CN116992137A - Interpretable ecological civilization mode recommendation method considering spatial heterogeneity - Google Patents
Interpretable ecological civilization mode recommendation method considering spatial heterogeneity Download PDFInfo
- Publication number
- CN116992137A CN116992137A CN202310954478.7A CN202310954478A CN116992137A CN 116992137 A CN116992137 A CN 116992137A CN 202310954478 A CN202310954478 A CN 202310954478A CN 116992137 A CN116992137 A CN 116992137A
- Authority
- CN
- China
- Prior art keywords
- region
- ecological civilization
- mode
- vector
- intention
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 239000013598 vector Substances 0.000 claims abstract description 141
- 230000002776 aggregation Effects 0.000 claims description 9
- 238000004220 aggregation Methods 0.000 claims description 9
- 230000004931 aggregating effect Effects 0.000 claims description 7
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 230000004927 fusion Effects 0.000 claims description 4
- 238000011161 development Methods 0.000 abstract description 19
- 238000010276 construction Methods 0.000 abstract description 5
- 230000001737 promoting effect Effects 0.000 abstract description 5
- 238000013468 resource allocation Methods 0.000 abstract description 3
- 238000005065 mining Methods 0.000 abstract 1
- 238000012163 sequencing technique Methods 0.000 abstract 1
- 230000007613 environmental effect Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 241000983354 Parsonsia heterophylla Species 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000011081 inoculation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000002352 surface water Substances 0.000 description 1
- 230000031068 symbiosis, encompassing mutualism through parasitism Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Computational Linguistics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an interpretable ecological civilization mode recommendation method considering space heterogeneity, which comprises the following steps: and collecting the ecological civilization mode cases. Constructing a knowledge graph in the form of < region, region feature and region >, and mining advanced semantic information on the knowledge graph to consider the spatial heterogeneity of different regions; an intent network in the form of < region, set of intent vectors, ecological civilization pattern > is constructed, and potential preferences for geographic features are analyzed using the intent vectors. The regional vectors obtained on the knowledge graph and the intention network are fused, the scores of the regions on the ecological civilization modes are obtained through the vector inner products of the regional vectors and the ecological civilization modes, and the ecological civilization modes suitable for development of the regions can be recommended for the regions through sequencing the scores. The invention can reveal the advantages and characteristics of different areas, thereby optimizing the resource allocation and enabling the resources of each area to be more reasonably and efficiently utilized. This helps to improve the overall resource utilization efficiency, promoting the sustainable development of ecological civilization construction.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to an interpretable ecological civilization mode recommendation method considering space heterogeneity.
Background
The ecological civilization mode is a social development mode focusing on ecological protection and environmental sustainable development, and aims to promote green development, protect ecological environment, promote living standard of people, promote social progress and promote realization of sustainable development. The ecological civilization mode has important significance in the context of global sustainable development, is not only a practice and exploration for actively responding to a global sustainable development target, but also provides a borrowable experience and concept. The recommendation of the ecological civilization mode is one of important measures for ecological civilization construction, and the recommendation of the proper ecological civilization mode can promote the coordinated development of economy, society and environment, is beneficial to realizing the sustainable development goal, improves the quality of life of people and improves the national competitiveness and image. The method is a comprehensive, long-term and sustainable development mode, accords with the concept of harmonious symbiosis of people and nature, and has positive significance for promoting sustainable development.
However, recommendations for ecological civilization patterns also face a number of challenges. First, in the field of geography, there is a concept of spatial heterogeneity, i.e., differences and diversity in particular properties or characteristics between different sites or regions in space. The spatial heterogeneity reflects the characteristics of unique natural environment, resource distribution, socioeconomic performance, cultural custom and the like among different areas. Because of the existence of spatial heterogeneity, the recommendation of the ecological civilization mode needs to consider different characteristics of different areas, and the same mode cannot be used simply by cutting at a glance. Therefore, a plurality of indexes and influence factors are required to be considered in constructing the model recommended by the ecological civilization mode, and the complexity is high. Moreover, due to the existence of spatial heterogeneity, a generic model may not be suitable for each county, requiring customization of the model. Secondly, the current recommendation of the ecological civilization mode mainly depends on manual operation, and the time cost is too high due to the fact that the number of the ecological civilization modes and the target areas to be recommended is too large. In addition, ecological civilization mode recommendation relates to a plurality of disciplines such as geography, ecology, environmental science and the like, has very high requirements on talent expertise, and is difficult to match corresponding talents for each region. Finally, although many sophisticated recommendation methods exist in the vertical direction of commodities, music and the like in the computer field at present, the recommendation scenes of the methods in the ecological civilization mode are poor in performance due to the existence of spatial heterogeneity. Meanwhile, because of the complex model and data, the recommended result is difficult to intuitively understand, so that people cannot infer and analyze the inoculation factor of a certain ecological civilization mode.
Disclosure of Invention
Aiming at the defects in the prior art, the interpretable ecological civilization mode recommendation method considering the space heterogeneity solves the problem that the current ecological civilization mode cannot consider the geographical characteristics of dynamic change, so that the recommended scene is poor in performance.
In order to achieve the aim of the invention, the invention adopts the following technical scheme:
an interpretable ecological civilization pattern recommendation method considering spatial heterogeneity is provided, which includes the steps of:
s1, collecting ecological civilization mode cases and regional features of different regions to form a regional set, an ecological civilization mode set and a regional feature set;
s2, generating a triplet of region, relation and region characteristics based on the region set, the ecological civilization mode set and the region characteristic set, and constructing a region-region characteristic knowledge graph;
s3, constructing an ecological civilization mode intention network based on the regional-regional characteristic knowledge graph;
s4, information aggregation is carried out on the regional-regional characteristic knowledge graph, and a first vector representation corresponding to each region is obtained;
s5, information aggregation is carried out on the ecological civilization mode intention network, and a second vector representation corresponding to each region is obtained;
and S6, fusing the first vector representation and the second vector representation corresponding to each region to obtain the ecological civilization mode score of the region, and selecting the ecological civilization mode corresponding to the highest ecological civilization mode score as a recommendation result of the region.
Further, step S1 comprises the sub-steps of:
s1-1, collecting the directory of different areas and the directory of developed ecological civilization modes by taking areas/counties as units to respectively obtain an area set and an ecological civilization mode set;
s1-2, collecting natural environment characteristics, resource distribution characteristics, socioeconomic characteristics and cultural custom characteristics of each county in the directory to obtain a regional characteristic set.
Further, the specific method of step S2 comprises the following sub-steps:
s2-1, modeling a knowledge graph body based on a region set, an ecological civilization mode set and a region feature set to form a relationship set R and an entity set E;
s2-2, combining the region, the relation and the region characteristics according to the constructed relation set to generate a triplet of the region, the relation and the region characteristics;
s2-3, importing the generated triples into a Neo4j graphic database to form a region-region characteristic knowledge graph.
Further, the specific method of step S3 comprises the following sub-steps:
s3-1, setting an intention vector group according to a relation set in the knowledge graph;
s3-2, respectively taking the region and the ecological civilization mode as the input of the neural network embedded layer to obtain the vector representation of the region and the vector representation of the ecological civilization mode;
s3-3, forming the vector representation of the region, the intention vector group and the vector representation of the ecological civilization mode into an intention network with the form of < region, intention vector and ecological civilization mode >.
Further, the specific method of step S4 comprises the following sub-steps:
s4-1, searching K-hop neighbors of the regional-regional characteristic knowledge graph by taking a node where any region is located as a central node; wherein K is a superparameter;
s4-2, according to the formula:
aggregating the node information of the searched neighbors and the path information from the neighbor node to the center node to obtain a first vector representation of the regionWherein (1)>Is a collection of pairs of relational entities (r, v), e, connected to a central node d r Is a vector representation of the relation r, e v Is a vector representation of the entity v, and by which is meant the multiplication of the corresponding elements of the two vectors.
Further, the specific method of step S5 comprises the following sub-steps:
s5-1, combining each pair of intention vectors and ecological civilization modes in the intention network;
s5-2, according to the formula:
calculating the importance of each intention vector in the intention vector group to any region, and performing softmax operation on the importance so that the sum of the importance of all the intention vectors is 1; wherein β (d, i) represents the importance of the intent vector i to region d; i is an intention vector group; e, e d Is a region vector representation obtained by embedding a region number; e, e i Is a vector representation of the intent vector i;
s5-3, according to the formula:
based on the importance, the corresponding<Intention vector, ecological civilization mode>Combining and aggregating to target node to obtain second vector representation of the regionWherein (1)>Is the ecological civilization pattern p and intention connected with the target area dA set of vectors i; e, e p Is a vector representation of the ecological civilization pattern p; e, e i Is a vector representation of the intent vector i; as indicated by the fact that the two vectors correspond to the element multiplication and fusion.
Further, the specific method of step S6 comprises the following sub-steps:
s6-1, for any region, directly adding the corresponding second vector representation and the second vector representation to obtain the final vector representation of the region;
s6-2, carrying out inner product on the final vector representation of the region and vector representations of different ecological civilization modes to obtain scores of different ecological civilization modes of the region;
s6-3, selecting an ecological civilization mode corresponding to the highest ecological civilization mode score as a recommendation result of the region.
The beneficial effects of the invention are as follows:
1. the invention can reveal the advantages and characteristics of different areas by considering the space heterogeneity, thereby optimizing the resource allocation and enabling the resources of each area to be more reasonably and efficiently utilized. This helps to improve the overall resource utilization efficiency, promoting the sustainable development of ecological civilization construction.
2. The method for recommending the interpretable ecological civilization mode taking account of the spatial heterogeneity can take the difference of each aspect among different areas into account, and can more accurately recommend the proper ecological civilization mode for the different areas, so that the recommendation result meets the actual needs more.
3. The ecological civilization mode recommendation according to local conditions can avoid the application of an unsuitable ecological civilization mode to unsuitable areas, thereby reducing the environmental risk and protecting the health of an ecological system.
4. The interpretable recommendation result can help policy makers to better understand the difference of each region, and provide scientific basis and decision support. The method is favorable for decision makers to make more intelligent and effective decisions, improves the scientificity and operability of policies, and customizes the proper ecological civilization modes for different areas, thereby promoting the regional differentiation development and realizing the optimal combination of sustainable development of each area. By explaining the basis of the recommendation result, a decision maker can better understand the trade-off and consideration of various factors by the recommendation model, so that the quality of decision making is improved.
Drawings
FIG. 1 is a schematic overall flow diagram of the present method;
figure 2 is a detailed flow chart of the stages of the method.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.
As shown in fig. 1 and 2, the method for recommending an interpretable ecological civilization pattern in consideration of spatial heterogeneity includes the steps of:
s1, collecting ecological civilization mode cases and regional features of different regions to form a regional set, an ecological civilization mode set and a regional feature set;
s2, generating a triplet of region, relation and region characteristics based on the region set, the ecological civilization mode set and the region characteristic set, and constructing a region-region characteristic knowledge graph;
s3, constructing an ecological civilization mode intention network based on the regional-regional characteristic knowledge graph;
s4, information aggregation is carried out on the regional-regional characteristic knowledge graph, and a first vector representation corresponding to each region is obtained;
s5, information aggregation is carried out on the ecological civilization mode intention network, and a second vector representation corresponding to each region is obtained;
and S6, fusing the first vector representation and the second vector representation corresponding to each region to obtain the ecological civilization mode score of the region, and selecting the ecological civilization mode corresponding to the highest ecological civilization mode score as a recommendation result of the region.
Step S1 comprises the following sub-steps:
s1-1, collecting the directory of different areas and the directory of developed ecological civilization modes by taking areas/counties as units to respectively obtain an area set and an ecological civilization mode set;
s1-2, collecting natural environment characteristics, resource distribution characteristics, socioeconomic characteristics and cultural custom characteristics of each county in the directory to obtain a regional characteristic set.
The specific method of the step S2 comprises the following substeps:
s2-1, modeling a knowledge graph body based on a region set, an ecological civilization mode set and a region feature set to form a relationship set R and an entity set E;
s2-2, combining the region, the relation and the region characteristics according to the constructed relation set to generate a triplet of the region, the relation and the region characteristics;
s2-3, importing the generated triples into a Neo4j graphic database to form a region-region characteristic knowledge graph.
The specific method of the step S3 comprises the following substeps:
s3-1, setting an intention vector group according to a relation set in the knowledge graph;
s3-2, respectively taking the region and the ecological civilization mode as the input of the neural network embedded layer to obtain the vector representation of the region and the vector representation of the ecological civilization mode;
s3-3, forming the vector representation of the region, the intention vector group and the vector representation of the ecological civilization mode into an intention network with the form of < region, intention vector and ecological civilization mode >.
The specific method of step S4 comprises the following sub-steps:
s4-1, searching K-hop neighbors of the regional-regional characteristic knowledge graph by taking a node where any region is located as a central node; wherein K is a superparameter;
s4-2, according to the formula:
aggregating the node information of the searched neighbors and the path information from the neighbor node to the center node to obtain a first vector representation of the regionWherein (1)>Is a collection of pairs of relational entities (r, v), e, connected to a central node d r Is a vector representation of the relation r, e v Is a vector representation of the entity v, and by which is meant the multiplication of the corresponding elements of the two vectors.
The specific method of step S5 comprises the following sub-steps:
s5-1, combining each pair of intention vectors and ecological civilization modes in the intention network;
s5-2, according to the formula:
calculating the importance of each intention vector in the intention vector group to any region, and performing softmax operation on the importance so that the sum of the importance of all the intention vectors is 1; wherein β (d, i) represents the importance of the intent vector i to region d; i is an intention vector group; e, e d Is a region vector representation obtained by embedding a region number; e, e i Is a vector representation of the intent vector i;
s5-3, according to the formula:
based on the importance, the corresponding<Intention vector, ecological civilization mode>Combining and aggregating to target node to obtain second vector representation of the regionWherein (1)>Is a set of ecological civilization pattern p and intent vector i connected to target region d; e, e p Is a vector representation of the ecological civilization pattern p; e, e i Is a vector representation of the intent vector i; as indicated by the fact that the two vectors correspond to the element multiplication and fusion.
The specific method of step S6 comprises the following sub-steps:
s6-1, for any region, directly adding the corresponding second vector representation and the second vector representation to obtain the final vector representation of the region;
s6-2, carrying out inner product on the final vector representation of the region and vector representations of different ecological civilization modes to obtain scores of different ecological civilization modes of the region;
s6-3, selecting an ecological civilization mode corresponding to the highest ecological civilization mode score as a recommendation result of the region.
In the specific implementation process, the classical paradigm (region and development mode) of ecological civilization mode development can be collected through various ways such as network searching, official directory, news report and the like when data are collected, and in order to achieve the purpose of taking account of spatial heterogeneity of different regions, the collected characteristics such as natural environment, resource distribution, social economy, cultural custom and the like of counties need to be collected and supplemented. And finally forming a county set and an ecological civilization mode set respectively.
Each vector in the set of intent vectors represents a potential purpose of the region in the ecological civilization mode recommendation process, specifically, each intent vector represents a weight size of all geographic features, and the intent vectors are independent from each other.
In one embodiment of the present invention, an implementation example of the county of Kaifeng in Zhejiang province, includes the following steps:
s11, collecting classical cases in the whole country of the China, wherein the classical cases comprise two parts of contents: firstly, collecting regional directories by taking a district/county as a unit, and secondly, collecting ecological civilization mode directories developed by the district/county;
s12, according to the regional directory collected in the S11, six indexes of feature design of each county in the directory are collected: background information, resource information, economic information, environmental information, social information, cultural information, respectively.
Taking Kaifeng county as an example, the collected indexes include provinces: zhejiang province, urban area: state city, division: east China, precipitation: 1582.030029mm, climate: subtropical humid areas in mountain areas of south of the Yangtze river, altitude: 64.975m, landform: coverage rate of low altitude alluvial terraces and grasslands: 0.2135%, water resource amount per person: 6937.435089 cubic meters, forest coverage: 55.0815%, GDP total: 1984745 Ten thousand yuan, average of people GDP:4.814864681 ten thousand yuan, urban residents and people are all saved: 3.769751972 kiloyuan, second industry yield specific gravity: 49.573%, third industry yield specific gravity: 52.0529338%, mass of surface water environment: 0mg/L, etc.
S13, finally forming a region set D, D epsilon D; an ecological civilization mode set P, P epsilon P; regional feature set F, F E F;
s2, constructing a regional knowledge graph:
s21, modeling a knowledge graph body to form a relation set R, wherein each relation represents a geographic feature type and an entity set E;
s22, converting the data collected in the S12 into triplet data according to the constructed relation set, and preparing for subsequent construction of a knowledge graph. Taking Kaifeng county as an example, a triplet of < Kaifeng county, a subtropical humid region in mountain regions of Jiangnan province, and shou ning county > may be formed because public security county and shou ning county have the same soil type. The method comprises the steps of carrying out a first treatment on the surface of the
S23, importing the triples in the previous step into a Neo4j graphic database to form a regional-regional characteristic knowledge graph, and realizing query and visualization operation of the knowledge graph.
S3, constructing an intention network:
s31, according to a relation set in the knowledge graph, an intention vector group is set, the number of the intention vectors is artificially set, and the intention vector group I is set and comprises 4 intention vectors I1, I2, I3 and I4;
s32, numbering the regional and ecological civilization mode directory in the S1, and obtaining vector representation of the ecological civilization mode and the regional through a neural network embedding layer;
s33, forming the region, the intention vector group and the ecological civilization mode into an intention network in the form of < region, intention vector and ecological civilization mode >. For example, a part of the county of open degree in the intention network is < county of open degree, I, national park mode >, < county of Kaifeng, I, beautiful village mode >.
S4, knowledge graph information aggregation:
s41, searching 4 jump-collar campaigns on the knowledge graph constructed in the S2 by taking Kaifeng county as a center;
s42, aggregating the collar node of the Kaifeng county, including node information and path information from the collar node to the central node, onto the Kaifeng county, extracting spatial heterogeneity characteristics, and obtaining a richer and more accurate node representation of the Kaifeng county
S43, performing aggregation operation on the knowledge graph to capture the space heterogeneity characteristics by adopting a similar method in each other region, and finally obtaining the representation of each region
S5, intention network information aggregation;
s51, merging each pair of < intention vector and ecological civilization mode > combinations in the intention network by utilizing the constructed intention network;
s52, calculating the importance beta of each intention vector in the intention vector group in the intention network to a target area, and adopting softmax operation to the importance degree so that the sum of the importance degrees of the vector group members is 1; the four intention vectors in Kaifeng county are respectively [0.25], [0.25], [0.25], [0.25], and the four intention vectors are of the same importance to Kaifeng county.
S53, according to the importance degree, the fused information is aggregated to regional nodes to obtain
S54, the same operation is adopted for each region, and finally the representation of each region on the intention network is obtained
S6, information fusion and recommendation are carried out;
region vectors obtained on the knowledge graph in S61 and S4And vector polymerized on the intended network in S5 +.>Belonging to different semantic spaces, directly adopting AddFusion, namely directly adding to fuse two kinds of expression to obtain final vector expression e of Kaifeng county, and adopting the same method to obtain e in other areas d ；
S62, using the regional vector representation e obtained in the previous step to turn on and using the ecological civilization mode vector representation e obtained in the S32 p And (3) carrying out inner product operation, and calculating scores of all ecological civilization modes of Kaifeng county, wherein the higher the scores are, the higher the suitable development degree is. The ecological civilization modes of the first five of the Kaihua county scores can be obtained respectively as follows: natural park mode, natural protection area mode, ecological tourism mode, ecological park complex mode and water and soil loss treatment mode;
and S63, adopting the method in S62 for other areas, and obtaining a scoring Matrix of all areas for all ecological civilization modes.
Each intention vector in the overall analysis intention vector group, i1 pays attention to resource information, i2 pays attention to background information, i3 pays attention to various information basically the same, and i4 pays attention to environment information. Since the importance degree of the 4 intention vectors of the Kaifeng county is consistent, the ecological civilization recommendation of the Kaifeng county is the result under the combined action of the several intention vectors and does not depend on single intention, namely, the dependence on the geographic characteristics of all aspects is basically consistent. In addition, the degree of dependence of a certain intention on a specific geographic feature is equal to the weight of the relationship in the intention vector on the knowledge graph, the weight relationship of each intention vector can be analyzed, a more detailed dependence is obtained, and support is provided for decision and research. For example, i1 places most importance on cultivated land area, i2 places most importance on climate, i3 places most importance on forest land coverage, and i4 places most importance on grassland coverage.
In summary, the invention can reveal the advantages and features of different regions, thereby optimizing the resource allocation and enabling the resources of each region to be more reasonably and efficiently utilized. This helps to improve the overall resource utilization efficiency, promoting the sustainable development of ecological civilization construction.
Claims (7)
1. An interpretable ecological civilization mode recommendation method considering spatial heterogeneity, comprising the steps of:
s1, collecting ecological civilization mode cases and regional features of different regions to form a regional set, an ecological civilization mode set and a regional feature set;
s2, generating a triplet of region, relation and region characteristics based on the region set, the ecological civilization mode set and the region characteristic set, and constructing a region-region characteristic knowledge graph;
s3, constructing an ecological civilization mode intention network based on the regional-regional characteristic knowledge graph;
s4, information aggregation is carried out on the regional-regional characteristic knowledge graph, and a first vector representation corresponding to each region is obtained;
s5, information aggregation is carried out on the ecological civilization mode intention network, and a second vector representation corresponding to each region is obtained;
and S6, fusing the first vector representation and the second vector representation corresponding to each region to obtain the ecological civilization mode score of the region, and selecting the ecological civilization mode corresponding to the highest ecological civilization mode score as a recommendation result of the region.
2. An interpretable ecological civilization pattern recommendation method accounting for spatial heterogeneity according to claim 1, wherein step S1 includes the sub-steps of:
s1-1, collecting the directory of different areas and the directory of developed ecological civilization modes by taking areas/counties as units to respectively obtain an area set and an ecological civilization mode set;
s1-2, collecting natural environment characteristics, resource distribution characteristics, socioeconomic characteristics and cultural custom characteristics of each county in the directory to obtain a regional characteristic set.
3. An interpretable ecological civilization pattern recommendation method accounting for spatial heterogeneity according to claim 1, wherein the specific method of step S2 includes the sub-steps of:
s2-1, modeling a knowledge graph body based on a region set, an ecological civilization mode set and a region feature set to form a relationship set R and an entity set E;
s2-2, combining the region, the relation and the region characteristics according to the constructed relation set to generate a triplet of the region, the relation and the region characteristics;
s2-3, importing the generated triples into a Neo4j graphic database to form a region-region characteristic knowledge graph.
4. An interpretable ecological civilization pattern recommendation method accounting for spatial heterogeneity according to claim 3, wherein the specific method of step S3 includes the sub-steps of:
s3-1, setting an intention vector group according to a relation set in the knowledge graph;
s3-2, respectively taking the region and the ecological civilization mode as the input of the neural network embedded layer to obtain the vector representation of the region and the vector representation of the ecological civilization mode;
s3-3, forming the vector representation of the region, the intention vector group and the vector representation of the ecological civilization mode into an intention network with the form of < region, intention vector and ecological civilization mode >.
5. An interpretable ecological civilization pattern recommendation method accounting for spatial heterogeneity according to claim 1, wherein the specific method of step S4 includes the sub-steps of:
s4-1, searching K-hop neighbors of the regional-regional characteristic knowledge graph by taking a node where any region is located as a central node; wherein K is a superparameter;
s4-2, according to the formula:
aggregating the node information of the searched neighbors and the path information from the neighbor node to the center node to obtain a first vector representation of the regionWherein (1)>Is a collection of pairs of relational entities (r, v), e, connected to a central node d r Is a vector representation of the relation r, e v Is a vector representation of the entity v, and by which is meant the multiplication of the corresponding elements of the two vectors.
6. The method of claim 4, wherein the specific method of step S5 comprises the following sub-steps:
s5-1, combining each pair of intention vectors and ecological civilization modes in the intention network;
s5-2, according to the formula:
calculating the importance of each intention vector in the intention vector group to any region, and performing softmax operation on the importance so that the importance of all the intention vectors is always the sameAnd is 1; wherein β (d, i) represents the importance of the intent vector i to region d; i is an intention vector group; e, e d Is a region vector representation obtained by embedding a region number; e, e i Is a vector representation of the intent vector i;
s5-3, according to the formula:
based on the importance, the corresponding<Intention vector, ecological civilization mode>Combining and aggregating to target node to obtain second vector representation of the regionWherein (1)>Is a set of ecological civilization pattern p and intent vector i connected to target region d; e, e p Is a vector representation of the ecological civilization pattern p; e, e i Is a vector representation of the intent vector i; as indicated by the fact that the two vectors correspond to the element multiplication and fusion.
7. The method of claim 4, wherein the specific method of step S6 comprises the following sub-steps:
s6-1, for any region, directly adding the corresponding second vector representation and the second vector representation to obtain the final vector representation of the region;
s6-2, carrying out inner product on the final vector representation of the region and vector representations of different ecological civilization modes to obtain scores of different ecological civilization modes of the region;
s6-3, selecting an ecological civilization mode corresponding to the highest ecological civilization mode score as a recommendation result of the region.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310954478.7A CN116992137A (en) | 2023-07-31 | 2023-07-31 | Interpretable ecological civilization mode recommendation method considering spatial heterogeneity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310954478.7A CN116992137A (en) | 2023-07-31 | 2023-07-31 | Interpretable ecological civilization mode recommendation method considering spatial heterogeneity |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116992137A true CN116992137A (en) | 2023-11-03 |
Family
ID=88522716
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310954478.7A Pending CN116992137A (en) | 2023-07-31 | 2023-07-31 | Interpretable ecological civilization mode recommendation method considering spatial heterogeneity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116992137A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117540102A (en) * | 2023-12-04 | 2024-02-09 | 中国科学院地理科学与资源研究所 | Ecological civilization mode recommendation method based on knowledge graph and multi-mode data |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113505234A (en) * | 2021-06-07 | 2021-10-15 | 中国科学院地理科学与资源研究所 | Construction method of ecological civilization geographical knowledge map |
| CN115062237A (en) * | 2022-06-16 | 2022-09-16 | 东北大学 | Culture resource recommendation method based on combination of graph neural network and knowledge graph |
| WO2023273182A1 (en) * | 2021-06-29 | 2023-01-05 | 北京邮电大学 | Multi-source knowledge graph fusion-oriented entity alignment method and apparatus, and system |
| CN115757897A (en) * | 2022-06-20 | 2023-03-07 | 东北大学 | Intelligent culture resource recommendation method based on knowledge graph convolution network |
| WO2023071845A1 (en) * | 2021-10-25 | 2023-05-04 | 支付宝(杭州)信息技术有限公司 | Knowledge graph processing |
-
2023
- 2023-07-31 CN CN202310954478.7A patent/CN116992137A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113505234A (en) * | 2021-06-07 | 2021-10-15 | 中国科学院地理科学与资源研究所 | Construction method of ecological civilization geographical knowledge map |
| WO2023273182A1 (en) * | 2021-06-29 | 2023-01-05 | 北京邮电大学 | Multi-source knowledge graph fusion-oriented entity alignment method and apparatus, and system |
| WO2023071845A1 (en) * | 2021-10-25 | 2023-05-04 | 支付宝(杭州)信息技术有限公司 | Knowledge graph processing |
| CN115062237A (en) * | 2022-06-16 | 2022-09-16 | 东北大学 | Culture resource recommendation method based on combination of graph neural network and knowledge graph |
| CN115757897A (en) * | 2022-06-20 | 2023-03-07 | 东北大学 | Intelligent culture resource recommendation method based on knowledge graph convolution network |
Non-Patent Citations (1)
| Title |
|---|
| 孙雨生 等: "国内基于知识图谱的信息推荐研究进展", 情报理论与实践, no. 12, 31 December 2019 (2019-12-31), pages 167 - 173 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117540102A (en) * | 2023-12-04 | 2024-02-09 | 中国科学院地理科学与资源研究所 | Ecological civilization mode recommendation method based on knowledge graph and multi-mode data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ge et al. | Farmland transition and its influences on grain production in China | |
| Yang et al. | Urban-rural spatial transformation process and influences from the perspective of land use: A case study of the Pearl River Delta Region | |
| Guaita-Pradas et al. | Analyzing territory for the sustainable development of solar photovoltaic power using GIS databases | |
| Amjad et al. | Identification and assessment of sites for solar farms development using GIS and density based clustering technique-A case of Pakistan | |
| Ma et al. | Overview and progress of Chinese geographical human settlement research | |
| Zhang et al. | Assessing the role of high-speed rail in shaping the spatial patterns of urban and rural development: A case of the Middle Reaches of the Yangtze River, China | |
| Badi et al. | Optimal site selection for sitting a solar park using a novel GIS-SWA’TEL model: A case study in Libya | |
| CN116992137A (en) | Interpretable ecological civilization mode recommendation method considering spatial heterogeneity | |
| Zhou et al. | Multiscale perspective research on the evolution characteristics of the ecosystem services supply-demand relationship in the chongqing section of the three gorges reservoir area | |
| Li et al. | Urban function orientation based on spatiotemporal differences and driving factors of urban construction land | |
| Yang et al. | Interval-thresholding effect of cooling and recreational services of urban parks in metropolises | |
| CN103593365A (en) | Device for real-time update of patent database on basis of Internet | |
| Peng et al. | Integrated urban land-use zoning and associated spatial development: Case study in Shenzhen, China | |
| Wang et al. | Method multi-criteria decision-making method for site selection analysis and evaluation of urban integrated energy stations based on geographic information system | |
| CN113076336A (en) | GIS macro-micro decision support system for site selection of water plant in remote area | |
| CN106951453A (en) | A kind of geographical entity coding method of quick renewal and data sharing | |
| Malemnganbi et al. | Solar power plant site selection: a systematic literature review on MCDM techniques used | |
| Arca et al. | Geographical information systems-based analysis of site selection for wind power plants in Kozlu District (Zonguldak-NW Turkey) by multi-criteria decision analysis method | |
| Fang | Site selection of ecotourism: a case study of Zhejiang province | |
| He et al. | Spatial-temporal dynamics and scenario simulation of land ecological security: a case study of Deyang, Sichuan Province, China | |
| Chen et al. | Uncommonly known change characteristics of land use pattern in Guangdong Province–Hong Kong–Macao, China: Space time pattern, terrain gradient effects and policy implication | |
| Zhou et al. | A water-land-energy-carbon nexus evaluation of agricultural sustainability under multiple uncertainties: The application of a multi-attribute group decision method determined by an interval-valued intuitionistic fuzzy set | |
| Zhao et al. | A multi-aspect coordination HDRED site selection framework under multi-type heterogeneous environments | |
| Liu et al. | Ecological source identification based on the PSR model framework and structural features: a case study in Tianjin, China | |
| Cao | Development potential evaluation for land resources of forest tourism based on fuzzy AHP method |
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 |