CN112667877A - Scenic spot recommendation method and equipment based on tourist knowledge map - Google Patents
Scenic spot recommendation method and equipment based on tourist knowledge map Download PDFInfo
- Publication number
- CN112667877A CN112667877A CN202011567776.3A CN202011567776A CN112667877A CN 112667877 A CN112667877 A CN 112667877A CN 202011567776 A CN202011567776 A CN 202011567776A CN 112667877 A CN112667877 A CN 112667877A
- Authority
- CN
- China
- Prior art keywords
- similarity
- entities
- vector
- entity
- knowledge
- 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
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A scenic spot recommendation method and device based on a tourist knowledge map are disclosed, wherein the recommendation method comprises the following steps: crawling data information of a tourism website, cleaning and arranging the data information, and constructing a tourism domain knowledge map; embedding the entities and the relations of the tourist field knowledge map into a low-dimensional vector, and calculating the similarity between the scenic spots; constructing a user-scenery spot scoring matrix, and calculating the similarity between scenery spots according to the user-scenery spot scoring matrix; fusing the similarity obtained by learning calculation according to knowledge representation and the similarity obtained by calculation of a user-scenery spot scoring matrix to obtain the final similarity; and obtaining the first K entities with the highest similarity with the target entity according to the final similarity to form a recommendation list. The method effectively solves the problems of cold start of the project and data sparsity in the traditional recommendation system, and meanwhile can more accurately make personalized recommendation for the target task.
Description
Technical Field
The invention belongs to the field of intelligent information processing, and particularly relates to a scenic spot recommendation method and device based on a tourist knowledge map.
Background
With the development of the information-based era and the continuous improvement of the living standard, more and more people choose to travel outside on holidays, but generally, on the premise that the holidays are short and the scenic spot information amount is large, tourists cannot make reasonable choices in a short time period, and therefore, more and more attention is paid to the research on a travel recommendation system. How to design a recommendation system according to the historical information of the user and provide more accurate recommendation for the user becomes a popular problem in the current research. On one hand, the conventional collaborative filtering recommendation algorithm often has the problems of cold start of projects and sparse data, and cannot make accurate recommendation for target tasks. On the other hand, the construction of the knowledge graph for a specific field is still in an immature stage, and a great problem still exists for recommendation research based on the travel knowledge graph. The knowledge graph is a semantic network which essentially links entities and relations in knowledge information, and can represent a large amount of knowledge and store the knowledge in a triple manner of 'entity-relation-entity'.
Disclosure of Invention
The invention aims to provide a scenic spot recommendation method and device based on a tourist knowledge map, aiming at the problem that tourist attractions cannot be personalized to a user for recommendation in the prior art, and helping the user make reasonable travel selection.
In order to achieve the above object, the present invention has the following technical means:
a scenic spot recommendation method based on a tourist knowledge map comprises the following steps:
s1, crawling data information of the tourism website, cleaning and sorting the data information, and constructing a tourism domain knowledge map;
s2, embedding the entities and the relations of the tourist field knowledge map into a low-dimensional vector, and calculating the similarity between the scenic spots;
s3, constructing a user-scenery spot scoring matrix, and calculating the similarity between scenery spots according to the user-scenery spot scoring matrix;
s4, fusing the similarity calculated in the step S2 and the step S3 to obtain the final similarity;
and S5, obtaining the first K entities with the highest similarity with the target entity according to the final similarity, and forming a recommendation list.
Preferably, the step S1 specifically includes the following steps:
crawling data information on a related tourism website by using a web crawler in Python, cleaning and sorting the data to ensure the integrity of the data, importing entities in the tourism data into a neo4j database, setting relationship attributes among the entities to form a tourism knowledge map, and obtaining a triple including structured knowledge such as an entity-attribute value.
Preferably, the step S2 specifically includes the following steps: the method comprises the steps of enabling a three-tuple set in a travel knowledge graph to be composed of a head entity h, a tail entity t and a relation vector r, representing all the three-tuple sets as 'head entity-relation-tail entity', namely (h, r, t), representing a learning algorithm through the knowledge graph, mapping the head entity vector h to the tail entity t through the relation vector r, learning by combining network structure characteristics of the knowledge graph, and representing the entities and the relations in the knowledge graph as dense low-dimensional real value vectors.
Preferably, the calculation of the semantic similarity between the entities by the represented low-dimensional dense vector comprises the following steps:
2.1) triple representation learning;
the triple representation learning is to obtain a triple set of entities and relations by constructing a travel knowledge map, map the entities and relations into a continuous low-dimensional dense vector space by a knowledge map representation algorithm, and represent the entities and relations in a numerical mode; initializing a sharing vector, and storing the learned knowledge representation vector into the sharing vector;
2.2) structural feature learning;
the structural feature learning is to accurately reflect a plurality of relationships between entities according to the characteristics of semantic network structures in the knowledge graph; representing a target entity vector using vectors of multi-order neighbor entities and relationships; inputting the entity vector and the relation vector in the neighbor structure of the target entity into a neural network structure, accumulating and summing the vectors of the entity and the relation in the neighbor structure, calculating the ratio of the sum occupying all the entity vector sums, namely the value of the target entity, and then updating the value into a shared vector;
2.3) calculating the similarity;
and calculating the similarity between the entities by utilizing a cosine similarity formula according to the shared vector to construct an entity similarity matrix.
Preferably, the step S3 of constructing the user-attraction score matrix specifically includes the following steps:
according to the crawled data information of the tourist sites, sorting score data of the tourists for the scenic spots, taking the tourists as rows of a matrix, taking the scenic spots as columns of the matrix, constructing a tourist-scenic spot score matrix, obtaining score vectors of each scenic spot based on the tourists, and then calculating the similarity between the scenic spots.
Preferably, the step S4 of calculating the similarity by fusing the weights specifically includes the following steps:
and adding fusion factors between the similarity calculated in the step S2 and the similarity calculated in the step S3 to obtain the final similarity between the scenic spots.
Preferably, the step S5 specifically includes the following steps:
setting a target entity, calculating the final similarity between the target entity and other entities, sorting according to the size of the similarity, and selecting the top k entities with the highest similarity as a recommendation list of the target entity.
The invention also provides a scenic spot recommendation system based on the tourist knowledge map, which comprises the following steps:
the knowledge map building module is used for building a knowledge map in the travel field by crawling data information of the travel website, cleaning and arranging;
the similarity knowledge representation calculation module is used for embedding the entities and the relations of the tourist field knowledge map into a low-dimensional vector by using a knowledge map representation learning algorithm and calculating the similarity between the scenic spots;
the similarity user-scenery spot scoring matrix calculation module is used for constructing a user-scenery spot scoring matrix according to the crawled tourism website data and calculating the similarity between scenery spots according to the user-scenery spot scoring matrix;
the similarity fusion weight calculation module is used for fusing the similarity calculated by the similarity knowledge representation calculation module and the similarity calculated by the similarity user-scenery spot scoring matrix calculation module to obtain the final similarity between the two scenery spots;
and the recommendation list generation module is used for obtaining the first K entities with the highest similarity with the target entity according to the similarity calculated by the similarity fusion weight calculation module to form a recommendation list.
The invention also provides a terminal, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to realize the steps of the scenic spot recommendation method based on the tourist map.
The invention also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the steps of the scenic spot recommendation method based on the tourist map.
Compared with the prior art, the invention has the following beneficial effects: by crawling data information related to the travel website and sorting the data, the integrity of the data is guaranteed, and entities, attributes and attribute values in the data are set, so that a knowledge graph is constructed. The entities and the relations can be embedded into the low-dimensional dense vector through knowledge graph representation learning, the relations between the entities are measured in a mathematical mode, the problem of the traditional recommendation technology is solved, and the problems of cold start of projects and data sparsity in a traditional recommendation system are effectively solved. The method integrates methods of knowledge graph representation learning, recommendation technology, similarity calculation and the like to complete recommendation of the target entity, and simultaneously considers the characteristics of the entity and the display feedback information of the user, so that the accuracy of the recommendation method is improved. The method and the system fuse the similarity calculated by combining the tourism knowledge map representation learning and the user-scenery spot scoring matrix, and can more accurately make personalized recommendation for the target task by adopting collaborative filtering compared with the traditional recommendation technology.
Drawings
FIG. 1 is a flow chart of the present invention for building a travel knowledge map;
FIG. 2 is a flow chart of a scenic spot recommendation method based on a tourist map of the present invention;
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
Referring to fig. 1, the tourism knowledge map is constructed as the first step of the scenic spot recommendation method based on the tourism knowledge map, the tourism knowledge map is constructed according to the requirements of the scenic spot recommendation method, the semantic relation network in the knowledge map is learned and applied to the recommendation technology research, the purpose of recommending articles which best meet the requirements of different users can be achieved, and the recommendation accuracy is improved.
The specific steps for constructing the tourist map are as follows:
step one, analyzing data requirements. According to the recommendation method based on the knowledge graph, entities, attributes and attribute values required in the recommendation process, such as names, addresses, levels, scores and the like of scenic spots, are analyzed, and then corresponding data information is crawled.
Step two, crawling the tourism data information.
And (4) designing a Python program to crawl the travel data information on the travel website and storing the information into the csv file.
And step three, processing the data information. And cleaning and integrating the information in csv files obtained by crawling from different travel websites, deleting wrong or unmatched information in the crawling process, integrating the information stored in different files according to attributes and attribute values, and supplementing the attributes or attribute values lacking in some entities in the files.
And step four, importing the neo4j database. And importing the sorted data information into a neo4j database according to the entities and the attributes respectively, and then establishing the relationship between the entities according to the actual relation between the entities.
And fifthly, exporting the data and constructing a triple. Exporting data in neo4j, storing the data as a csv file, reading the data in the csv file, deleting redundant information, and constructing a triple set by using spaces for separating entities, relations and entities, wherein the triple storage format is 'entity-relation-entity'.
Referring to fig. 2, the knowledge-graph-based recommendation method is designed according to the constructed knowledge graph. The knowledge graph-based recommendation technology is designed by combining knowledge representation learning with an article-based collaborative filtering algorithm by utilizing an improved TransE algorithm TransGraph. The method comprises the steps of firstly learning through knowledge representation, simultaneously embedding triples into low-dimensional dense vectors in combination with network structure characteristics of a knowledge graph in the learning process, calculating similarity between entities, then calculating the similarity between the entities based on user scores through a collaborative filtering algorithm, finally performing fusion calculation on the similarity calculated by the two methods, performing descending ordering on similarity values calculated by a target entity and other entities, and recommending the top K bits which are most similar to the target entity.
The recommendation technical method based on the knowledge graph comprises the following specific steps:
step one, knowledge representation learning.
And (3) forming a triad set obtained by the constructed knowledge graph by a head entity h, a tail entity t and a relation vector r, wherein all triads are expressed as 'head entity-relation-tail entity', namely (h, r, t). And expressing a learning algorithm through a knowledge graph, mapping a head entity vector h to a tail entity t through a relation vector r, and expressing the entities and the relation in the knowledge graph as dense low-dimensional real value vectors by combining network structure characteristic learning of the knowledge graph. The expression contains original semantic information in the knowledge graph, the characteristics of the entities are fully considered, and the semantic similarity between the entities is calculated through the expressed low-dimensional dense vectors. The method comprises the following steps:
1. constructing a Huffman tree according to the triple set;
and taking the frequency of the entity and the relation appearing in the triple as the weight of the leaf node of the Huffman tree.
2. Initializing a sharing vector;
assuming that there are m entities and n relations in the triple, initializing vectors of the entities and relations to form a shared vector matrix V(m+n)xdInitializing vectors of non-leaf nodes in the Huffman tree and forming an auxiliary vector momentThe matrix theta.
3. The triples represent learning;
the triple representation learning is to acquire a triple set of entities and relations by constructing a travel knowledge map. Training and learning by using a TransE algorithm, defining a triplet (h, r, t) in a knowledge graph as a positive sample, randomly replacing a head entity or a tail entity of the triplet of the positive sample with other entities to obtain a negative sample (h ', r, t'), and training according to a loss function formula (1):
wherein S represents a triplet set, S' represents a negative sample set of the triplet set S, l is a spacing distance parameter, [ x ]]+Represents a positive function of x, and d (h + r, t) is the distance between the vector h + r and the vector t.
4. Learning structural features;
the structural feature learning is to accurately reflect a plurality of relationships between entities according to the characteristics of semantic network structures in the knowledge graph. The target entity vector is represented using vectors of multi-order neighbor entities and relationships. And accumulating and summing the entity vectors and the relationship vectors in the neighbor structure of the target entity to obtain an intermediate vector, inquiring the path from the root node to the target node in the Huffman tree, classifying branches every time, multiplying the probabilities generated by the branches to obtain the probability of the target entity, calculating an updated gradient value, and updating the gradient value into a shared vector. The specific implementation steps are as follows:
1) computing the intermediate vector Xt。
Obtaining a contiguous structure N of an entity tr_n(t) accumulating the entity vectors in the adjacent structure to obtain an intermediate vector Xt；
Wherein N isr_n(t)＝{(h1,r1,r2,…,rn),(h2,r1,r2,…,rn),…,(hi,r1,r2,…,rn)}，hiA contiguous entity of t, riIs the relationship between entity t and the adjacent entities.
2) Calculating the probability J of the current entity tt。
Inquiring a path from a root node to an entity t in the constructed Huffman tree, carrying out secondary classification on each branch on the path, generating a probability, and multiplying the probabilities generated by the branches in sequence to obtain the probability J of the current entity ttWherein:
wherein d ist j1 denotes that the jth branch on the path from the root node to t (entity or relationship) is divided into positive classes, dj t0 means that the branch is classified as negative.
3) The auxiliary vector theta and the gradient are updated. The value of the auxiliary vector is updated according to equation (2), and the value of the gradient is updated according to equation (3).
4) The vectors of entities and relationships are updated. And distributing the u value updated in the previous step to a shared vector corresponding to the entity and the relation in the adjacent structure, namely adding the updated gradient value to the value of each entity and relation of the adjacent structure in the shared vector.
5. Cross training;
and updating the vectors of the entities and the relations obtained by the triple representation learning into a shared vector, training by taking the shared vector as an initial vector during the structural feature learning, updating the result obtained by the training into the shared vector, and continuing the triple representation learning by taking the shared vector as the initial vector, thereby realizing the cross training.
And step two, calculating the similarity of knowledge representation.
The triple representation learning and the structural feature learning can truly reflect a plurality of relationships between entities, and the problem that the similarity of calculation between the entities is wrong due to the fact that only a single relationship between the entities can be learned through traditional knowledge representation learning is avoided. According to the shared vector obtained by knowledge representation learning, the similarity between the entities is calculated by utilizing a cosine similarity formula, and the larger the calculated similarity value is, the more similar the two entities are, so that an entity similarity matrix is constructed. Equation (4) is as follows:
wherein, Ii，IjVector representing two sight entities, where Ii＝(C1i,C2i,…,Cmi)T，CijRepresenting the user i's score for sight j.
And step three, constructing a user-item scoring matrix.
And judging the similarity between the items according to the data fed back by the item history by using an item-based collaborative filtering algorithm. And acquiring scoring data of the tourist for the scenic spots through the crawled data, and constructing a tourist-scenic spot scoring matrix. Considering that different users have different scores for the scenic spots, when calculating the similarity between the scenic spots based on the scores of the tourists, the average score of the scenic spots is calculated first, and the cosine similarity between the scenic spots is calculated by combining the average score. Equation (5) is as follows:
and step four, fusing the weights to calculate the similarity.
Similarity Sim between entities to be calculated by knowledge representation learning1(Ii，Ij) And similarity Sim calculated by collaborative filtering algorithm2(Ii，Ij) And adding the fusion weights x to obtain a final similarity matrix, wherein the selection of the fusion weights is to verify the recall rate and the accuracy rate according to a plurality of experimental results and select the fusion weight with the best performance. The calculation formula (6) is as follows:
Sim＝x*Sim1(Ii,Ij)+(1-x)*Sim2(Ii,Ij) (6)
and step five, calculating a recommendation list. The knowledge graph is introduced into the traditional recommendation technology through knowledge graph representation learning, and the accuracy of the recommendation system is improved. Setting a target entity, calculating the fusion similarity of the target entity and other entities according to a similarity fusion method, sequencing the similarity values of the target entity according to a finally calculated similarity matrix, and selecting the top k entities with the highest similarity values as a recommendation list of the target entity.
The invention also provides a scenic spot recommendation system based on the tourist knowledge map, which comprises the following steps:
the knowledge map building module is used for building a knowledge map in the travel field by crawling data information of the travel website, cleaning and arranging;
the similarity knowledge representation calculation module is used for embedding the entities and the relations of the tourist field knowledge map into a low-dimensional vector by using a knowledge map representation learning algorithm and calculating the similarity between the scenic spots;
the similarity user-scenery spot scoring matrix calculation module is used for constructing a user-scenery spot scoring matrix according to the crawled tourism website data and calculating the similarity between scenery spots according to the user-scenery spot scoring matrix;
the similarity fusion weight calculation module is used for fusing the similarity calculated by the similarity knowledge representation calculation module and the similarity calculated by the similarity user-scenery spot scoring matrix calculation module to obtain the final similarity between the two scenery spots;
and the recommendation list generation module is used for obtaining the first K entities with the highest similarity with the target entity according to the similarity calculated by the similarity fusion weight calculation module to form a recommendation list.
The invention also provides a terminal, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to realize the steps of the scenic spot recommendation method based on the tourist map.
The invention also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the steps of the scenic spot recommendation method based on the tourist map.
The computer program may be divided into one or more modules/units, which are stored in the memory and executed by the processor to perform the method of the invention. The terminal can be a desktop computer, a notebook, a palm computer, a cloud server and other computing equipment, and can also be a processor and a memory.
The processor may be a Central Processing Unit (CPU), other general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The memory may be used to store the computer programs and/or modules, and the processor may implement the various functions of the tourist map-based attraction recommendation system module by running or executing the computer programs and/or modules stored in the memory and invoking the data stored in the memory.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the technical solution of the present invention, and it should be understood by those skilled in the art that the technical solution can be modified and replaced by a plurality of simple modifications and replacements without departing from the spirit and principle of the present invention, and the modifications and replacements also fall within the protection scope covered by the claims.
Claims (10)
1. A scenic spot recommendation method based on a travel knowledge map is characterized by comprising the following steps:
s1, crawling data information of the tourism website, cleaning and sorting the data information, and constructing a tourism domain knowledge map;
s2, embedding the entities and the relations of the tourist field knowledge map into a low-dimensional vector, and calculating the similarity between the scenic spots;
s3, constructing a user-scenery spot scoring matrix, and calculating the similarity between scenery spots according to the user-scenery spot scoring matrix;
s4, fusing the similarity calculated in the step S2 and the step S3 to obtain the final similarity;
and S5, obtaining the first K entities with the highest similarity with the target entity according to the final similarity, and forming a recommendation list.
2. The scenic spot recommendation method based on tourist map as claimed in claim 1, wherein the step S1 specifically comprises the steps of: crawling data information on a related tourism website by using a web crawler in Python, cleaning and sorting the data to ensure the integrity of the data, importing entities in the tourism data into a neo4j database, setting relationship attributes among the entities to form a tourism knowledge map, and obtaining a triple including structured knowledge such as an entity-attribute value.
3. The scenic spot recommendation method based on tourist map as claimed in claim 1, wherein the step S2 specifically comprises the steps of: the method comprises the steps of enabling a three-tuple set in a travel knowledge graph to be composed of a head entity h, a tail entity t and a relation vector r, representing all the three-tuple sets as 'head entity-relation-tail entity', namely (h, r, t), representing a learning algorithm through the knowledge graph, mapping the head entity vector h to the tail entity t through the relation vector r, learning by combining network structure characteristics of the knowledge graph, and representing the entities and the relations in the knowledge graph as dense low-dimensional real value vectors.
4. The method of claim 3, wherein the calculating semantic similarity between entities by the represented low-dimensional dense vector comprises the steps of:
2.1) triple representation learning;
the triple representation learning is to obtain a triple set of entities and relations by constructing a travel knowledge map, map the entities and relations into a continuous low-dimensional dense vector space by a knowledge map representation algorithm, and represent the entities and relations in a numerical mode; initializing a sharing vector, and storing the learned knowledge representation vector into the sharing vector;
2.2) structural feature learning;
the structural feature learning is to accurately reflect a plurality of relationships between entities according to the characteristics of semantic network structures in the knowledge graph; representing a target entity vector using vectors of multi-order neighbor entities and relationships; inputting the entity vector and the relation vector in the neighbor structure of the target entity into a neural network structure, accumulating and summing the vectors of the entity and the relation in the neighbor structure, calculating the ratio of the sum occupying all the entity vector sums, namely the value of the target entity, and then updating the value into a shared vector;
2.3) calculating the similarity;
and calculating the similarity between the entities by utilizing a cosine similarity formula according to the shared vector to construct an entity similarity matrix.
5. The scenic spot recommendation method based on the tourist map as claimed in claim 1, wherein the step S3 of constructing the user-scenic spot scoring matrix specifically comprises the steps of: according to the crawled data information of the tourist sites, sorting score data of the tourists for the scenic spots, taking the tourists as rows of a matrix, taking the scenic spots as columns of the matrix, constructing a tourist-scenic spot score matrix, obtaining score vectors of each scenic spot based on the tourists, and then calculating the similarity between the scenic spots.
6. The scenic spot recommendation method based on the tourist map as claimed in claim 1, wherein the step S4 is performed by fusing weights to calculate similarity, and the method specifically comprises the following steps: and adding fusion factors between the similarity calculated in the step S2 and the similarity calculated in the step S3 to obtain the final similarity between the scenic spots.
7. The scenic spot recommendation method based on tourist map as claimed in claim 1, wherein the step S5 specifically comprises the steps of: setting a target entity, calculating the final similarity between the target entity and other entities, sorting according to the size of the similarity, and selecting the top k entities with the highest similarity as a recommendation list of the target entity.
8. A scenic spot recommendation system based on a travel knowledge map is characterized by comprising:
the knowledge map building module is used for building a knowledge map in the travel field by crawling data information of the travel website, cleaning and arranging;
the similarity knowledge representation calculation module is used for embedding the entities and the relations of the tourist field knowledge map into a low-dimensional vector by using a knowledge map representation learning algorithm and calculating the similarity between the scenic spots;
the similarity user-scenery spot scoring matrix calculation module is used for constructing a user-scenery spot scoring matrix according to the crawled tourism website data and calculating the similarity between scenery spots according to the user-scenery spot scoring matrix;
the similarity fusion weight calculation module is used for fusing the similarity calculated by the similarity knowledge representation calculation module and the similarity calculated by the similarity user-scenery spot scoring matrix calculation module to obtain the final similarity between the two scenery spots;
and the recommendation list generation module is used for obtaining the first K entities with the highest similarity with the target entity according to the similarity calculated by the similarity fusion weight calculation module to form a recommendation list.
9. A terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: the processor when executing the computer program performs the steps of the method for tourist map based attraction recommendation of any of claims 1-7.
10. A computer-readable storage medium storing a computer program, characterized in that: the computer program when being executed by a processor performs the steps of the method for tourist map based attraction recommendation according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011567776.3A CN112667877A (en) | 2020-12-25 | 2020-12-25 | Scenic spot recommendation method and equipment based on tourist knowledge map |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011567776.3A CN112667877A (en) | 2020-12-25 | 2020-12-25 | Scenic spot recommendation method and equipment based on tourist knowledge map |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112667877A true CN112667877A (en) | 2021-04-16 |
Family
ID=75409635
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011567776.3A Pending CN112667877A (en) | 2020-12-25 | 2020-12-25 | Scenic spot recommendation method and equipment based on tourist knowledge map |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112667877A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113220905A (en) * | 2021-05-27 | 2021-08-06 | 哈尔滨理工大学 | Service recommendation method fusing knowledge graph |
| CN113220833A (en) * | 2021-05-07 | 2021-08-06 | 支付宝(杭州)信息技术有限公司 | Entity association degree identification method and device |
| CN113312854A (en) * | 2021-07-19 | 2021-08-27 | 成都数之联科技有限公司 | Type selection recommendation method and device, electronic equipment and readable storage medium |
| CN113407731A (en) * | 2021-06-16 | 2021-09-17 | 浙江工商大学 | API recommendation method based on knowledge graph and collaborative filtering |
| CN114399251A (en) * | 2021-12-30 | 2022-04-26 | 淮阴工学院 | Cold-chain logistics recommendation method and device based on semantic network and cluster preference |
| CN114528482A (en) * | 2022-01-25 | 2022-05-24 | 北京三快在线科技有限公司 | Method and device for determining recommended object, electronic equipment and storage medium |
| CN114996490A (en) * | 2022-08-02 | 2022-09-02 | 江西中业智能科技有限公司 | Movie recommendation method, system, storage medium and device |
| WO2023179689A1 (en) * | 2022-03-24 | 2023-09-28 | 重庆邮电大学工业互联网研究院 | Knowledge graph-based recommendation method for internet of things |
| CN117407606A (en) * | 2023-12-14 | 2024-01-16 | 青岛理工大学 | Tourist route recommendation method based on large language model and knowledge graph |
-
2020
- 2020-12-25 CN CN202011567776.3A patent/CN112667877A/en active Pending
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113220833A (en) * | 2021-05-07 | 2021-08-06 | 支付宝(杭州)信息技术有限公司 | Entity association degree identification method and device |
| CN113220905A (en) * | 2021-05-27 | 2021-08-06 | 哈尔滨理工大学 | Service recommendation method fusing knowledge graph |
| CN113407731A (en) * | 2021-06-16 | 2021-09-17 | 浙江工商大学 | API recommendation method based on knowledge graph and collaborative filtering |
| CN113312854A (en) * | 2021-07-19 | 2021-08-27 | 成都数之联科技有限公司 | Type selection recommendation method and device, electronic equipment and readable storage medium |
| CN114399251A (en) * | 2021-12-30 | 2022-04-26 | 淮阴工学院 | Cold-chain logistics recommendation method and device based on semantic network and cluster preference |
| CN114399251B (en) * | 2021-12-30 | 2023-05-30 | 淮阴工学院 | Cold-chain logistics recommendation method and device based on semantic web and cluster preference |
| CN114528482A (en) * | 2022-01-25 | 2022-05-24 | 北京三快在线科技有限公司 | Method and device for determining recommended object, electronic equipment and storage medium |
| WO2023179689A1 (en) * | 2022-03-24 | 2023-09-28 | 重庆邮电大学工业互联网研究院 | Knowledge graph-based recommendation method for internet of things |
| CN114996490A (en) * | 2022-08-02 | 2022-09-02 | 江西中业智能科技有限公司 | Movie recommendation method, system, storage medium and device |
| CN117407606A (en) * | 2023-12-14 | 2024-01-16 | 青岛理工大学 | Tourist route recommendation method based on large language model and knowledge graph |
| CN117407606B (en) * | 2023-12-14 | 2024-03-05 | 青岛理工大学 | Tourist route recommendation method based on large language model and knowledge graph |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112667877A (en) | Scenic spot recommendation method and equipment based on tourist knowledge map | |
| CN110555112B (en) | Interest point recommendation method based on user positive and negative preference learning | |
| CN111428147A (en) | Social recommendation method of heterogeneous graph volume network combining social and interest information | |
| CN110462612A (en) | The method and apparatus for carrying out machine learning using the network at network node with ageng and ranking then being carried out to network node | |
| US20120215795A1 (en) | System and Method For Intelligent Job Hunt | |
| CN112380453B (en) | Article recommendation method and device, storage medium and equipment | |
| CN112905801A (en) | Event map-based travel prediction method, system, device and storage medium | |
| CN109726747A (en) | Recommend the data fusion sort method of platform based on social networks | |
| CN110851664A (en) | Theme-oriented social network node importance degree evaluation method | |
| CN114265986A (en) | Information pushing method and system fusing knowledge graph structure and path semantics | |
| Ming-Te et al. | Using data mining technique to perform the performance assessment of lean service | |
| CN112330510A (en) | Volunteer recommendation method and device, server and computer-readable storage medium | |
| Antelmi et al. | Analyzing, exploring, and visualizing complex networks via hypergraphs using SimpleHypergraphs. jl | |
| CN112131261A (en) | Community query method and device based on community network and computer equipment | |
| CN110457706B (en) | Point-of-interest name selection model training method, using method, device and storage medium | |
| CN108154380A (en) | The method for carrying out the online real-time recommendation of commodity to user based on extensive score data | |
| CN117312680A (en) | Resource recommendation method based on user-entity sub-graph comparison learning | |
| CN116610874A (en) | Cross-domain recommendation method based on knowledge graph and graph neural network | |
| CN115344794A (en) | Scenic spot recommendation method based on knowledge map semantic embedding | |
| CN115587192A (en) | Relationship information extraction method, device and computer readable storage medium | |
| CN112214683B (en) | Mixed recommendation model processing method, system and medium based on heterogeneous information network | |
| CN111291904B (en) | Preference prediction method and device and computer equipment | |
| CN114429384A (en) | Intelligent product recommendation method and system based on e-commerce platform | |
| TWI419071B (en) | Active knowledge management system, method and computer program product for problem solving | |
| CN113407704A (en) | Text matching method, device and equipment and computer readable storage medium |
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 |