CN110334221A - A kind of interpretation recommended method in knowledge based map path - Google Patents
A kind of interpretation recommended method in knowledge based map path Download PDFInfo
- Publication number
- CN110334221A CN110334221A CN201910649318.5A CN201910649318A CN110334221A CN 110334221 A CN110334221 A CN 110334221A CN 201910649318 A CN201910649318 A CN 201910649318A CN 110334221 A CN110334221 A CN 110334221A
- Authority
- CN
- China
- Prior art keywords
- path
- entity
- user
- paths
- pond
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—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
-
- 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
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)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of interpretation recommended method in knowledge based map path, the interactive history that this method passes through acquisition user, using interactive history as the subset of knowledge mapping, user-item data set is obtained from subset, under the premise of obtaining subset, the ternary group polling of corresponding knowledge mapping is carried out to subset, and these triples are extracted, the semanteme of composite entity and relationship in triplet information is generated into path representation, is made inferences according to path to infer user preference;After determining a triple path, under the premise of constrained Path length is 4, the routing head entity is inquired to tail entity others path, is indicated with multiple triples;After finding mulitpath, pond operation is carried out to each path to distinguish the contribution that prediction is recommended in different paths;Selection contribution point maximum path is recommended by explaining property of user.It is high that this method recommends precision, and solves the problems, such as the opacity of recommendation.
Description
Technical field
The present invention relates to proposed algorithm technical field, the interpretation recommendation side in specifically a kind of knowledge based map path
Method.
Background technique
Nowadays with the improvement of living standards, people can go more to excavate oneself potential interest, and existing skill
Art is also to aid in people and looks for point of interest,, will be auxiliary in recommended technology by some potential activities or commercial product recommending to user
Supplementary information knowledge mapping introduces recommender system and receives more and more attention, and this mode improves the accuracy recommended, still
Accompanying problem is that: for example, why people will buy the commodity of system recommendation or participate in the activity of system recommendation? this gives system
Recommendation brings opacity problem.The key for solving the problems, such as opacity is that the recommendation is allowed to have interpretation, for example, when one
A recommender system to user's Recommendations or activity when, path and the cause for telling user to recommend the commodity push away user
It recommends system and generates more trusts, to improve the efficiency of recommender system.
In view of the above-mentioned problems, researcher is by exploring in knowledge mapping before this in order to make to recommend to have interpretation
Chain type connection, the connection between user and item can be found to be path, provide supplement abundant to interact for user
Information.This connectivity does not only disclose the semanteme of entity and relationship, and helps to understand the interest of user.However, existing
Work sufficiently studying this connectivity to infer the preference of user, especially in the order dependent to inside track
Property and in terms of whole semanteme modeled.For above-mentioned limitation, the present invention proposes a kind of method, suitable in path by utilizing
Sequence dependence, we allow to carry out effective reasoning to path, to infer user-item interaction basic principle.In addition, setting
Having counted a kind of new weighting pond operation makes our model to distinguish advantage of the different paths in terms of connection user and commodity
With certain interpretation.
Summary of the invention
It is an object of the invention to overcome the deficiencies in the prior art, and provide a kind of knowledge graph based on location-based service field
The recommended method of spectrum, this method recommends precision higher, and solves the problems, such as the opacity of recommendation.
Realizing the technical solution of the object of the invention is:
A kind of interpretation recommended method in knowledge based map path, includes the following steps:
1) the Extraction Projects entity from the interactive history of user record, obtains user-item entity data set, solid data
Collect the subset as knowledge mapping KG;
2) entity data set for obtaining step 1), it is corresponding with knowledge mapping KG entity, corresponding ternary group polling is carried out,
Entity-relation table is obtained, going out to be used together triple with Relation extraction for the entity in table indicates;
3) triple obtained according to step 2) generates road with the semanteme of composite entity and relationship in triplet information
Diameter indicates, and infers user preference according to path, determines a favored pathway;
4) favored pathway obtained according to step 3) inquires the routing head entity under the premise of constrained Path length is 4
Path is reached to tail entity others, obtains mulitpath, each path is indicated with multiple triples;
5) mulitpath obtained according to step 4) to each path distinguish different paths to prediction recommend contribution not
Together, it then carries out pond operation and calculates different path contributions, and the maximum paths of contribution point is selected to carry out interpretable push away
It recommends.
The path is that the triple in knowledge mapping KG clearly illustrates the attribute of a relation of direct or indirect item,
The attribute constitutes given one or more path of user and item between, is expressed as follows:
It is the sequence of entity and relationship, sequence table by the path formal definition of user u to project i in knowledge mapping KG
Up to formula are as follows:Wherein e1=u, eL=i;(el,rl,el+1) it is first of triple in p, and
L indicates the number of triple in path.
The contribution that prediction is recommended in the different paths of differentiation in step 5), specific as follows:
User-item entity is to a series of path for usually having paths to connect them in knowledge mapping, if s={ s1,
s2,…,skBe K paths prediction score, P (u, i)={ p1,p2,…,pkIt is path set of the connection user-item to (u, i)
It closes, wherein each element is the average value for being predicted as all path scores calculated according to formula, last, indicates are as follows:
WhereinIndicate that the prediction average mark in all paths, σ (x) indicate activation primitive, wherein x expression parameter, K are road
Diameter item number.
The prediction score is projected to end-state in the prediction score of output using two full communicating layers,
Calculation is as follows:
Wherein W1And W2The respectively coefficient weights in first layer path and second layer path, ReLU (x) are activation primitive,
Middle x is activation parameter, pkIndicate kth paths.
In step 5), the pond operation calculates different path contributions, specific as follows:
A weighting pond operation is designed to polymerize the score in all paths, pond function is defined as follows:
Wherein log [] is mathematical function, and K is path number, and exp () is exponential function, SkScore, γ are predicted for path
It is hyper parameter to control each weighted index;The importance in path is distinguished, the importance in path is determined by gradient, pressure gradient expression formula
It is as follows:
Wherein exp () is exponential function, ∑kIndicate that summation meets, SkIt is path prediction score, γ is hyper parameter to control
Each weighted index, calculated result is directly proportional to the score of each path in backpropagation step, and pond function assigns final prediction
Greater flexibility, in particular, the pond function in maximum pond may degenerate, however, it can be by setting when setting γ → 0
Setting γ → ∞ reduces average pond.
A kind of interpretation recommended method in knowledge based map path provided by the invention, this method is by knowledge mapping road
Diameter is combined with recommendation, solves the problems, such as the opacity of conventional recommendation, is improved and is recommended precision and make to recommend to have interpretable
Property;And using the auxiliary information knowledge mapping of specific network layered structure, so that recommendation results more have a sense of hierarchy, has and push away
It is good to recommend effect, the characteristics of sustainable utilization.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the interpretation recommended method in knowledge based map of the invention path;
Fig. 2 is case result figure.
Specific embodiment
The content of present invention is further elaborated with reference to the accompanying drawings and examples, but is not limitation of the invention.
Embodiment:
A kind of interpretation recommended method in knowledge based map path, process is as shown in Figure 1, below with reference to cinematic data
For specifically introduced, cinematic data is as shown in table 1 below, in conjunction with introduced for film user u4825 method include it is as follows
Step:
1) it watches from user or with Extraction Projects entity movie name in the interactive history record for clicking film, is used
Family-item entity data set, subset of the entity set as film knowledge mapping KG;
2) entity data set obtained according to step 1), it is corresponding with knowledge mapping KG entity, it carries out corresponding triple and looks into
It askes, obtains entity-relation table, going out to be used together triple with Relation extraction for the entity in table indicates;
3) triple (user, relationship, film) information obtained according to step 2), with the composite entity in triplet information
Path representation is generated with the semanteme of relationship, and infers film user preference according to path, determines a favored pathway;
4) favored pathway obtained according to step 3) inquires the routing head entity under the premise of constrained Path length is 4
Path is reached to tail entity others, obtains mulitpath, each path is indicated with multiple triples;
5) mulitpath obtained according to step 4) to each path distinguish different paths to prediction recommend contribution not
Together, it then carries out pond operation and calculates different path contributions, and the maximum paths of contribution point is selected to carry out interpretable push away
It recommends, result is as shown in Fig. 2, we can choose the maximum paths progress interpretation of a contribution point as can be seen from the results
Recommendation can preferably excavate the interest of user.
6) according to above-mentioned steps, We conducted experimental analyses, for two item number of AUC (clicking rate) and ACC (accuracy rate)
According to being analyzed, the two has respectively reached 0.924 and 0.849, this embodies the feasibility and accuracy of our methods.
The path is that the triple in knowledge mapping KG clearly illustrates direct or indirect (i.e. multi-step) item
Attribute of a relation, the attribute constitute given one or more path of user and item between, are expressed as follows:
It is the sequence of entity and relationship by the path formal definition of user u to project i in knowledge mapping KG:Wherein e1=u, eL=i;(el,rl,el+1) it is first of triple in p, and L indicates road
The number of triple in diameter.
The contribution that prediction is recommended in the different paths of differentiation in step 5), specific as follows:
User-item entity is to a series of path for usually having paths to connect them in knowledge mapping, if s={ s1,
s2,…,skBe K paths prediction score, P (u, i)={ p1,p2,…,pkIt is path set of the connection user-item to (u, i)
It closes, wherein each element is the average value for being predicted as all path scores calculated according to formula, last, indicates are as follows:
WhereinIndicate that the prediction average mark in all paths, σ (x) indicate activation primitive, wherein x expression parameter, K are path
Item number.
The prediction score is projected to end-state in the prediction score of output using two full communicating layers,
Calculation is as follows:
Wherein W1And W2The respectively coefficient weights in first layer path and second layer path, ReLU (x) are activation primitive,
Middle x is activation parameter, pkIndicate kth paths.
In step 5), the pond operation calculates different path contributions, specific as follows:
A weighting pond operation is designed to polymerize the score in all paths, pond function is defined as follows:
Wherein log [] is mathematical function, and K is path number, and exp () is exponential function, SkScore, γ are predicted for path
It is hyper parameter to control each weighted index;The importance in path is distinguished, the importance in path is determined by gradient, pressure gradient expression formula
It is as follows:
Wherein exp () is exponential function, ∑kIndicate that summation meets, SkIt is path prediction score, γ is hyper parameter to control
Each weighted index, calculated result is directly proportional to the score of each path in backpropagation step, and pond function assigns final prediction
Greater flexibility, in particular, the pond function in maximum pond may degenerate, however, it can be by setting when setting γ → 0
Setting γ → ∞ reduces average pond.
1 book data collection of table
Claims (5)
1. a kind of interpretation recommended method in knowledge based map path, which comprises the steps of:
1) the Extraction Projects entity from the interactive history of user record, obtains user-item entity data set, and entity data set is made
For the subset of knowledge mapping KG;
2) entity data set for obtaining step 1), it is corresponding with knowledge mapping KG entity, corresponding ternary group polling is carried out, is obtained
Entity-relation table, going out to be used together triple with Relation extraction for the entity in table indicates;
3) triple obtained according to step 2) generates routing table with the semanteme of composite entity and relationship in triplet information
Show, and user preference is inferred according to path, determines a favored pathway;
4) favored pathway obtained according to step 3) inquires the routing head entity to tail under the premise of constrained Path length is 4
Entity others reach path, obtain mulitpath, each path is indicated with multiple triples;
5) mulitpath obtained according to step 4) distinguishes the difference for the contribution that prediction is recommended in different paths to each path, into
The operation of row pond calculates different path contributions, and the maximum paths of contribution point is selected to carry out interpretable recommendation.
2. a kind of interpretation recommended method in knowledge based map path according to claim 1, which is characterized in that institute
The path stated is that the triple in knowledge mapping KG clearly illustrates the attribute of a relation of direct or indirect item, which is constituted
Given one or more path of user and item between, is expressed as follows:
It is the sequence of entity and relationship, sequence expression formula by the path formal definition of user u to project i in knowledge mapping KG
Are as follows:Wherein e1=u, eL=i;(el,rl,el+1) it is first of triple in p, and L table
Show the number of triple in path.
3. a kind of interpretation recommended method in knowledge based map path according to claim 1, which is characterized in that institute
The contribution that prediction is recommended in the different paths of differentiation in step 5) is stated, specific as follows:
User-item entity is to a series of path for having paths to connect them in knowledge mapping, if s={ s1,s2,…,skIt is K
The prediction score of paths, P (u, i)={ p1,p2,…,pkIt is set of paths of the connection user-item to (u, i), wherein each
Element is the average value for being predicted as all path scores calculated according to formula, last, is indicated are as follows:
WhereinIndicate that the prediction average mark in all paths, σ (x) indicate activation primitive, wherein x expression parameter, K is path item
Number.
4. a kind of interpretation recommended method in knowledge based map path according to claim 3, which is characterized in that institute
The prediction score stated is projected to end-state in the prediction score of output using two full communicating layers, and calculation is such as
Under:
Wherein W1And W2The respectively coefficient weights in first layer path and second layer path, ReLU (x) are activation primitive, and wherein x is
Activation parameter, pkIndicate kth paths.
5. a kind of interpretation recommended method in knowledge based map path according to claim 1, which is characterized in that step
It is rapid 5) in, the described pond operation calculates different path contributions, specific as follows:
A weighting pond operation is designed to polymerize the score in all paths, pond function is defined as follows:
Wherein log [] is mathematical function, and K is path number, and exp () is exponential function, SkScore is predicted for path, and γ is super ginseng
Number is to control each weighted index;The importance in path is distinguished, the importance in path is determined that pressure gradient expression formula is as follows by gradient:
Wherein exp () is exponential function, ∑kIndicate that summation meets, SkIt is path prediction score, γ is that hyper parameter is each to control
Weighted index, calculated result is directly proportional to the score of each path in backpropagation step, and it is bigger that pond function assigns final prediction
Flexibility, when setting γ → 0, the pond function in maximum pond may degenerate, and pass through setting γ → ∞ reduction and be averaged pond.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910649318.5A CN110334221B (en) | 2019-07-18 | 2019-07-18 | Interpretability recommendation method based on knowledge graph path |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910649318.5A CN110334221B (en) | 2019-07-18 | 2019-07-18 | Interpretability recommendation method based on knowledge graph path |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110334221A true CN110334221A (en) | 2019-10-15 |
CN110334221B CN110334221B (en) | 2022-12-09 |
Family
ID=68145892
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910649318.5A Active CN110334221B (en) | 2019-07-18 | 2019-07-18 | Interpretability recommendation method based on knowledge graph path |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110334221B (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111147375A (en) * | 2019-12-31 | 2020-05-12 | 秒针信息技术有限公司 | Network operation method, device, computer equipment and medium |
CN111291265A (en) * | 2020-02-10 | 2020-06-16 | 青岛聚看云科技有限公司 | Recommendation information generation method and device |
CN112085559A (en) * | 2020-08-18 | 2020-12-15 | 山东大学 | Interpretable commodity recommendation method and system based on time-sequence knowledge graph |
CN112115210A (en) * | 2020-09-18 | 2020-12-22 | 北京明略昭辉科技有限公司 | Recommendation method and system based on knowledge graph |
CN112115358A (en) * | 2020-09-14 | 2020-12-22 | 中国船舶重工集团公司第七0九研究所 | Personalized recommendation method using multi-hop path features in knowledge graph |
CN112149004A (en) * | 2020-10-30 | 2020-12-29 | 中国科学院计算技术研究所 | Personalized recommendation method based on collaborative knowledge map |
CN112328802A (en) * | 2020-09-30 | 2021-02-05 | 支付宝(杭州)信息技术有限公司 | Data processing method and device and server |
CN112632296A (en) * | 2020-12-31 | 2021-04-09 | 上海交通大学 | Knowledge graph-based paper recommendation method and system with interpretability and terminal |
CN112687388A (en) * | 2021-01-08 | 2021-04-20 | 中山依数科技有限公司 | Interpretable intelligent medical auxiliary diagnosis system based on text retrieval |
CN112766415A (en) * | 2021-02-09 | 2021-05-07 | 第四范式(北京)技术有限公司 | Method, device and system for explaining artificial intelligence model |
CN113254550A (en) * | 2021-06-29 | 2021-08-13 | 浙江大华技术股份有限公司 | Knowledge graph-based recommendation method, electronic device and computer storage medium |
CN113298426A (en) * | 2021-06-17 | 2021-08-24 | 华能澜沧江水电股份有限公司 | Knowledge graph driven dam safety evaluation weight dynamic drafting method and system |
CN113656709A (en) * | 2021-08-24 | 2021-11-16 | 东北大学 | Interpretable interest point recommendation method fusing knowledge graph and time sequence characteristics |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140032109A1 (en) * | 2012-07-30 | 2014-01-30 | Institute For Information Industry | Route recommendation system and method thereof |
CN106503172A (en) * | 2016-10-25 | 2017-03-15 | 天闻数媒科技(湖南)有限公司 | The method and apparatus that learning path recommended by knowledge based collection of illustrative plates |
CN108303108A (en) * | 2017-12-05 | 2018-07-20 | 华南理工大学 | A kind of personalized route recommendation method based on vehicle historical track |
CN108920527A (en) * | 2018-06-07 | 2018-11-30 | 桂林电子科技大学 | A kind of personalized recommendation method of knowledge based map |
CN109241341A (en) * | 2018-05-09 | 2019-01-18 | 上海大学 | A kind of video display of knowledge based map preferentially recommender system and method |
CN112836120A (en) * | 2021-01-27 | 2021-05-25 | 深圳大学 | Multi-mode knowledge graph-based movie recommendation method, system and terminal |
CN113094587A (en) * | 2021-04-23 | 2021-07-09 | 东南大学 | Implicit recommendation method based on knowledge graph path |
CN113326384A (en) * | 2021-06-22 | 2021-08-31 | 四川大学 | Construction method of interpretable recommendation model based on knowledge graph |
-
2019
- 2019-07-18 CN CN201910649318.5A patent/CN110334221B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140032109A1 (en) * | 2012-07-30 | 2014-01-30 | Institute For Information Industry | Route recommendation system and method thereof |
CN106503172A (en) * | 2016-10-25 | 2017-03-15 | 天闻数媒科技(湖南)有限公司 | The method and apparatus that learning path recommended by knowledge based collection of illustrative plates |
CN108303108A (en) * | 2017-12-05 | 2018-07-20 | 华南理工大学 | A kind of personalized route recommendation method based on vehicle historical track |
CN109241341A (en) * | 2018-05-09 | 2019-01-18 | 上海大学 | A kind of video display of knowledge based map preferentially recommender system and method |
CN108920527A (en) * | 2018-06-07 | 2018-11-30 | 桂林电子科技大学 | A kind of personalized recommendation method of knowledge based map |
CN112836120A (en) * | 2021-01-27 | 2021-05-25 | 深圳大学 | Multi-mode knowledge graph-based movie recommendation method, system and terminal |
CN113094587A (en) * | 2021-04-23 | 2021-07-09 | 东南大学 | Implicit recommendation method based on knowledge graph path |
CN113326384A (en) * | 2021-06-22 | 2021-08-31 | 四川大学 | Construction method of interpretable recommendation model based on knowledge graph |
Non-Patent Citations (5)
Title |
---|
GUOYONG CAI ET.AL: "An Improved Collaborative Method for Recommendation and Rating Prediction", 《2014 IEEE INTERNATIONAL CONFERENCE ON DATA MINING WORKSHOP》 * |
刘勤等: "基于知识图谱用户偏好传播的实体推荐模型", 《计算机应用研究》 * |
宋秀来: "基于知识图谱的个性化推荐方法研究", 《中国优秀硕士学位论文全文数据库(电子期刊)》 * |
李浩等: "融合循环知识图谱和协同过滤电影推荐算法", 《计算机工程与应用》 * |
秦川等: "基于知识图谱的推荐***研究综述", 《中国科学:信息科学》 * |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111147375A (en) * | 2019-12-31 | 2020-05-12 | 秒针信息技术有限公司 | Network operation method, device, computer equipment and medium |
CN111291265A (en) * | 2020-02-10 | 2020-06-16 | 青岛聚看云科技有限公司 | Recommendation information generation method and device |
CN111291265B (en) * | 2020-02-10 | 2023-10-03 | 青岛聚看云科技有限公司 | Recommendation information generation method and device |
CN112085559B (en) * | 2020-08-18 | 2024-07-02 | 山东大学 | Interpretable commodity recommendation method and system based on time sequence knowledge graph |
CN112085559A (en) * | 2020-08-18 | 2020-12-15 | 山东大学 | Interpretable commodity recommendation method and system based on time-sequence knowledge graph |
CN112115358A (en) * | 2020-09-14 | 2020-12-22 | 中国船舶重工集团公司第七0九研究所 | Personalized recommendation method using multi-hop path features in knowledge graph |
CN112115358B (en) * | 2020-09-14 | 2024-04-16 | 中国船舶重工集团公司第七0九研究所 | Personalized recommendation method utilizing multi-hop path characteristics in knowledge graph |
CN112115210A (en) * | 2020-09-18 | 2020-12-22 | 北京明略昭辉科技有限公司 | Recommendation method and system based on knowledge graph |
CN112328802A (en) * | 2020-09-30 | 2021-02-05 | 支付宝(杭州)信息技术有限公司 | Data processing method and device and server |
CN112149004A (en) * | 2020-10-30 | 2020-12-29 | 中国科学院计算技术研究所 | Personalized recommendation method based on collaborative knowledge map |
CN112149004B (en) * | 2020-10-30 | 2022-10-25 | 中国科学院计算技术研究所 | Personalized recommendation method based on collaborative knowledge map |
CN112632296A (en) * | 2020-12-31 | 2021-04-09 | 上海交通大学 | Knowledge graph-based paper recommendation method and system with interpretability and terminal |
CN112687388A (en) * | 2021-01-08 | 2021-04-20 | 中山依数科技有限公司 | Interpretable intelligent medical auxiliary diagnosis system based on text retrieval |
CN112687388B (en) * | 2021-01-08 | 2023-09-01 | 中山依数科技有限公司 | Explanatory intelligent medical auxiliary diagnosis system based on text retrieval |
CN112766415B (en) * | 2021-02-09 | 2023-01-24 | 第四范式(北京)技术有限公司 | Method, device and system for explaining artificial intelligence model |
WO2022171037A1 (en) * | 2021-02-09 | 2022-08-18 | 第四范式(北京)技术有限公司 | Method and apparatus for interpreting artificial intelligence model, and system |
CN112766415A (en) * | 2021-02-09 | 2021-05-07 | 第四范式(北京)技术有限公司 | Method, device and system for explaining artificial intelligence model |
CN113298426A (en) * | 2021-06-17 | 2021-08-24 | 华能澜沧江水电股份有限公司 | Knowledge graph driven dam safety evaluation weight dynamic drafting method and system |
CN113254550A (en) * | 2021-06-29 | 2021-08-13 | 浙江大华技术股份有限公司 | Knowledge graph-based recommendation method, electronic device and computer storage medium |
CN113656709B (en) * | 2021-08-24 | 2023-07-25 | 东北大学 | Interpretable interest point recommendation method integrating knowledge graph and time sequence characteristics |
CN113656709A (en) * | 2021-08-24 | 2021-11-16 | 东北大学 | Interpretable interest point recommendation method fusing knowledge graph and time sequence characteristics |
Also Published As
Publication number | Publication date |
---|---|
CN110334221B (en) | 2022-12-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110334221A (en) | A kind of interpretation recommended method in knowledge based map path | |
CN110516160B (en) | Knowledge graph-based user modeling method and sequence recommendation method | |
CN109241440A (en) | It is a kind of based on deep learning towards implicit feedback recommended method | |
CN109977283A (en) | A kind of the tourism recommended method and system of knowledge based map and user's footprint | |
Pan et al. | BIM log mining: Exploring design productivity characteristics | |
US20100211308A1 (en) | Identifying interesting locations | |
CN108256093A (en) | A kind of Collaborative Filtering Recommendation Algorithm based on the more interest of user and interests change | |
CN109785062A (en) | A kind of hybrid neural networks recommender system based on collaborative filtering model | |
CN106096047B (en) | User partition preference calculation method and system based on Information Entropy | |
CN114510653B (en) | Social group recommendation method, system, device and storage medium | |
CN109921921B (en) | Method and device for detecting aging-stable community in time-varying network | |
CN103034963B (en) | A kind of service selection system and system of selection based on correlation | |
CN114817663A (en) | Service modeling and recommendation method based on class perception graph neural network | |
CN109190030A (en) | Implicit feedback recommendation method fusing node2vec and deep neural network | |
Lee et al. | An approach for discovering firm-specific technology opportunities: Application of link prediction to F-term networks | |
CN108648058A (en) | Model sequencing method and device, electronic equipment, storage medium | |
CN109840702A (en) | A kind of new projects' collaborative recommendation method based on multi-core integration | |
CN109242533A (en) | The online motivational techniques of car networking intelligent perception user based on Game Theory | |
CN103198432B (en) | Detection method and detection system of network groups in online social network | |
KR102518389B1 (en) | Method, Apparatus and System for Item Recommendation Using Consumer Bigdata | |
CN115456707A (en) | Method and device for providing commodity recommendation information and electronic equipment | |
US20240039797A1 (en) | Method for predicting quality of service in service network environment | |
CN112364245B (en) | Top-K movie recommendation method based on heterogeneous information network embedding | |
CN106846127A (en) | A kind of time-based Products Show method and system | |
CN114417166A (en) | Continuous interest point recommendation method based on behavior sequence and dynamic social influence |
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 |