be crawled through websites of resident guests, original property and the like.
4) The location attribute data is the distance between the community and a Central Business Center (CBD), and specifically, the distance between each community and the CBD is calculated by using a neighbor analysis tool in ArcGIS software.
5) The treatment attributes comprise government treatment and resident participation, and data collection is carried out by accessing a website of the people's government, including but not limited to the number of living committees, the number of street government agencies, the expenditure of urban and rural communities, the expenditure of social security and employment, the expenditure of social welfare, the number of active government open contents, the number of department files, the dynamic street and town updating frequency, the times of activity development and lecture propaganda, the times of community care/entertainment activity organization, the times of community micro-improvement activities, the times of dispute handling and the like, so as to comprehensively evaluate the treatment level of the community.
And S102, constructing a community evaluation index system.
Dividing relevant indexes of the complete community into nine types of education, medical treatment, commercial service, public transportation, environment, economy, administration, location and sports leisure according to concepts of the complete community, related academic research, national standards and media reports, selecting nine items of education, medical treatment, commercial service, sports leisure, public transportation, environment, economy, location and administration as a criterion layer of an evaluation system according to the equal division standards of 'five minutes, ten minutes and fifteen minutes living circle' community service facilities and convenience service facilities in living areas in GB50180-2018 'planning and designing standards of urban living areas', subdividing the indexes of the facility layer into three evaluation indexes of fifty minutes, such as the related contents of the national standard about five minutes, ten minutes and fifteen minutes living circles, and subdividing the indexes of the facility layer (education, medical treatment, commercial service, location and administration), Public transport, etc.) are subdivided into indices such as the number of construction points in five minutes, ten minutes, and fifteen minutes.
In order to further reduce index system data for subsequent model training, a certain number of resident samples (the larger the sample size is, the better the sample size is) covering information of different sexes, ages, professions, incomes, education levels and the like are randomly obtained through Snowball Sampling (Snowball Sampling), the residents are asked to sort every two of nine criterion layer indexes and all indexes in all the criterion layer indexes, value assignment is carried out according to priority, and finally the index with the highest total score is listed as the highest priority index.
Then, based on an Analytic Hierarchy Process (AHP), yaahp software is used to calculate the weight values of each rule layer and each index inside the rule layer. The method specifically comprises five steps of establishing a hierarchical structure model, constructing a judgment matrix, performing hierarchical single sequencing, checking consistency and performing hierarchical total sequencing.
The first step is to divide the object into the highest layer, the middle layer and the lowest layer according to the decision target, the factor (decision criterion) and the relation between the decision objects, and draw a hierarchical structure diagram to build a hierarchical structure model.
The second step is to construct a decision matrix using a uniform matrix method, namely: all factors are not put together for comparison, but are compared with each other two by two; for this time, relative scales are adopted to reduce the difficulty of comparing different factors of the properties with each other as much as possible, so as to improve the accuracy.
The third step is that the feature vector corresponding to the maximum feature root λ max of the judgment matrix is normalized (the sum of each element in the vector is 1), and then is denoted as W, and the element of W is an ordering weight value of the relative importance of the same level element to the previous level factor, and the process is called level single ordering.
The fourth step introduces a random consistency index RI for measuring the CI size, defining a consistency ratio: when the consistency ratio CR is considered to be less than 0.1, the normalized feature vector of the judgment matrix can be used as the weight vector through the consistency check, otherwise, the judgment matrix is reconstructed and adjusted.
The fifth step is to calculate the weight values of relative importance of all factors of a certain level to the highest level (total target) from the highest level to the lowest level in turn, which is called total ranking of levels.
And finally, respectively obtaining nine criterion layer indexes and weight values of all evaluation indexes in all the criterion layer indexes, normalizing data of all the evaluation indexes in all the criterion layer indexes, multiplying the normalized data by the sum of the weights to obtain coefficients of the nine criterion layer indexes, recording the coefficients as nine indexes of education, medical treatment, commercial service, public transportation, environment, treatment and the like, multiplying and adding the nine indexes and corresponding weights (wherein economic and regional indexes are negative values) and normalizing the coefficients to obtain a complete community index which is used as an index for comprehensively evaluating the integrity of the community, and storing the nine indexes and the complete community index in a community database for further use, and providing guidance and reference for monitoring the community quality in a large range and fine granularity by the government.
S103, training by using a back propagation neural network model by using the resident socioeconomic attribute data and the community attribute data as influencing factors and the subentry indexes and the complete community indexes of each community as predicted indexes to obtain a trained model.
The Back Propagation Neural Network model (hereinafter referred to as "BP Neural Network") has strong nonlinear mapping capability, is suitable for solving the problem of complex internal mechanism, has high self-learning and self-adaption capability, generalization capability and fault-tolerant capability, can apply learning results to new knowledge, and can support and realize the function of community recommendation to residents.
The community recommendation algorithm used in this embodiment is to recommend a corresponding community for a user with a specific attribute based on preferences of different residents for various attributes of the community, take social and economic attribute data of the residents and the attribute data of the community as influencing factors, take subentry indexes and complete community indexes of various communities as predicted indexes, train by using a back propagation neural network model to obtain a trained model, and when the social and economic attributes of the user are input, predict a predicted value of each index of the community corresponding to the user through the trained model.
Specifically, the structure of the BP neural network model is as shown in fig. 2 and includes an input layer, a hidden layer, and an output layer, where a weight coefficient matrix from the input layer to the hidden layer is a first weight coefficient matrix denoted as w, a bias from the hidden layer is a first bias denoted as θ, a weight coefficient matrix from the hidden layer to the output layer is a second weight coefficient matrix denoted as v, and a bias from the output layer is a second bias denoted as Φ; forming a training set based on the resident socioeconomic attribute data, the community attribute data, and the subentry index and the complete community index of each community, and recording the training set as D { (X) if the training set has P samples1,T1),(X2,T2),…,(XP,TP) Each input sample comprising n elements, corresponding to n nodes of the input layer,
each target output sample contains m elements, corresponding to m nodes of the output layer,
the hidden layer has l nodes, and the input of the jth node is set as
The first bias is thetajOutput is
The activation function is denoted as f (x),
then
The calculation formula of (a) is as follows:
the output layer has m nodes, and the input of the kth node is set as
The second bias is phikOutput is
The actual value is
Then
The calculation formula of (a) is as follows:
the mean square error at sample point p is as follows:
the cumulative error is as follows:
the number of parameters to be determined in the whole network is as follows:
(n×l+l)+(l×m+m)＝(n+m+1)l+m
the updating formula of the parameters is as follows: e ← E + Δ E, where Δ E is updated according to the gradient descent strategy, the updating algorithm is divided into standard BP algorithm and cumulative BP algorithm, each based on E of a single samplepOr calculating according to the principle of minimizing the accumulated error.
In the aspect of algorithm implementation, a learning rate eta, a termination threshold value epsilon, a maximum step number N, a first weight coefficient matrix w, a second weight coefficient matrix v, a first bias theta and a second bias phi are set; inputting the training set into a BP neural network model, calculating input values and output values of each layer, calculating gradients of weight coefficients and offsets (standard BP algorithm)/required partial derivatives (accumulated BP algorithm), and updating a first weight coefficient matrix, a second weight coefficient matrix, a first offset and a second offset; and continuously iterating in the actual training to correct the first weight coefficient matrix, the second weight coefficient matrix, the first bias and the second bias to obtain a trained model.
And S104, acquiring socioeconomic attribute data input by the user, and predicting by using the trained model to obtain a predicted value of each index of the community corresponding to the user.
And S105, matching the predicted value of each index of the community corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result.
And S106, outputting a community recommendation list according to the matching result.
In the embodiment, a recommendation platform is formed based on a terminal and a cloud server, the terminal is used for acquiring social and economic attribute information input by a user and performing community recommendation, and the cloud server is used as a background and is used in a place where a BP neural network model and a cosine similarity algorithm are operated.
In order to realize the acquisition of socio-economic attribute information input by a user and the community recommendation, the terminal can adopt the modes of WeChat applet, APP, webpage and the like, the user can input information such as gender, age, education level, income level, occupation, working place and the like through interfaces such as WeChat applet, APP, webpage and the like, the socio-economic attribute information of the user is sent to a cloud server, the cloud server acquires the socio-economic attribute information of the user and then predicts the information by using a model after back propagation neural network training to obtain the predicted value of each index of a community corresponding to the user, the predicted value is used as a first sentence, nine indexes of education, medical treatment and the like and a complete community index of each community in an original community database are used as a second sentence, the matching degree of the first sentence and the second sentence is calculated by a similarity calculation method, and sorting the matching degree from high to low, and recording the result as a first-layer community recommendation list.
The cosine similarity is a measure for measuring the difference between two individuals by using cosine values of included angles of two vectors in a vector space, the closer the cosine values are to 1, the closer the included angles are to 0 degree, namely the more similar the two vectors are, the cosine similarity is called, for a non-right-angled triangle, a, b and c respectively represent three sides of the triangle, and a cosine calculation formula of an included angle theta between the sides a and b is as follows:
in the triangle represented by the vector, assuming that the a vector is (x1, y1) and the b vector is (x2, y2), the cosine of the angle θ between the vector a and the vector b is calculated as:
if the vectors a and b are not two-dimensional but n-dimensional, the above cosine calculation method is still correct, and assuming that a and b are two n-dimensional vectors, the cosine of the included angle θ between a and b is equal to:
the first sentence is denoted as { x1,…,x9The second sentence is denoted as { y }1,…,y9Imagine sentences a and B as two line segments in space, both from the origin ([0, 0. ].]) Starting, pointing to different directions; an included angle is formed between the two line segments, if the included angle is 0 degree, the direction is the same, the line segments are overlapped, and the texts represented by the two vectors are completely equal; if the included angle is 90 degrees, the right angle is formed, and the directions are completely dissimilar; if the included angle is 180 degreesMeaning the directions are exactly opposite. Therefore, the similarity of the vectors can be judged by the size of the angle, i.e. the smaller the angle, the closer the value of cos (θ) is to 1, and the more similar the first sentence and the second sentence. The concrete formula is as follows:
and then further defining a range by taking the working place as the center of a circle and the acceptable commuting distance as the radius based on the working place filled by the user and the acceptable commuting distance, screening out only the community of the first-layer recommendation list in the range, and returning the information to a community recommendation interface of the WeChat applet/APP/webpage, thereby providing community recommendation for the user inputting the specific attribute, and further supplementing the influence factors of the original BP neural network model based on additional information provided by the user to perform more accurate community recommendation based on the socio-economic attribute information of more community residents.
It should be noted that while the method operations of the above-described embodiments are described in a particular order, this does not require or imply that these operations must be performed in that particular order, or that all of the illustrated operations must be performed, to achieve desirable results. Rather, the depicted steps may change the order of execution. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions.
Example 2:
in the embodiment, a Guangzhou city is used as an application example, and a WeChat applet is used as a recommendation platform for acquiring data and displaying a community recommendation result to verify the community recommendation method, which comprises the following steps:
1) and acquiring resident socioeconomic attribute data and community attribute data.
The resident socioeconomic attributes comprise gender, age, occupation, entertainment preference, telephone charge level and interest, the resident socioeconomic attribute data are mainly obtained based on mobile phone signaling data, the mobile phone signaling data are intelligent footprint data provided by Unicom China, the intelligent footprint data comprise gender, age, occupation, entertainment preference, telephone charge level and interest, the behavior track of each resident at each moment is recorded, and the residence, the working place and the entertainment place of the resident are respectively defined based on the following standards. The statistical mode of the residence is as follows: 1) setting an observation period to be 21:00 to 8:00 of the next day; 2) the number of seconds observed by a user in an observation time period every day is accumulated monthly, ranking is carried out, and the place with the highest ranking is taken as the residence of the user; 3) and the condition that the number of days of working days exceeds 10 days in one month is met; the statistical mode of the workplace is as follows: 1) Setting an observation period to be 9:00 to 17: 00; 2) accumulating monthly data of the seconds observed in the observation time period on the working day of the user, and ranking; taking the place with the highest rank as the employment place of the user; 3) and the condition that the number of days of working days exceeds 10 days in one month is met; the statistical mode of the entertainment place is as follows: all the sites which reside for more than 1 hour on weekends and are not in work and live, wherein the longest resident site is defined as the recreation ground. And (3) associating the coordinates of the residence, the work place and the entertainment place with the nearest community, company enterprise and recreation and entertainment POI, and determining the community, the employment company and the entertainment place of the resident.
The community attributes comprise five aspects of facility, environment, economy, location and governance attributes. 1) The facility attributes comprise education, medical treatment, commercial service, sports leisure and public transportation facilities, the data of the facilities crawl 56.14 ten thousand POI data of the wide-state city range of No. 4 months in 2018 by calling a Gade map open platform API (https:// lbs. amap. com/API/webservice/guide/API/search), and the data cover 15 categories of catering, shopping, company enterprises, transportation facility service and the like, and each row of data comprises information of name, large category, medium category, small category, address, belonging province area, longitude and latitude coordinates and the like. 2) Environmental attributes include greenfield and water systems. The used satellite remote sensing data comprise remote sensing images of Landsat 8 satellite digital products in 2018, 3, month and 14 days (the cloud cover is less than 5%, and the spatial resolution is 30 m). And (4) importing the grid images into ArcGIS software for processing, respectively extracting water body and green space grid images, and converting the grid images into vector graphics data. 3) Economic attributes include community housing prices and rentals. The house price and rent data were crawled through the resident (https:// guangzhou.anjuke.com /), native (https:// gz.centreanet.com/. 4) Location attribute is the social Business distance from the Central Business District (CBD). Specifically, a Zhujiang New City of a river area is set as CBD in ArcGIS software, a specific space range is defined, and a neighbor analysis tool is used for calculating the distance between each community and the CBD. 5) The treatment attributes include both government treatment and resident participation. Data collection is carried out by visiting the website of the people's government in each administrative district of Guangzhou city, including the number of living committees, the number of street government institutions, the expenditure of urban and rural communities, the social security and employment expenditure, the expenditure of social welfare, the number of active government open contents, the number of department files, the dynamic street and town updating frequency, the times of developing activities and lecture propaganda, the times of community care/entertainment organization, the times of community micro-improvement activities, the times of dispute treatment and the like, so as to comprehensively evaluate the treatment level of the community
2) And constructing a community evaluation index system.
The method comprises the steps of dividing relevant indexes of the complete community into nine types including education, medical treatment, commercial service, sports and leisure, public transportation, environment, economy, location and treatment according to concepts of the complete community, related academic research and national standards, selecting nine types including education, medical treatment, commercial service, sports and leisure, public transportation, environment, economy, location and treatment as a criterion layer of an evaluation system according to the dividing standards of 'five-minute, ten-minute and fifteen-minute living circle' community service facilities and convenience service facilities in a residential area in GB50180-2018 'urban residential area planning and designing standard', and subdividing fifty three evaluation indexes downwards. For example, according to the national standard, the life circle of five minutes, ten minutes and fifteen minutes, the indexes (education, medical treatment, commercial service, public transportation and the like) at the facility level are subdivided into the indexes of the number of construction points or the nearest distance in five minutes, ten minutes and fifteen minutes.
In order to make the evaluation result more scientific, 101 network questionnaires including 86 effective questionnaires were obtained by Snowball Sampling (Snowball Sampling). After 7 investigators are used as basic investigation units to perform investigation, the investigators are asked to perform investigation on about 15 other investigation objects in the relationship network, the investigation objects are asked to compare every two of nine criteria layer indexes and all indexes in all criteria layer indexes, assignment is performed according to priority, the index with the highest total score is listed as the most preferred index, then the analytic hierarchy process is used for extracting the weight values of all indexes, and specific numerical values of nine criteria layer indexes and complete community indexes of Guangzhou city communities are calculated.
3) And (5) constructing and training a BP neural network model.
The influencing factors and all indexes of the community are normalized, Matlab is adopted as a software platform, and the construction, training and simulation of the BP neural network model are realized by programming the neural network toolbox function. The network is created by applying newff, the number of the hidden neurons is set to be 10, and the function adopted for training the network performance is a train default function which is suitable for a medium-sized network, has large demand on a memory and has high convergence rate; setting the training times to 10000 times, the convergence error to 0.0000001, the learning rate to 0.01, the momentum factor to 0.9, the display interval times to 25, and selecting a BP neural network model with a three-layer structure as shown in FIG. 3; the collected sample was divided into three fractions: randomly selected 70% of the data from the samples as training data, 15% of the data as test samples, and the remaining 15% of the data as validation samples.
In the operation results of the BP neural network model, the mean square deviations of the training set, the verification set and the test set are 0.0029, 0.0031 and 0.0040, the R values are 0.95, 0.94 and 0.93, respectively, the overall R value is 0.94, the fitting degree is good, and the operation results of the training set, the verification set, the test set and the overall operation results are shown in fig. 4 to 7, respectively.
4) Community recommendation based on cosine similarity algorithm.
And further inputting user sample data, predicting the predicted value of each index of the community through the trained model, calculating by using a cosine similarity recommendation algorithm to obtain a community recommendation result aiming at the user sample, and demonstrating in a WeChat small program mode.
The user sample is three samples randomly extracted from a network questionnaire and recorded as a user A, B, C, a WeChat applet obtains a user information page as shown in FIG. 8, specific information obtained through the applet is shown in the following table 1, and the specific information is input into a trained model to obtain a predicted value of each index of the community as shown in the following table 2; then, a cosine correlation function and a sorting function of python are called in the cloud server, a community recommendation list corresponding to three target users is obtained and is presented in the form of a WeChat applet page, wherein a community recommendation list page and a community detail page of a user A are respectively shown in fig. 9 and fig. 10, a community recommendation list page and a community detail page of a user B are respectively shown in fig. 11 and fig. 12, and a community recommendation list page and a community detail page of a user C are respectively shown in fig. 13 and fig. 14.
TABLE 1 socio-economic Attribute information for user A, B, C
TABLE 2 prediction of Community index of user A, B, C
Example 3:
as shown in fig. 15, the present embodiment provides a recommendation system for community selection, which includes an obtaining module 1501, a constructing module 1502, a training module 1503, a predicting module 1504, a matching module 1505, and a recommending module 1506, where the specific functions of each module are as follows:
the obtaining module 1501 is configured to obtain the social and economic attribute data of the residents and the community attribute data.
A construction module 1502 for constructing a community evaluation index system; the community evaluation index system comprises a subentry index and a complete community index of each community.
The training module 1503 is used for training by using a back propagation neural network model by using the resident socioeconomic attribute data and the community attribute data as influencing factors and using the subentry index and the complete community index of each community as predicted indexes to obtain a trained model.
The prediction module 1504 is used for acquiring the socioeconomic attribute data input by the user, and predicting by using the trained model to obtain the predicted value of each index of the community corresponding to the user.
The matching module 1505 is used for matching the predicted value of each community index corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result.
And the recommending module 1506 is used for outputting a community recommendation list according to the matching result.
It should be noted that the system provided in this embodiment is only illustrated by the division of the functional modules, and in practical applications, the functions may be distributed by different functional modules according to needs, that is, the internal structure is divided into different functional modules to complete all or part of the functions described above.
Example 4:
the present embodiment provides a computer device, which is a cloud server, as shown in fig. 16, and includes a processor 1602, a memory and a network interface 1603 connected by a system bus 1601, the processor is used for providing computing and control capability, the memory includes a nonvolatile storage medium 1604 and an internal memory 1605, the nonvolatile storage medium 1604 stores an operating system, computer programs and a database, the internal memory 1605 provides an environment for the operating system and the computer programs in the nonvolatile storage medium to run, and when the processor 1602 executes the computer programs stored in the memory, the recommended method of the above embodiment 1 is implemented as follows:
acquiring resident socioeconomic attribute data and community attribute data;
constructing a community evaluation index system; the community evaluation index system comprises a subentry index and a complete community index of each community;
using the resident socioeconomic attribute data and the community attribute data as influencing factors, using the subentry index and the complete community index of each community as a predicted index, and training by using a back propagation neural network model to obtain a trained model;
acquiring socioeconomic attribute data input by a user, and predicting by using the trained model to obtain a predicted value of each index of the community corresponding to the user;
matching the predicted value of each index of the community corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result;
and outputting a community recommendation list according to the matching result.
Example 5:
the present embodiment provides a storage medium, which is a computer-readable storage medium, and stores a computer program, and when the computer program is executed by a processor, the recommendation method of embodiment 1 is implemented as follows:
acquiring resident socioeconomic attribute data and community attribute data;
constructing a community evaluation index system; the community evaluation index system comprises a subentry index and a complete community index of each community;
using the resident socioeconomic attribute data and the community attribute data as influencing factors, using the subentry index and the complete community index of each community as a predicted index, and training by using a back propagation neural network model to obtain a trained model;
acquiring socioeconomic attribute data input by a user, and predicting by using the trained model to obtain a predicted value of each index of the community corresponding to the user;
matching the predicted value of each index of the community corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result;
and outputting a community recommendation list according to the matching result.
It should be noted that the computer readable storage medium of the present embodiment may be a computer readable signal medium or a computer readable storage medium or any combination of the two. The computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
In conclusion, the community evaluation index system comprises the subentry index and the complete community index of each community, the subentry index and the complete community index of each community are used as influencing factors, the subentry index and the complete community index of each community are used as predicted indexes, a back propagation neural network model is used for training, the prediction is carried out through the trained model and is matched with the subentry index and the complete community index of each community, a recommendation scheme can be provided for community selection of citizens according to a matching result, the community evaluation index system is beneficial to providing research foundation and methodology tests for planning and construction of the complete community, and also provides guidance for government departments and enterprises to accurately grasp customer images and decision and practice for participating in urban space and service reconstruction, and provides reference for community selection behavior of citizens.
The above description is only for the preferred embodiments of the present invention, but the protection scope of the present invention is not limited thereto, and any person skilled in the art can substitute or change the technical solution of the present invention and the inventive concept thereof within the scope of the present invention.
Claims (10)
1. A community recommendation method, the method comprising:
acquiring resident socioeconomic attribute data and community attribute data;
constructing a community evaluation index system; the community evaluation index system comprises a subentry index and a complete community index of each community;
using the resident socioeconomic attribute data and the community attribute data as influencing factors, using the subentry index and the complete community index of each community as a predicted index, and training by using a back propagation neural network model to obtain a trained model;
acquiring socioeconomic attribute data input by a user, and predicting by using the trained model to obtain a predicted value of each index of the community corresponding to the user;
matching the predicted value of each index of the community corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result;
and outputting a community recommendation list according to the matching result.
2. The recommendation method according to claim 1, wherein the constructing of the community evaluation index system specifically comprises:
for each community, education, medical treatment, commercial service, sports and leisure, public transportation, environment, economy, location and management are selected as criteria layer indexes of an evaluation system, and are subdivided downwards into fifty-three evaluation indexes;
calculating nine criteria layer indexes and weight values of evaluation indexes in the criteria layer indexes by using an analytic hierarchy process, wherein the calculation method specifically comprises the following steps: acquiring a certain amount of resident samples randomly through snowball type sampling, and requesting each resident to sort every two of nine criterion layer indexes and importance degrees of each evaluation index in each criterion layer index, integrating opinions of all visited residents, sorting every two of the nine criterion layer indexes and each evaluation index subdivided by the nine criterion layer indexes, assigning values according to priorities, listing the highest total score as the most preferred index, and calculating weight values of the nine criterion layer indexes and each evaluation index in each criterion layer index;
normalizing the data of each evaluation index in each criterion layer index, multiplying the data by the sum of weights to obtain the coefficients of nine criterion layer indexes, and taking the coefficients as the subentry indexes of nine communities;
and multiplying and adding the nine indexes and the corresponding weights, and performing normalization processing to obtain a complete community index which is used as an index for comprehensively evaluating the community.
3. The recommendation method according to claim 2, wherein the calculating of the nine criterion layer indicators and the weight values of the evaluation indicators in the criterion layer indicators based on the analytic hierarchy process specifically includes:
dividing the decision into a highest layer, a middle layer and a lowest layer according to the decision target, the considered factors and the interrelation among decision objects, and drawing a hierarchical structure diagram to establish a hierarchical structure model;
constructing a judgment matrix by using a consistent matrix method;
the feature vector corresponding to the maximum feature root of the judgment matrix is normalized and recorded as W, and the element of W is a sequencing weight of the relative importance of the same-level element to the previous-level factor;
introducing a random consistency index RI, defining a consistency ratio CR to be CI/RI, if the consistency ratio CR is less than 0.1, using a normalized feature vector of a judgment matrix as a weight vector through consistency test, otherwise, reconstructing the judgment matrix for adjustment;
and sequentially calculating the weight values of all factors of a certain level relative to the relative importance of the highest level from the highest level to the lowest level so as to obtain nine criteria level indexes and the weight values of all evaluation indexes in all criteria level indexes.
4. The recommendation method according to claim 1, wherein the back propagation neural network model comprises an input layer, a hidden layer and an output layer, the weight coefficient matrix from the input layer to the hidden layer is a first weight coefficient matrix, the bias of the hidden layer is a first bias, the weight coefficient matrix from the hidden layer to the output layer is a second weight coefficient matrix, and the bias of the output layer is a second bias;
the method comprises the following steps of training by using a back propagation neural network model by using resident socioeconomic attribute data and community attribute data as influencing factors and using subentry indexes and complete community indexes of various communities as predicted indexes to obtain a trained model, and specifically comprises the following steps:
the method comprises the following steps of (1) forming a training set by taking resident socioeconomic attribute data and community attribute data as influencing factors and taking the subentry index and the complete community index of each community as a predicted index;
setting a learning rate, a termination threshold, a maximum step number, a first weight coefficient matrix, a second weight coefficient matrix, a first bias and a second bias;
inputting the training set into a back propagation neural network model, calculating the input value and the output value of each layer, calculating the gradient/required partial derivative of the weight coefficient and the bias, and updating a first weight coefficient matrix, a second weight coefficient matrix, a first bias and a second bias;
and continuously iterating to correct the first weight coefficient matrix, the second weight coefficient matrix, the first bias and the second bias to obtain the trained model.
5. The recommendation method of claim 4, wherein the training set is denoted as D { (X)1,T1),(X2,T2),…,(XP,TP) }; wherein each input sample comprises n elements, corresponding to n nodes of the input layer,
each target output sample contains m elements, corresponding to m nodes of the output layer,
the hidden layer has l nodes, and the input of the jth node is set as
The first bias is thetajOutput is
The activation function is noted as f (x),
then
The calculation formula of (a) is as follows:
the output layer has m nodes, and the input of the kth node is set as
The second bias is phikOutput is
The actual value is
Then
The calculation formula of (a) is as follows:
the mean square error at sample point p is as follows:
the cumulative error is as follows:
the number of parameters to be determined in the whole network is as follows:
(n×l+l)+(l×m+m)＝(n+m+1)l+m
the updating formula of the parameters is as follows: e ← e + Δ e, where Δ e is updated according to a gradient descent strategy.
6. The recommendation method according to claim 1, wherein the matching of the predicted value of each community index corresponding to the user with the subentry index and the complete community index of each community to obtain the matching result specifically comprises:
taking the predicted value of each community index corresponding to the user as a first sentence, and taking the subentry index and the complete community index of each community as a second sentence;
calculating the matching degree of the first sentence and the second sentence by a cosine similarity algorithm, which is as follows:
wherein x isiRepresenting individual word vectors, y, in a first sentenceiRepresenting the respective word vectors in the second sentence, cos (theta) representing the cosine similarity of the angle theta.
7. The recommendation method according to any one of claims 1 to 6, wherein the resident socio-economic attribute data includes gender, age, occupation, entertainment preference, telephone charge level and hobbies, and the resident socio-economic attribute data is acquired based on mobile phone signaling data;
the community attribute data comprises facility attributes, environment attributes, location attributes, economic attributes and governance attributes, the facility attributes comprise education, medical treatment, commercial service, sports leisure and public transportation facilities, the environment attributes comprise greenbelts and water systems, the economic attributes comprise community housing prices and rent, the location attributes use the distance representation of the community from a central business center, and the governance attributes comprise government governance and resident participation.
8. A recommendation system for community selection, the system comprising:
the acquisition module is used for acquiring the social and economic attribute data of residents and the community attribute data;
the building module is used for building a community evaluation index system; the community evaluation index system comprises a subentry index and a complete community index of each community;
the training module is used for training by using a back propagation neural network model by taking the resident socioeconomic attribute data and the community attribute data as influencing factors and the subentry index and the complete community index of each community as predicted indexes to obtain a trained model;
the prediction module is used for acquiring socioeconomic attribute data input by a user, and predicting by using the trained model to obtain a predicted value of each index of the community corresponding to the user;
the matching module is used for matching the predicted value of each index of the community corresponding to the user with the subentry index and the complete community index of each community to obtain a matching result;
and the recommending module is used for outputting a community recommending list according to the matching result.
9. A computer device comprising a processor and a memory for storing a program executable by the processor, wherein the processor implements the recommendation method of any one of claims 1-7 when executing the program stored in the memory.
10. A storage medium storing a program, wherein the program, when executed by a processor, implements the recommendation method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011142989.1A CN112380425B (en) | 2020-10-23 | 2020-10-23 | Community recommendation method, system, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011142989.1A CN112380425B (en) | 2020-10-23 | 2020-10-23 | Community recommendation method, system, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112380425A true CN112380425A (en) | 2021-02-19 |
| CN112380425B CN112380425B (en) | 2023-11-14 |
Family
ID=74580755
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011142989.1A Active CN112380425B (en) | 2020-10-23 | 2020-10-23 | Community recommendation method, system, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112380425B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113010578A (en) * | 2021-03-22 | 2021-06-22 | 华南理工大学 | Community data analysis method and device, community intelligent interaction platform and storage medium |
| CN114399202A (en) * | 2022-01-17 | 2022-04-26 | 青岛文达通科技股份有限公司 | Big data visualization system for urban community |
| CN115394394A (en) * | 2022-10-27 | 2022-11-25 | 曹县人民医院 | Resident health service reservation method and system based on big data processing technology |
| CN115713145A (en) * | 2022-11-04 | 2023-02-24 | 哈尔滨工业大学（深圳） | Village and town community efficiency prediction method and device, computer equipment and storage medium |
| CN116822727A (en) * | 2023-06-16 | 2023-09-29 | 深圳慧锐通智能技术股份有限公司 | Smart community cloud platform-based refined community management method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015074428A1 (en) * | 2013-11-22 | 2015-05-28 | 华为技术有限公司 | Neural network system, and image parsing method and device based on same |
| WO2017062194A1 (en) * | 2015-10-09 | 2017-04-13 | Mastercard International Incorporated | Method and system for providing a housing recommendation |
| CN106895975A (en) * | 2017-01-17 | 2017-06-27 | 苏州大学 | Method for Bearing Fault Diagnosis based on Stacked SAE deep neural networks |
| CN109767270A (en) * | 2019-01-17 | 2019-05-17 | 建信养老金管理有限责任公司 | The old information recommendation method of housing support and system are deposited based on artificial intelligence |
| CN110990542A (en) * | 2019-10-11 | 2020-04-10 | 北京华跃博弈科技有限公司 | Client requirement matching system of real estate agency service platform |
| CN111274476A (en) * | 2020-01-16 | 2020-06-12 | 长春每房科技有限公司 | Room source matching method, device and equipment based on face recognition and storage medium |
-
2020
- 2020-10-23 CN CN202011142989.1A patent/CN112380425B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015074428A1 (en) * | 2013-11-22 | 2015-05-28 | 华为技术有限公司 | Neural network system, and image parsing method and device based on same |
| WO2017062194A1 (en) * | 2015-10-09 | 2017-04-13 | Mastercard International Incorporated | Method and system for providing a housing recommendation |
| CN106895975A (en) * | 2017-01-17 | 2017-06-27 | 苏州大学 | Method for Bearing Fault Diagnosis based on Stacked SAE deep neural networks |
| CN109767270A (en) * | 2019-01-17 | 2019-05-17 | 建信养老金管理有限责任公司 | The old information recommendation method of housing support and system are deposited based on artificial intelligence |
| CN110990542A (en) * | 2019-10-11 | 2020-04-10 | 北京华跃博弈科技有限公司 | Client requirement matching system of real estate agency service platform |
| CN111274476A (en) * | 2020-01-16 | 2020-06-12 | 长春每房科技有限公司 | Room source matching method, device and equipment based on face recognition and storage medium |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113010578A (en) * | 2021-03-22 | 2021-06-22 | 华南理工大学 | Community data analysis method and device, community intelligent interaction platform and storage medium |
| CN113010578B (en) * | 2021-03-22 | 2024-03-15 | 华南理工大学 | Community data analysis method and device, community intelligent interaction platform and storage medium |
| CN114399202A (en) * | 2022-01-17 | 2022-04-26 | 青岛文达通科技股份有限公司 | Big data visualization system for urban community |
| CN115394394A (en) * | 2022-10-27 | 2022-11-25 | 曹县人民医院 | Resident health service reservation method and system based on big data processing technology |
| CN115713145A (en) * | 2022-11-04 | 2023-02-24 | 哈尔滨工业大学（深圳） | Village and town community efficiency prediction method and device, computer equipment and storage medium |
| CN115713145B (en) * | 2022-11-04 | 2024-06-11 | 哈尔滨工业大学（深圳） | Village and town community efficiency prediction method, village and town community efficiency prediction device, computer equipment and storage medium |
| CN116822727A (en) * | 2023-06-16 | 2023-09-29 | 深圳慧锐通智能技术股份有限公司 | Smart community cloud platform-based refined community management method and device |
| CN116822727B (en) * | 2023-06-16 | 2024-03-22 | 深圳慧锐通智能技术股份有限公司 | Smart community cloud platform-based refined community management method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112380425B (en) | 2023-11-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kang et al. | Understanding house price appreciation using multi-source big geo-data and machine learning | |
| Boeing et al. | New insights into rental housing markets across the United States: Web scraping and analyzing craigslist rental listings | |
| CN112380425B (en) | Community recommendation method, system, computer equipment and storage medium | |
| Oueslati et al. | Determinants of urban sprawl in European cities | |
| Martin et al. | Developing a flexible framework for spatiotemporal population modeling | |
| Mimis et al. | Property valuation with artificial neural network: the case of Athens | |
| Petrişor et al. | Applications of principal component analysis integrated with GIS | |
| Ünel et al. | Valuations of building plots using the AHP method | |
| Haas et al. | Untangling housing cost and transportation interactions: The location affordability index model—Version 2 (LAIM2) | |
| Shmueli et al. | Neural network analysis of travel behavior: evaluating tools for prediction | |
| Liu et al. | A generalized framework for measuring pedestrian accessibility around the world using open data | |
| Aydinoglu et al. | Implementing a mass valuation application on interoperable land valuation data model designed as an extension of the national GDI | |
| Cabello Eras et al. | Comparative study of the urban quality of life in Cuban first-level cities from an objective dimension | |
| Abildtrup et al. | Combining RP and SP data while accounting for large choice sets and travel mode–an application to forest recreation | |
| Lu et al. | Linking socioeconomic development, sea level rise, and climate change impacts on urban growth in New York City with a fuzzy cellular automata-based Markov chain model | |
| Boin et al. | What communicates quality to the spatial data consumer | |
| Li et al. | Integrating conformance and performance for the evaluation of urban planning implementation from a goal-oriented perspective | |
| Stang et al. | From human business to machine learning—methods for automating real estate appraisals and their practical implications | |
| Li et al. | Applied Research Methods in Urban and Regional Planning | |
| Deaconu et al. | Real estate valuation models performance in price prediction | |
| Chen | Addressing measurement error bias in GDP with nighttime lights and an application to infant mortality with Chinese county data | |
| Nasser et al. | Application of selected MCDM methods for developing a multi-functional framework for Eco-hotel planning in Yemen | |
| CN106779181A (en) | Method is recommended by a kind of medical institutions based on linear regression factor Non-negative Matrix Factorization model | |
| Correia et al. | Applying a structured simulation-based methodology to assess carpooling time–space potential | |
| Zhou et al. | GPS2space: An open-source python library for spatial measure extraction from GPS 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |