CN110851705A - Project-based collaborative storage recommendation method and recommendation device thereof - Google Patents
Project-based collaborative storage recommendation method and recommendation device thereof Download PDFInfo
- Publication number
- CN110851705A CN110851705A CN201910953264.1A CN201910953264A CN110851705A CN 110851705 A CN110851705 A CN 110851705A CN 201910953264 A CN201910953264 A CN 201910953264A CN 110851705 A CN110851705 A CN 110851705A
- Authority
- CN
- China
- Prior art keywords
- neighborhood
- project
- representation
- similarity
- information
- 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 52
- 230000007246 mechanism Effects 0.000 claims abstract description 27
- 238000005457 optimization Methods 0.000 claims abstract description 8
- 238000004364 calculation method Methods 0.000 claims abstract description 5
- 230000006870 function Effects 0.000 claims description 50
- 230000015654 memory Effects 0.000 claims description 20
- 239000011159 matrix material Substances 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 8
- 230000003993 interaction Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 13
- 238000002474 experimental method Methods 0.000 description 10
- 230000001186 cumulative effect Effects 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000000354 decomposition reaction Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- 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
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a collaborative storage recommendation method and device based on projects, wherein the method comprises the following steps: embedding the user information and the project information respectively to obtain similarity; adding the similarity into an attention mechanism to set a weighting function, and operating the formed weight and the project neighborhood to form neighborhood representation; setting a score function through project information embedding and neighborhood representation; setting a loss function using bayesian personalization-ranking optimization; and realizing the collaborative storage recommendation according to the score function and the loss function. The device comprises: the device comprises a similarity calculation module, a neighborhood representation module, a setting module and a recommendation module. The method effectively overcomes the limitation that only the connection between users is considered, can fully consider the distribution of potential influences between the items and the users, strengthens the relation between a specific item and the neighborhood, considers the potential relation between the items and the users, adopts a higher-order method, enhances the expandability of recommendation, and improves the accuracy and precision of information retrieval and results.
Description
Technical Field
The invention relates to the field of natural language processing and information retrieval, in particular to a collaborative storage recommendation method and a recommendation device based on projects.
Background
The recommendation system recommends information and commodities which are interesting for a user according to the interest characteristics and historical behaviors of the user, and the user is submerged in the problem of information overload in the process of browsing a large amount of irrelevant information and products. The recommendation method based on the content is the continuation and development of the information filtering technology, which is established to make recommendations on the content information of the item, and does not need to obtain the user interest information from the case of the feature description about the content by a machine learning method according to the evaluation opinions of the user on the item, but the feature content is required to have good structure. The recommendation technology based on collaborative filtering is one of the earliest and the most successful technologies applied in a recommendation system, generally adopts the nearest neighbor technology, calculates the distance between users by using the historical preference information of the users, then predicts the preference degree of a target user for a specific commodity by using the weighted evaluation value of the nearest neighbor user of the target user for commodity evaluation, and the system recommends the target user according to the preference degree.
In the recommendation, an attention mechanism can be used for bringing a huge promotion effect on a sequence learning task, an attention model can perform data weighted transformation on a data sequence, hard attention only pays attention to a certain position of the model sequence at each moment, soft attention pays attention to all positions at each time, and the weight of each position is different. Local attention can be seen as a mixture of hard and soft attention in advantage, unlike hard attention, which is nearly everywhere differentiable, easy to train, focusing only a fraction of the source point locations at a time, while global attention needs to scan the full source hidden state at a time.
However, in the prior art, such as a content-based recommendation method, content is required to be easily extracted into meaningful features, the feature content is required to have good structure, and the preference of a user must be expressed in a content feature form, so that the situation of other users cannot be obtained in a display manner, while the collaborative filtering technology has a sparse problem and an extensible problem in the recommendation process, the recommendation method has poor extensibility, and various problems such as cold start are encountered.
Disclosure of Invention
The invention provides a project-based collaborative storage recommendation method and a recommendation device thereof, which are used for ranking and recommending network retrieval information, can effectively overcome the limitation of only considering the connection between users, can fully consider the distribution of potential influences between projects and users, strengthen the relation between a specific project and a neighborhood, consider the potential relation between the projects and the users, adopt a higher-order method to enhance the expandability of recommendation, and improve the accuracy and precision of information retrieval and results in the recommendation process, and are described in detail as follows:
a method of project-based collaborative storage recommendation, the method comprising:
embedding the user information and the project information respectively to obtain similarity;
adding the similarity into an attention mechanism to set a weighting function, and operating the formed weight and the project neighborhood to form neighborhood representation;
setting a score function through project information embedding and neighborhood representation; setting a loss function using bayesian personalization-ranking optimization;
and realizing the collaborative storage recommendation according to the score function and the loss function.
Further, the embedding processing is respectively performed on the user information and the project information, and the obtaining of the similarity specifically includes:
where t is a specific item in the neighborhood of items with implicit feedback, N (u) is a set of all items providing implicit feedback to the user u, muSlicing of user memory components, eiIs a slice of item memory.
Wherein, the adding the similarity into the attention mechanism sets a weighting function, and the formed weight and the project neighborhood are operated to form neighborhood representation specifically as follows:
the learning process of the attention mechanism is established by adopting the self-adaptive weighting function, different weights are given to the contents and items with different influences, and the unique contribution x of the learning itemsiut；
Operating the weights with the neighborhood by using the weights formed by the attention mechanism to form a neighborhood representation siuNote that the weights formed by the force mechanism are selectively weighted in different ways into the items and their neighbors to obtain the final weighted neighborhood representation.
Wherein the unique contribution xiutThe method specifically comprises the following steps:
further, the neighborhood representation siuThe method specifically comprises the following steps:
siu＝∑t∈N(u)xiut×ft
wherein f istIs the vector of the embedded matrix of the entry, x is the matrix-level operation.
The setting of the score function through the project information embedding and the neighborhood information specifically includes:
capturing local structures of neighborhoods of the projects and the users and global interaction information between the projects and the users, and establishing a relation between the projects and the users;
the potential relationship between the item and the user is reflected by the non-linearity to produce a ranking score that is focused on the item's influence.
An apparatus for project-based collaborative storage recommendation, the apparatus comprising:
the similarity calculation module is used for respectively embedding the user information and the project information to obtain similarity;
the neighborhood representation module is used for adding the similarity into an attention mechanism to set a weighting function, and operating the formed weight and the project neighborhood to form neighborhood representation;
the setting module is used for setting a score function through project information embedding and neighborhood representation; setting a loss function using bayesian personalization-ranking optimization;
and the recommending module is used for realizing collaborative storage recommendation according to the score function and the loss function.
The technical scheme provided by the invention has the beneficial effects that:
1. according to the method, the potential influence of the project on the user is considered on the project level, so that better and more comprehensive learning can be realized, and the accuracy of information acquisition in the recommendation process can be effectively improved;
2. the method focuses on the information and the characteristics of the items, can capture the attraction of the items to the user more comprehensively and in detail, improves the accuracy of the recommendation result, and improves the expandability of the recommendation, and the final experimental result shows that the method is more accurate than the recommendation algorithm result which only focuses on the relation between the historical behaviors of the user and the user.
Drawings
FIG. 1 is a flow diagram of a method for collaborative storage recommendation based on projects;
FIG. 2 is an overall architecture diagram of a collaborative storage recommendation method based on projects;
FIG. 3 is a schematic structural diagram of a collaborative storage recommendation device based on projects.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention are described in further detail below.
Example 1
In order to achieve the above object, an embodiment of the present invention provides a collaborative storage recommendation method based on projects, and referring to fig. 1 and fig. 2, the method includes the following steps:
101: embedding the user information and the project information respectively to obtain similarity;
102: adding the obtained similarity into an attention mechanism to set a weighting function, and operating the formed weight and the project neighborhood to form neighborhood representation;
103: setting a score function through project information embedding and neighborhood information;
104: setting a loss function using bayesian personalization-ranking optimization;
105: and evaluating and verifying the effect of the project-based collaborative storage recommendation method according to the score function and the loss function experiment.
In one embodiment, step 101 performs embedding processing on user and project information respectively, and includes the following specific steps:
the input user and item information is embedded into the user and item memory components, respectively, the user memory component being M, the item memory component being E, the user's preferences being stored in a user memory component slice MuWherein the features of item i are stored in item memory slice eiThe similarity a in a given neighborhood for a particular item and a particular user is calculatediut。
In an embodiment, step 102 sets a weighting function by using an attention mechanism (well known to those skilled in the art, which is not described herein in this embodiment of the present invention) based on step 101, and performs an operation on a weight and a neighborhood formed by using the attention mechanism to form a neighborhood representation, which includes the following specific steps:
built using adaptive weighting functionsThe learning process of the attention mechanism can endow different weights to the contents and items with different influences, highlight the importance of the contents with high influence and the unique contribution x of the learning itemsiutThe weights are then manipulated with the neighborhood using weights formed using an attention mechanism to form a neighborhood representation siuNote that the weights formed by the force mechanism are selectively weighted in different ways into the items and their neighbors to obtain the final weighted neighborhood representation.
In one embodiment, step 103 sets a score function through item information embedding and neighborhood information on the basis of step 101 and step 102, and includes the following specific steps:
capturing local structures of neighborhoods of the items and the users and global interaction information between the items and the users, establishing deeper and extensive relation between the items and the users, and then reflecting potential relation between the items and the users more comprehensively through a nonlinear method to generate ranking scores concerning the influence of the items
In one embodiment, step 104 sets a loss function and optimizes the model by the following specific steps:
and (4) optimizing a set loss function and optimizing the model by using Bayes personalized ranking.
In one embodiment, step 105 performs an experiment on the collaborative storage recommendation method based on the project, and includes the following specific steps:
and calculating the hit rate and the normalized breaking and loss accumulation gain of the experiment so as to realize the evaluation of the model effect, and evaluating and verifying the effect of the algorithm by adopting eight baseline comparison experiments in order to better balance the two indexes.
Example 2
The scheme of example 1 is further described below with reference to specific calculation formulas and examples, which are described in detail below:
201: embedding the user information and the project information respectively, and embedding the input user information and the input project information into memory components of the user information and the project information respectively;
where the user's memory component is M, the item's memory component is E, and the user's preferences are stored in a user memory component slice MuWherein the features of item i are stored in item memory slice eiTo find the similarity a in a given neighborhood thereof for a particular item and a particular useriutAs shown in equation (1).
Where t in equation (1) is a particular item in the neighborhood of items with implicit feedback, and n (u) is the set of all items that provide implicit feedback to user u.
202: the attention mechanism is adopted to set the weighting function, the embodiment of the invention adopts the self-adaptive weighting function to establish the learning process of the attention mechanism, and the unique contribution x of the learning itemiutAs shown in equation (2).
Wherein, a in the formula (2)iutIs the similarity in a given neighborhood for a particular item and a particular user.
203: the weights are operated on with the neighborhood using weights formed by an attention mechanism to form a neighborhood representation siuAs shown in equation (3).
siu＝∑t∈N(u)xiut×ft(3)
F in formula (3)tFor the features of the item, the weights formed by the attention mechanism are selectively weighted in different ways into the item and its neighborhood to obtain the final weighted neighborhood.
203: the method comprises the following steps of setting a score function through project information embedding and neighborhood information:
capturing local structures of neighborhoods of the items and the users and global interaction information between the items and the users, establishing deeper and extensive relation between the items and the users, then reflecting potential relation between the items and the users more comprehensively by a nonlinear method, and generating ranking scores of attention item influence by linear projection U
As shown in equation (4), equation (5) and equation (6).
Wherein, in the formulas (4), (5) and (6)
Is a memory, and is used as a memory,is the similarity matrix at level y, v, b are the parameters to be learned,
is the neighborhood updated by the project neighborhood,
for non-linear mapping between memories, etIn order to be a unique contribution of the user t,
for the neighborhood matrix of the fourth layer, σ (x) ═ 1/(1+ exp (-x)) is the nonlinear activation sigmoid function, and y is the layer of the memory.
Wherein, y is 4, W is a weight matrix, which maps the characteristic of the item to the potential space and combines with the information of the previous layer, and is the operation of element level, after the element product of the item and the user is obtained, the neighborhood representation and the parameter obtained by learning are operated in the way of linear projection. Using non-linear activation of ReLU functions
204: and (4) optimizing and setting a loss function by adopting Bayes personalized ranking, and optimizing a recommendation effect as shown in a formula (7).
σ (x) ═ 1/(1+ exp (-x)) in formula (7) is a logical sigmoid function.
205: the effect of the method of the invention can be evaluated and verified by calculating the hit rate and the normalized depreciation cumulative gain.
Example 3
An embodiment of the present invention provides a collaborative storage recommendation apparatus based on a project, and referring to fig. 3, the apparatus includes:
the similarity calculation module is used for respectively embedding the user information and the project information to obtain similarity;
that is, the item embedding matrix E, F and the user embedding matrix M are obtained.
The neighborhood representation module is used for adding the similarity into an attention mechanism to set a weighting function, and operating the formed weight and the project neighborhood to form neighborhood representation;
the setting module is used for setting a score function through project information embedding and neighborhood representation; setting a loss function using bayesian personalization-ranking optimization;
and the recommending module is used for realizing collaborative storage recommendation according to the score function and the loss function.
Example 4
In the experiment of the project-based collaborative storage recommendation method and the recommendation device thereof, the final score of each project is calculated through a score function, and ranking and recommendation are performed through the score.
According to experimental effects, after potential influences of projects on users are introduced, the collaborative storage recommendation method and the recommendation device based on the projects have certain linguistic significance and excellent effects through attention mechanism and optimization of loss functions.
In the experiment of the project-based collaborative storage recommendation method and the recommendation device thereof, the comparison experiment is carried out through eight baseline experiments, the number of negative samples in the experiment is set to be 4, the weight attenuation of l2 in pre-training is set to be 0.001, the weight attenuation of l2 in training is set to be 0.1, the embedding size d of the memory is 50, and both the project information embedding matrix and the user information embedding matrix are automatically generated randomly.
The method and the device evaluate the effect of a collaborative storage recommendation method based on projects by using two evaluation indexes of Hit Rate (HR) and normalized breaking cumulative gain (NDCG) values, wherein the hit rate is calculated mainly for evaluating the proportion of the number of the projects in a top-N recommendation list obtained by a model in a test set, and the normalized breaking cumulative gain is calculated mainly for evaluating the comparison between the effect of the recommendation list generated by the model and the effect of the recommendation list generated in an ideal state. The formula for calculating the Hit Rate (HR) is shown in formula (8), and the formula for calculating the normalized loss cumulative gain (NDCG) is shown in formula (9).
The data value | GT | in equation (8) refers to all test sets, and numberfhits @ K is the sum of the number of test sets belonging to the top K recommendation list for each user. DCG in equation (9) is the average discounted cumulative gain, and ideal DCG @ K is the maximum DCG value under ideal conditions. The hit rate and the normalized breaking accumulated gain are both larger, and the effect is better.
Table 1 is an evaluation index table
The experimental results of nine baseline contrast experiments are shown in table 1, where SVD + + is a hybrid model that combines neighborhood-based similarity with a latent factor model. The Generalized Matrix Factorization (GMF) model is a potential factor model for non-linear generalization. KNN is used in a neighborhood-based approach to compute cosine item-item similarity. Bayesian Personalized Ranking (BPR) is a matrix decomposition model of implicit feedback. Neural matrix decomposition (NeuMF), a matrix decomposition model for item ordering by a multi-layered perceptron model. A cooperative de-noising auto-encoder (CDAE) is a deep learning model based on terms. A Factor Item Similarity Model (FISM) enables the decomposition of a similarity matrix of item-item pairs, optimizing a neighborhood-based model of the loss function. CMN is a model that fuses memory components with attention mechanisms. According to the experimental result, the hit rate and the normalized depreciation cumulative gain of the experimental effect are highest by using the project-based collaborative storage recommendation method and the recommendation device thereof, and the effectiveness of the model is good.
Those skilled in the art will appreciate that the drawings are only schematic illustrations of preferred embodiments, and the above-described embodiments of the present invention are merely provided for description and do not represent the merits of the embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (7)
1. A collaborative storage recommendation method based on projects, characterized in that the method comprises:
embedding the user information and the project information respectively to obtain similarity;
adding the similarity into an attention mechanism to set a weighting function, and operating the formed weight and the project neighborhood to form neighborhood representation;
setting a score function through project information embedding and neighborhood representation; setting a loss function using bayesian personalization-ranking optimization;
and realizing the collaborative storage recommendation according to the score function and the loss function.
2. The project-based collaborative storage recommendation method according to claim 1, wherein the embedding processing is performed on the user and project information respectively, and the obtaining of the similarity specifically comprises:
where t is a specific item in the neighborhood of items with implicit feedback, N (u) is a set of all items providing implicit feedback to the user u, muSlicing of user memory components, eiIs a slice of item memory.
3. The project-based collaborative storage recommendation method according to claim 2, wherein the adding of the similarity into the attention mechanism sets a weighting function, and the formed weights are operated with project neighborhoods, and the formed neighborhood representation is specifically:
the learning process of the attention mechanism is established by adopting the self-adaptive weighting function, different weights are given to the contents and items with different influences, and the unique contribution x of the learning itemsiut；
Operating the weights with the neighborhood by using the weights formed by the attention mechanism to form a neighborhood representation siuNote that the weights formed by the force mechanism are selectively weighted in different ways into the items and their neighbors to obtain the final weighted neighborhood representation.
4. The method of claim 3, wherein the unique contribution x is a collaborative storage recommendationiutThe method specifically comprises the following steps:
5. the method of claim 3, wherein the neighborhood representation s is a representation of a user's location in the neighborhoodiuThe method specifically comprises the following steps:
siu＝∑t∈N(u)xiut×ft
wherein f istIs the vector of the embedded matrix of the entry, x is the matrix-level operation.
6. The method of claim 1, wherein the setting of the score function by the project information embedding and the neighborhood information is specifically:
capturing local structures of neighborhoods of the projects and the users and global interaction information between the projects and the users, and establishing a relation between the projects and the users;
the potential relationship between the item and the user is reflected by the non-linearity to produce a ranking score that is focused on the item's influence.
7. An apparatus for project-based collaborative storage recommendation, the apparatus comprising:
the similarity calculation module is used for respectively embedding the user information and the project information to obtain similarity;
the neighborhood representation module is used for adding the similarity into an attention mechanism to set a weighting function, and operating the formed weight and the project neighborhood to form neighborhood representation;
the setting module is used for setting a score function through project information embedding and neighborhood representation; setting a loss function using bayesian personalization-ranking optimization;
and the recommending module is used for realizing collaborative storage recommendation according to the score function and the loss function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910953264.1A CN110851705A (en) | 2019-10-09 | 2019-10-09 | Project-based collaborative storage recommendation method and recommendation device thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910953264.1A CN110851705A (en) | 2019-10-09 | 2019-10-09 | Project-based collaborative storage recommendation method and recommendation device thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110851705A true CN110851705A (en) | 2020-02-28 |
Family
ID=69596311
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910953264.1A Pending CN110851705A (en) | 2019-10-09 | 2019-10-09 | Project-based collaborative storage recommendation method and recommendation device thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110851705A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111597432A (en) * | 2020-04-03 | 2020-08-28 | 浙江工业大学 | Hash-based quick candidate generation recommendation method |
| CN115019933A (en) * | 2022-06-16 | 2022-09-06 | 浙江工业大学 | Amblyopia training scheme recommendation method fusing GMF and CDAE |
| US11869015B1 (en) | 2022-12-09 | 2024-01-09 | Northern Trust Corporation | Computing technologies for benchmarking |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740430A (en) * | 2016-01-29 | 2016-07-06 | 大连理工大学 | Personalized recommendation method with socialization information fused |
| CN109241424A (en) * | 2018-08-29 | 2019-01-18 | 陕西师范大学 | A kind of recommended method |
| CN109670121A (en) * | 2018-12-18 | 2019-04-23 | 辽宁工程技术大学 | Project level and feature level depth Collaborative Filtering Recommendation Algorithm based on attention mechanism |
| CN110060097A (en) * | 2019-04-01 | 2019-07-26 | 苏州市职业大学 | User behavior sequence of recommendation method based on attention mechanism and convolutional neural networks |
| US20190251435A1 (en) * | 2018-02-09 | 2019-08-15 | Daniel Shiebler | Matching cross domain user affinity with co-embeddings |
| CN110232480A (en) * | 2019-03-01 | 2019-09-13 | 电子科技大学 | The item recommendation method and model training method realized using the regularization stream of variation |
-
2019
- 2019-10-09 CN CN201910953264.1A patent/CN110851705A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740430A (en) * | 2016-01-29 | 2016-07-06 | 大连理工大学 | Personalized recommendation method with socialization information fused |
| US20190251435A1 (en) * | 2018-02-09 | 2019-08-15 | Daniel Shiebler | Matching cross domain user affinity with co-embeddings |
| CN109241424A (en) * | 2018-08-29 | 2019-01-18 | 陕西师范大学 | A kind of recommended method |
| CN109670121A (en) * | 2018-12-18 | 2019-04-23 | 辽宁工程技术大学 | Project level and feature level depth Collaborative Filtering Recommendation Algorithm based on attention mechanism |
| CN110232480A (en) * | 2019-03-01 | 2019-09-13 | 电子科技大学 | The item recommendation method and model training method realized using the regularization stream of variation |
| CN110060097A (en) * | 2019-04-01 | 2019-07-26 | 苏州市职业大学 | User behavior sequence of recommendation method based on attention mechanism and convolutional neural networks |
Non-Patent Citations (2)
| Title |
|---|
| MEI YU， YUE HU等: "Paper Recommendation with Item-Level Collaborative Memory Network", 《IN PROCEEDINGS OF THE 12TH INTERNATIONAL CONFERENCE ON KNOWLEDGE SCIENCE, ENGINEERING AND MANAGEMENT》 * |
| 刘剑波等: "基于SVD++与行为分析的社交推荐", 《计算机应用》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111597432A (en) * | 2020-04-03 | 2020-08-28 | 浙江工业大学 | Hash-based quick candidate generation recommendation method |
| CN115019933A (en) * | 2022-06-16 | 2022-09-06 | 浙江工业大学 | Amblyopia training scheme recommendation method fusing GMF and CDAE |
| US11869015B1 (en) | 2022-12-09 | 2024-01-09 | Northern Trust Corporation | Computing technologies for benchmarking |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110162703B (en) | Content recommendation method, training device, content recommendation equipment and storage medium | |
| AU2016256753B2 (en) | Image captioning using weak supervision and semantic natural language vector space | |
| US9792534B2 (en) | Semantic natural language vector space | |
| GB2547068B (en) | Semantic natural language vector space | |
| CN112163165B (en) | Information recommendation method, device, equipment and computer readable storage medium | |
| US8965891B1 (en) | Training scoring models optimized for highly-ranked results | |
| US10489688B2 (en) | Personalized digital image aesthetics in a digital medium environment | |
| CN110781409B (en) | Article recommendation method based on collaborative filtering | |
| CN108509573B (en) | Book recommendation method and system based on matrix decomposition collaborative filtering algorithm | |
| CN111931062A (en) | Training method and related device of information recommendation model | |
| CN111460130A (en) | Information recommendation method, device, equipment and readable storage medium | |
| CN110851705A (en) | Project-based collaborative storage recommendation method and recommendation device thereof | |
| CN111737578A (en) | Recommendation method and system | |
| CN110827115B (en) | Attention mechanism-based product recommendation method and system | |
| Hu et al. | Learning peer recommendation using attention-driven CNN with interaction tripartite graph | |
| CN111695024A (en) | Object evaluation value prediction method and system, and recommendation method and system | |
| AlShourbaji et al. | An efficient churn prediction model using gradient boosting machine and metaheuristic optimization | |
| CN116976505A (en) | Click rate prediction method of decoupling attention network based on information sharing | |
| CN105677838A (en) | User profile creating and personalized search ranking method and system based on user requirements | |
| CN110727872A (en) | Method and device for mining ambiguous selection behavior based on implicit feedback | |
| Latha et al. | Product recommendation using enhanced convolutional neural network for e-commerce platform | |
| CN118245745A (en) | Task processing, object recommendation, model training method and computing device | |
| CN112084406B (en) | Short message processing method, device, electronic equipment and storage medium | |
| Kumar et al. | Next-item recommendation within a short session using the combined features of horizontal and vertical convolutional neural network | |
| Salehi et al. | A hybrid recommendation approach based on attributes of products using genetic algorithm and naive Bayes classifier |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200228 |
|
| RJ01 | Rejection of invention patent application after publication |