CN114529227B - Rural joyful comprehensive service platform based on big data and deep learning - Google Patents
Rural joyful comprehensive service platform based on big data and deep learning Download PDFInfo
- Publication number
- CN114529227B CN114529227B CN202210432686.6A CN202210432686A CN114529227B CN 114529227 B CN114529227 B CN 114529227B CN 202210432686 A CN202210432686 A CN 202210432686A CN 114529227 B CN114529227 B CN 114529227B
- Authority
- CN
- China
- Prior art keywords
- service
- portrait
- accurate
- data
- client
- 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.)
- Active
Links
- 238000013135 deep learning Methods 0.000 title claims abstract description 16
- 230000008676 import Effects 0.000 claims abstract description 10
- 239000013598 vector Substances 0.000 claims description 27
- 238000004364 calculation method Methods 0.000 claims description 16
- 230000006870 function Effects 0.000 claims description 15
- 239000000178 monomer Substances 0.000 claims description 13
- 230000004913 activation Effects 0.000 claims description 10
- 238000007726 management method Methods 0.000 claims description 9
- 230000007123 defense Effects 0.000 claims description 5
- 230000009977 dual effect Effects 0.000 claims description 5
- 230000002776 aggregation Effects 0.000 claims description 4
- 238000004220 aggregation Methods 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000000926 separation method Methods 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000002955 isolation Methods 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 claims description 3
- 230000000644 propagated effect Effects 0.000 claims description 3
- 125000004432 carbon atom Chemical group C* 0.000 claims description 2
- 238000013523 data management Methods 0.000 claims description 2
- 238000009826 distribution Methods 0.000 claims description 2
- 230000004927 fusion Effects 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 238000004080 punching Methods 0.000 claims description 2
- 230000005540 biological transmission Effects 0.000 claims 1
- 238000003384 imaging method Methods 0.000 claims 1
- 238000005516 engineering process Methods 0.000 abstract description 4
- 238000012423 maintenance Methods 0.000 abstract description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000008520 organization Effects 0.000 description 7
- 238000012549 training Methods 0.000 description 7
- 238000000034 method Methods 0.000 description 6
- 230000009471 action Effects 0.000 description 3
- 238000000513 principal component analysis Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000007418 data mining Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000006116 polymerization reaction Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 238000003892 spreading Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000005406 washing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- 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
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
-
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- 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
- G06Q50/26—Government or public services
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Development Economics (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Software Systems (AREA)
- Operations Research (AREA)
Abstract
The invention discloses a country happy comprehensive service platform based on big data and deep learning, which comprises a service data import module, a big data platform, a single accurate portrait application, an accurate single portrait, a server side service portrait and an accurate service push network, belongs to the technical field of artificial intelligence and big data, and particularly relates to a country happy comprehensive service platform based on big data and deep learning; the invention simultaneously pushes the latest policy to the potential target group and the related target group capable of handling the service to the service institution by utilizing the deep learning algorithm, thereby better serving country people, integrating platform data of all parties by utilizing the advanced big data technology, completely portraying individuals or groups, integrating social resources, and realizing the accurate maintenance and management of individuals or groups by government affairs and third party service institution gridders on the premise of meeting the individual demands and the individual services of the masses.
Description
Technical Field
The invention belongs to the technical field of artificial intelligence and big data, and particularly relates to a rural vogue accurate service platform based on big data and deep learning.
Background
Under the big background of the village joyful strategy, the management problem of the rural areas is particularly important, but in many rural laggard areas, vast farmers cannot know the preferential policies due to message blocking, so that the preferential policies enjoyed by the farmers in the aspects of education, medical treatment, government affairs, social security and the like are unclear and unknown, and the implementation of the policies is lagged.
The crowd portrayal is a marketing phrase, is one of the growing interest of product users, can help people to explore the reasons behind product index numbers, and the main step of constructing the crowd portrayal is to analyze user characteristics such as age, hobby, education degree, living environment, economic income and the like.
At present, the problem of data islanding exists for government affairs and other third-party service organizations in rural areas, data separation exists among all departments, all-round accurate portrait of peasants cannot be performed, so that service objects are difficult to understand comprehensively, latest preferential policies issued by relevant departments often need to be applied actively by the peasants, in addition, crowd portrait is difficult to realize for communities, enterprises, villages and towns and the like in rural areas, and accurate maintenance and management is difficult; for the masses, the periodic and real-time business handling problems of the masses of farmers are time-consuming and labor-consuming due to the separation of system data of each mechanism.
Disclosure of Invention
Technical problem to be solved
In order to solve the problems in the prior art, the invention provides a village happy comprehensive service platform based on big data and deep learning, which aims to integrate platform data of each party by using a leading-edge big data technology, completely figure individuals or groups, integrate social resources, realize accurate maintenance and management of the individuals or the groups by government affairs and third-party service organization network members on the premise of meeting individual demands and individual services of the groups, and simultaneously push the latest policies to potential target groups and related target groups capable of handling the services to service organizations by using a deep learning algorithm, thereby better serving village people.
(II) technical scheme
In order to achieve the purpose, the invention adopts the technical scheme that: the utility model provides a rural area joy integrated service platform based on big data and degree of depth study, includes that service data imports module, big data platform, the accurate application of portraying of monomer, accurate monomer portrays, server side service portrays, accurate service propelling movement network and the code application is swept to the polygon, service data imports module links to each other with big data platform, the accurate application of portraying of monomer and accurate service propelling movement network are connected with big data platform respectively, server side service portrays and accurate monomer portrays and accurate service propelling movement network both way junction, accurate monomer portrays are connected with big data platform, accurate service propelling movement network and the accurate application of portraying of monomer sweep the code application with the polygon and are connected.
Further, the leading-in module of service data includes government affair data and third party service data, the government affair data is leading-in by the manual work of government affair staff, third party service data is leading-in by the manual work of third party service staff, big data platform integrates government affair data and third party service institution data, big data platform adopts big data analysis, data mining, data washing, principal component analysis and data aggregation algorithm to rural area peasant household, enterprise, community and villages and towns in the support area generate label data to keep label data to big data platform, principal component analysis utilizes following correlation coefficient matrix to carry out principal component selection:
in the above formula, the first and second carbon atoms are,r ij is the ith row and the jth column element, x in the correlation coefficient matrix ki And x kj For the normalized label data, n is the dimension of each label data, and m represents the total data number of the label data; the big data platform shows the representation in the form of a label.
Furthermore, the service terminal service portrait comprises a bank service portrait, a community service portrait, a third party organization service portrait and a government service portrait, wherein the service terminal service portrait is obtained by accurately portraying various services provided by the service data import module, and is described in the form of label data; the big data platform further utilizes a machine learning algorithm to generate accurate single portrait from the label data, the accurate single portrait comprises a farmer portrait, an enterprise portrait, a village and town portrait and a community portrait, and the portrait is displayed in a label form after the big data platform integrates the data; the farmer portrait comprises farmer names, ages, household nationalities, professions, land existence, planting areas, types and respective planting areas of planted crops, family population number, low security, marriage and child and daughter number labels; the enterprise portrait comprises enterprise type, operation range and staff number labels; the village and town portrait comprises geographic positions, the total population number, male and female proportions, a population structure and a population GDP label; the community picture comprises a geographical position, the number of buildings and a cell name label.
Further, the accurate service push network comprises a client memory model, a server memory model, a double-association model and a network prediction end, wherein the client memory model, the server memory model and the double-association model are calculated in parallel, the respective output weights of the client memory model, the server memory model and the double-association model are connected to the output end of the network, the essence of the client memory model and the server memory model is a generalized linear model and is responsible for the connection between the memory service requirement and the service provision, the essence of the double-association model is a BP neural network and is responsible for recommending relevant services, the client memory model comprises a client input and a client output, and the client input is connected with the client output in a weight manner; the server-side memory model comprises a server-side input and a server-side output, and the server-side input is connected with the server-side output in a right way; the double-thinking model comprises a deep layer input, dense discrete features, a middle hidden layer and a double-thinking model output end, the dense discrete features are connected with the middle hidden layer in a weight mode, and the middle hidden layer is connected with the double-thinking model output in a weight mode; and the client output, the server output and the double-association model output end are connected with the network prediction end right.
Further, deep input of the double-association model comprises two categories of discrete features and continuous features, and Onehot coding is firstly required for the discrete features; performing discretization treatment on the continuous features by adopting box separation operation; processing the deep input by using an Embedding operation, so that extremely sparse vectors of the deep input are converted into low-dimensional dense vectors, the low-dimensional dense vectors are combined to become 1000-dimensional dense discrete features, the dense discrete features are spread forwards through weight vectors, enter an intermediate hidden layer and finally reach the output end of the double-association model, and the forward spreading calculation mode is as follows:
in the above formula, x is a dense discrete feature,w (0)andb (0)for the weight vector and offset between the dense discrete features and the intermediate hidden layer,w (1)andb (1)the weight vector and the offset inside the middle hidden layer,w (2)andb (2)for the weight vector and bias between the middle hidden layer to the output of the dual phantom, activation function ƒ is a Sigmoid activation function,α (3)values at the output of the dual phantom.
Further, the client input of the client memory model includes 3 cross features, and a vector formed by the 3 cross features input by the client is propagated forwards through the weight vector, and the calculation mode of the forward propagation is as follows:
in the above formula, x1An input vector input for the client is used,w 1andb 1weight vector and bias, activation function of client memory model respectivelyδThe function is activated for the Sigmoid and,y 1a value output for the client.
Further, a server input of the server memory model includes 3 cross features, and a vector formed by the 3 cross features input by the server is propagated forwards through a weight vector, and the specific calculation method is as follows:
in the above formula, x2Is input for the client side, and the client side inputs,w 2andb 2the weight vector and the bias of the memory model of the server side and the activation function are respectivelyδIs SThe function is activated by the igmoid in such a way that,y 2is the value at the output end of the double-ideal model.
Further, the calculation results y1, y2 and alpha of the client output, the server output and the double-association model output are finally obtained(3)Performing polymerization calculation, and obtaining final output at a network prediction end, wherein the polymerization calculation method comprises the following steps:
in the above-mentioned formula, the compound has the following structure,w 1、w2and w3Respectively, the client output, the dual-association model output and the connection weight coefficient between the server output and the network prediction end, the prediction output of the network prediction end uses the Relu activation function,Yis a binarized category label; the accurate service push network adopts strategies of off-line learning and on-line learning, the double association model adopts a stochastic gradient descent algorithm for training, and the client memory model and the server memory model adopt FTRL and L1 regular direction association training; accurate service propelling movement module, usable government affair data and peasant household's the service condition of handling simultaneously to peasant household and the corresponding service data of third party service mechanism propelling movement, inform on the one hand that the peasant household provides relevant preferential service, remind the service that the staff can handle to the peasant household on the one hand, and provide the form fill-in work of certain degree authority within range, also made things convenient for peasant household and third party staff when accomplishing to remind the peasant household to handle the business, and alleviateed staff's work load.
Preferably, the multi-role code scanning application uses a single accurate portrait application accurate service push network to generate four types of single two-dimensional codes by using a data fusion algorithm, wherein the four types of single two-dimensional codes are divided into farmer codes, enterprise codes, village and town codes and community codes, and related identity verification and services are facilitated.
Furthermore, the multi-role code scanning application adopts a data security isolation technology to enable each organization grid member to only see data in the authority range of the organization grid member, and tourists can view public data in a grid area by scanning codes; the householder and the family members can view the joint defense situation of ten households, third-party service data and family information in the authority besides the visitor; the third-party service gridding member scans codes through a server side of the third-party service gridding member, so that client basic portrait data in the authority range can be viewed, and other data which are agreed to be shared by other organizations can be viewed.
(III) advantageous effects
The invention relates to a village joy comprehensive service platform based on big data and deep learning, which aims to reduce the comprehensive social service cost in rural areas, ensure people in rural areas to enjoy convenient service brought by digital construction, analyze and aggregate data from multiple channels by utilizing big data technology, and accurately figure a single body in a data label mode The service function is perfected, basic treatment is integrated, convenience channels are unblocked, and digital villages and intelligent communities which enable management to be more intelligent and service to be more convenient are constructed.
Drawings
FIG. 1 is a flow chart of a rural voyage integrated service platform based on big data and deep learning according to the present invention;
fig. 2 is a schematic diagram of a neural network connection of the precision service push network according to the present invention.
Wherein, 1, government affair data, 2, third party service data, 3, big data platform, 4, single accurate portrait application, 6, accurate service push network, 7, rural credit service portrait, 8, community service portrait, 9, third party service portrait, 10, government affair service portrait, 11, farmer portrait, 12, enterprise portrait, 13, community portrait, 14, village and town portrait, 15, accurate single portrait, 16, multi-role code scanning application, 17, farmer code, 18, enterprise code, 19, village and town code, 20, community code, 21, service portrait, 22, service data import module, 23, deep layer input, 24, client input, 25, service input, 26, dense discrete characteristic, 27, middle hidden layer, 28, dual-linked thinking model output, 29, network forecast end, 30, single two-dimensional code, 31, client memory model, 32. a server memory model 33, a double-association model 34, a client output 35 and a server output.
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention.
Detailed Description
As shown in the figures 1 and 2, the invention utilizes a data import form to import government affair data 1 and third party service data 2, the large data platform 3 adopts data mining, data cleaning, principal component analysis and data aggregation algorithms to generate data labels for peasant households, enterprises, communities and villages and towns in a rural happy holding area, and stores the classified data labels to the large data platform 3, the large data platform 3 utilizes a machine learning algorithm to generate an accurate single portrait 15, and meanwhile, the large data platform 3 describes a rural credit agency service portrait 7, a community service portrait 8, a third party institution service portrait 9 and a government affair service portrait 10 in a service terminal service portrait 21 in the form of data labels.
Constructing a structure of the accurate service push network 6 and training the accurate service push network 6 aiming at specific services, wherein the forward propagation calculation mode of the duplex model 33 is as follows:
the calculation method of the client memory model 31 forward propagation is as follows:
the calculation method of the forward propagation of the server-side memory model 32 is as follows:
the calculation method for the network predicting end 29 to forward propagate includes:
building a network model according to a forward propagation calculation mode of the double-association model 33, the client memory model 31, the server memory model 32 and the network prediction end 29; the method comprises the steps of training an accurate service push network 6 by using a Tensorflow deep learning framework, wherein the accurate service push network 6 adopts strategies of off-line learning and on-line learning, a dual-association model 33 adopts a stochastic gradient descent algorithm for training, and a client memory model 31 and a server memory model 32 adopt FTRL and L1 regular direction combined training.
The accurate service push network 6 can be put into use after the construction and training of the accurate service push network 6 are completed, and the accurate service push network 6 carries out service push of a specific target and reverse push meeting transacted services according to the accurate monomer portrait 15 and the server side service portrait 21.
As an embodiment of the present invention, the precise monomer image 15 is exemplified by a farmer image 11, the server service image 21 is exemplified by a rural credit agency service image 7, the deep input 23 is from the farmer image 11, the client input 24 and the server input 25 are from the cross feature composed of the farmer image 11 and the rural credit agency service image 7; the accurate service push network 6 pushes the relevant services of the rural credit society to the peasant households, and meanwhile, the accurate service push network 6 pushes the peasant households meeting the service handling conditions to the rural credit society.
As an embodiment of the invention, accurate monomer representation 15 is exemplified by an enterprise representation 12, server service representation 21 is exemplified by a government service representation 10, deep input 23 is derived from enterprise representation 11, client input 24 and server input 25 are derived from cross-over features formed by cross-over transformation of enterprise representation 11 and government service representation 7; the accurate service push network 6 pushes relevant services of government departments to the enterprises, and meanwhile, the accurate service push network 6 pushes the enterprises meeting service handling conditions to the government departments.
For convenience of application of each organization, the multi-role code scanning application 16 generates 4 types of single two-dimensional codes 30 which comprise farmer codes 17, enterprise codes 18, village and town codes 19 and community codes 20 and facilitate related identity verification and services, the accurate service push network 6 continuously pushes related services to farmers, meanwhile, farmer portrait 11 is updated according to service handling conditions of the farmers, the big data platform 3 automatically updates portrait of the farmer 11 and server side service portrait 21 according to the service handling conditions pushed by the accurate service push network 6 and recommends new services according to government data 1, and meanwhile, the related farmer codes 17 are also correspondingly updated to facilitate related identity verification and services.
The rural happy precise service platform based on big data and deep learning adopts data security isolation to ensure that each organization grid member can only see data in the authority range, tourists can check public data in a grid area by scanning codes, such as government grid managers, third-party service organization managers, appointed handling, related business appointment handling of the third-party service organizations, on-line job seeking and finding work, epidemic situation prevention and control registration and hometown return reporting and standby information are carried out; the farmer and family members can view the joint defense situation of ten households, third-party service data and family information of the farmer and family members besides the visitor; the third-party service gridding personnel scans the basic portrait data of the client in the authority range of the third-party service gridding personnel through the management end of the third-party service gridding personnel and other data which are agreed to be shared by other organizations.
As an embodiment of the invention, bank staff apply for checking financial data and code scanning business conversion analysis of all customers after analysis and integration through a big data platform through own client, and each gridder collects updated basic data.
As an embodiment of the invention, a government affair gridder can check population distribution conditions and ten-family joint defense conditions through code scanning of a management terminal, can carry out card-punching and visiting management, and carries out basic data management on line; all new data are generated and then returned to the big data platform for iterative analysis, and the accurate single image 15 is updated in real time to form a closed loop for iterative updating of data.
The specific working process of the invention is described above, and the steps are repeated when the device is used next time.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
The present invention and its embodiments have been described above, and the description is not intended to be limiting, and the drawings are only one embodiment of the present invention, and the actual structure is not limited thereto. In summary, those skilled in the art should appreciate that they can readily use the disclosed conception and specific embodiments as a basis for designing or modifying other structures for carrying out the same purposes of the present invention without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (4)
1. A rural happy comprehensive service platform based on big data and deep learning is characterized in that: the system comprises a service data import module (22), a big data platform (3), a single accurate portrait application (4), an accurate single portrait (15), a server side service portrait (21), an accurate service push network (6) and a multi-angle color code scanning application (16), wherein the service data import module (22) is connected with the big data platform (3), the single accurate portrait application (4) and the accurate service push network (6) are respectively connected with the big data platform (3), the server side service portrait (21) and the accurate single portrait (15) are in two-way connection with the accurate service push network (6), the accurate single portrait (15) is connected with the big data platform (3), and the accurate service push network (6) and the single accurate portrait application (4) are connected with the multi-angle color code scanning application (16); the service-side service image (21) comprises a rural credit agency service image (7), a community service image (8), a third-party institution service image (9) and a government service image (10), the service-side service image (21) is obtained by accurately imaging various services provided by a service data import module (22), and the service-side service image (21) is described in the form of label data; the big data platform (3) further utilizes a machine learning algorithm to generate accurate single portrait (15) from the label data, the accurate single portrait (15) comprises a farmer portrait (11), an enterprise portrait (12), a community portrait (13) and a village and town portrait (14), and the portrait is displayed in a label form after the big data platform (3) integrates data; the farmer portrait (11) comprises farmer names, ages, household nationalities, professions, whether land exists or not, planting areas, types of planted crops, respective planting areas, family population numbers, whether low insurance exists or not, whether marriage exists or not and labels of numbers of children and women; the enterprise image (12) comprises enterprise type, operation range and staff number labels; the community picture (13) comprises a geographical position, the number of buildings and a cell name label; the village and town portrait (14) comprises a geographic position, a total population number, a male-female ratio, a population structure and a per-capita GDP label; the accurate service push network (6) comprises a client memory model (31), a server memory model (32), a double-association model (33) and a network prediction end (29), wherein the client memory model (31), the server memory model (32) and the double-association model (33) are calculated in parallel, the essence of the client memory model (31) and the server memory model (32) is a generalized linear model, the essence of the double-association model (33) is a BP neural network, the client memory model (31) comprises a client input (24) and a client output (34), and the client input (24) is connected with the client output (34) in a right way; the server memory model (32) comprises a server input (25) and a server output (35), and the server input (25) and the server output (35) are connected in right; the double-associative model (33) comprises a deep layer input (23), dense discrete features (26), an intermediate hidden layer (27) and a double-associative model output end (28), wherein the dense discrete features (26) are connected with the intermediate hidden layer (27) in a weight mode, and the intermediate hidden layer (27) is connected with the double-associative model output end (28) in a weight mode; the client output (34), the server output (35) and the double-association model output end (28) are connected with the network prediction end (29) in a right-weighted mode; the deep input (23) of the double-linkage model (33) comprises two categories of discrete features and continuous features, and Onehot coding is firstly required for the discrete features; performing discretization treatment on the continuous features by adopting box separation operation; processing the deep input (23) using an Embedding operation to convert the extremely sparse vectors of the deep input (23) into dense vectors of low dimensionality which are combined to become dense discrete features (26) of 1000 dimensionalities, the dense discrete features (26) being propagated forward via weight vectors into intermediate hidden layers (27) and finally to a double-associative model output (28), the forward propagation being calculated as:
in the above formula, x is a dense discrete feature (26),w (0)andb (0)for the weight vectors and offsets between the dense discrete features (26) and the intermediate hidden layer (27),w (1)andb (1)is the weight vector and the offset inside the middle hidden layer (27),w (2)andb (2)for the weight vector and bias between the middle hidden layer to the dual phantom output (28), the activation function ƒ is a Sigmoid activation function,α (3)is the value at the output (28) of the double-associative model; the client input (24) of the client memory model (31) comprises 3 cross features, a vector formed by the 3 cross features of the client input (24) is transmitted forwards through a weight vector, and the calculation mode of the forward transmission is as follows:
in the above formula, x1An input vector for the client input (24),w 1andb 1a weight vector and a bias, respectively, activation function of the client memory model (31)δThe function is activated for the Sigmoid and,y 1outputting (34) the value for the client; the server input (25) of the server memory model (32) comprises 3 cross features, and a vector formed by the 3 cross features of the server input (25) is transmitted forwards through a weight vector, and the specific calculation mode is as follows:
in the above formula, x2For the client input (24),w 2andb 2weight vector and bias, activation function of the server memory model (32) respectivelyδThe function is activated for the Sigmoid and,y 2the values at the outputs (28) of the dual phantom are used.
2. The rural happy integrated service platform based on big data and deep learning of claim 1, wherein: and finally, performing aggregation calculation on the calculation results y1, y2 and alpha (3) of the client output (34), the server output (35) and the dual-association model output end (28), and obtaining final output at the network prediction end (29), wherein the aggregation calculation method comprises the following steps:
in the above formula, the first and second carbon atoms are,w 1、w2and w3Respectively, the client output, the dual-association model output and the connection weight coefficient between the server output and the network prediction end, the prediction output of the network prediction end uses Relu activation function,Yis a binarized category label; the accurate service push network (6) adopts strategies of off-line learning and on-line learning, the dual-association model (33) is trained by adopting a stochastic gradient descent algorithm, and the client memory model (31) and the server memory model (32) are jointly trained by adopting FTRL and L1 regular direction.
3. The rural voyage integrated service platform based on big data and deep learning of claim 2, wherein: code application (16) are swept to many roles combines that the accurate picture of monomer is used (4) and accurate service propelling movement network (6) utilize data fusion algorithm to generate four types of monomer two-dimensional code (30), monomer two-dimensional code (30) include peasant household code (17), enterprise code (18), villages and small towns code (19) and community code (20).
4. The village happy comprehensive service platform based on big data and deep learning of claim 3, characterized in that: the multi-role code scanning application (16) adopts data security isolation to ensure that a gridder can only see data in the authority range of the gridder, a tourist code can only view public data in a gridding area, and farmers and family members can view ten-family joint defense conditions and family information in the authority besides the tourist code; the third-party service gridding personnel can check recommended customer basic portrait data in the authority range and other service data which are agreed to be shared by other organizations by scanning codes through own client; the government affair gridder can check key population distribution conditions and ten-family joint defense conditions through code scanning of the management terminal, can carry out card-punching and visiting management and can carry out basic data management on line; all new data are generated and then returned to the big data platform (3) for iterative analysis, and the accurate monomer portrait (15) is updated in real time to form a closed loop for iterative updating of data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210432686.6A CN114529227B (en) | 2022-04-24 | 2022-04-24 | Rural joyful comprehensive service platform based on big data and deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210432686.6A CN114529227B (en) | 2022-04-24 | 2022-04-24 | Rural joyful comprehensive service platform based on big data and deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114529227A CN114529227A (en) | 2022-05-24 |
| CN114529227B true CN114529227B (en) | 2022-07-15 |
Family
ID=81627972
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210432686.6A Active CN114529227B (en) | 2022-04-24 | 2022-04-24 | Rural joyful comprehensive service platform based on big data and deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114529227B (en) |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11080616B2 (en) * | 2016-09-27 | 2021-08-03 | Clarifai, Inc. | Artificial intelligence model and data collection/development platform |
| CN111723256A (en) * | 2020-06-03 | 2020-09-29 | 开普云信息科技股份有限公司 | Government affair user portrait construction method and system based on information resource library |
| CN112699297A (en) * | 2020-12-23 | 2021-04-23 | 平安银行股份有限公司 | Service recommendation method, device and equipment based on user portrait and storage medium |
| CN112766878A (en) * | 2020-12-30 | 2021-05-07 | 贵州人和致远数据服务有限责任公司 | Digital country population information management system based on big data technology processing mode |
| CN113516500A (en) * | 2021-04-23 | 2021-10-19 | 深圳市威斯登信息科技有限公司 | Implementation method and system based on big data business and travel operation platform |
| CN113554538A (en) * | 2021-05-28 | 2021-10-26 | 四川社智雲科技有限公司 | Digital information integrated system for urban and rural community management |
-
2022
- 2022-04-24 CN CN202210432686.6A patent/CN114529227B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN114529227A (en) | 2022-05-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Safford et al. | Linking knowledge to action: the role of boundary spanners in translating ecology | |
| Howarth et al. | Understanding barriers to decision making in the UK energy-food-water nexus: The added value of interdisciplinary approaches | |
| Cooper | Location-allocation problems | |
| Chandy et al. | Big data for good: Insights from emerging markets | |
| Valencia-Sandoval et al. | Participatory landscape planning and sustainable community development: Methodological observations from a case study in rural Mexico | |
| Sahani | Application of hybrid SWOT-AHP-FuzzyAHP model for formulation and prioritization of ecotourism strategies in Western Himalaya, India | |
| Williams et al. | Planning for climate change: identifying minimum‐dispersal corridors for the Cape Proteaceae | |
| Brugnach et al. | Toward a relational concept of uncertainty: about knowing too little, knowing too differently, and accepting not to know | |
| Johnson | Rarity in the tropics: latitudinal gradients in distribution and abundance in Australian mammals | |
| Li et al. | Intelligent GIS for solving high‐dimensional site selection problems using ant colony optimization techniques | |
| Härtel et al. | Understanding and responding to the climate change issue: Towards a whole-of-science research agenda | |
| Eden | Lessons on the generation of usable science from an assessment of decision support practices | |
| Verhagen et al. | Predictive spatial modelling | |
| Absalon et al. | Impact assessment of farm tourism sites using a hybrid MADM-based composite sustainability index | |
| Shahzad et al. | Role of organizational vision and adaptability in knowledge management | |
| Cuenca-Garcia et al. | The soil science & archaeo-geophysics alliance (SAGA): Going beyond prospection | |
| CN114529227B (en) | Rural joyful comprehensive service platform based on big data and deep learning | |
| Baulenas et al. | User selection and engagement for climate services coproduction | |
| Barra et al. | A model controversy: Using environmental competency groups to inform coastal restoration planning in Louisiana | |
| Kraus-Polk et al. | Affective ecologies, adaptive management and restoration efforts in the Sacramento-San Joaquin Delta | |
| Martin et al. | Environmental information for total catchment management: incorporating local knowledge | |
| Kabakuş et al. | Comparative analysis of trip generation models according to household characteristics for developed, developing and non-developed provinces in Turkey | |
| Swihart et al. | MODELING INTERACTIONS OF PRIVATE OWNERSHIP AND BIOLOGICAL DIVERSITY | |
| CN109145069A (en) | Network icon basic theory and method | |
| Machmury | Management of Popo village development strategy as a marine tourism village in Takalar Regency |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20240115 Address after: No. 23, 16th Floor, Building 12, Phase I Business District, Guiyang International Financial Center, northwest corner of Lincheng Road and Changling Road, Changling Street Office, Guanshanhu District, Guiyang City, Guizhou Province, 550081 Patentee after: Guizhou Rongfeng Information Technology Co.,Ltd. Address before: Room 150, floor 1, building 14, No. 29 and 31, Baiziwan Road, Chaoyang District, Beijing 100124 Patentee before: Kyushu Yinfeng (Beijing) Technology Co.,Ltd. |