CN112052388A - Method and system for recommending gourmet stores - Google Patents
Method and system for recommending gourmet stores Download PDFInfo
- Publication number
- CN112052388A CN112052388A CN202010841508.XA CN202010841508A CN112052388A CN 112052388 A CN112052388 A CN 112052388A CN 202010841508 A CN202010841508 A CN 202010841508A CN 112052388 A CN112052388 A CN 112052388A
- Authority
- CN
- China
- Prior art keywords
- user
- vector
- representation
- food
- similarity
- 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 30
- 239000013598 vector Substances 0.000 claims abstract description 99
- 238000003062 neural network model Methods 0.000 claims abstract description 42
- 238000004364 calculation method Methods 0.000 claims description 24
- 235000019640 taste Nutrition 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 8
- 239000000796 flavoring agent Substances 0.000 claims description 4
- 235000019634 flavors Nutrition 0.000 claims description 4
- 230000006399 behavior Effects 0.000 description 14
- 239000013604 expression vector Substances 0.000 description 6
- 238000013135 deep learning Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/12—Hotels or restaurants
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Tourism & Hospitality (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Evolutionary Biology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method and a system for recommending food stores, wherein the embodiment of the invention respectively calculates the similarity between a representation vector of a user portrait and a representation vector of each food store based on the acquired data of the user portrait and the acquired data of a plurality of food stores; inputting the calculated similarity and the obtained historical scoring data of the user into a trained self-encoder neural network model for scoring prediction to obtain a score of the gourmet shop; and obtaining the food stores recommended for the user according to the grading sequence of the food stores. According to the embodiment of the invention, because the data of the user image is not only the comment of the user, but also the multi-dimensional user characteristic data, the embodiment of the invention can accurately recommend the food shop based on the multi-dimensional user characteristic data.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a method and a system for recommending a cate shop.
Background
In the big data era, mass data brings richer information to users, and simultaneously brings great difficulty to the acquisition of user information meeting requirements. Therefore, the user information required by recommendation in the personalized recommendation mode appears, and when the user information is recommended in the personalized recommendation mode, the information which the user is interested in can be provided according to the historical behavior of the user, so that the user information acquisition efficiency is greatly improved. At present, when performing personalized recommendation, a collaborative filtering method, a content-based recommendation method, and a hybrid recommendation method may be employed. Of these, the collaborative filtering method is the most classical method. In recent years, deep learning has made a breakthrough in the fields of image processing, natural language understanding, speech recognition and the like, and has become a hot trend of artificial intelligence, thereby bringing new opportunities for the research of personalized recommendation. In the personalized recommendation mode based on deep learning, the most common neural network is a self-encoder neural network, the neural network is mainly applied to learning hidden layer feature representation of users and items in the personalized recommendation mode, the hidden representation of the users is obtained by learning user grading data, and the items are predicted based on the hidden representation of the users.
The personalized recommendation based on deep learning has strong deep semantic representation learning capability and better performance in the recommendation field. However, in the food store recommendation for the user, in addition to the scoring data of the user for the food store, other user characteristic data, such as the taste preference, the cuisine preference, the evaluation text, the featured dish and the like of the user, have a great influence on the food store selection of the user. However, the current personalized recommendation mode based on deep learning only considers the scoring data of the user, and is difficult to satisfy the multi-dimensional user characteristic consideration of the user recommended by the gourmet shop; secondly, the existing personalized recommendation method based on deep learning cannot well solve the cold start problem, namely when user scoring data is almost not available, accurate food shop recommendation cannot be made for the user scoring data.
Disclosure of Invention
In view of this, the embodiment of the invention provides a method for recommending a food store, which can accurately recommend the food store based on multi-dimensional user feature data.
The embodiment of the invention also provides a system for recommending the gourmet stores, which can accurately recommend the gourmet stores based on the multi-dimensional user characteristic data.
The embodiment of the invention provides a method for recommending a cate shop, which comprises the following steps:
respectively calculating the similarity between the representation vector of the user portrait and the representation vector of each food store based on the acquired data of the user portrait and the acquired data of the food stores;
inputting the calculated similarity and the obtained historical scoring data of the user into a trained self-encoder neural network model for scoring prediction to obtain a score of the gourmet shop;
and obtaining the food stores recommended for the user according to the grading sequence of the food stores.
Preferably, the data of the user portrait comprises registration information and behavior information of the user, and the acquisition mode is acquired through a background;
the implementation of the user representation vector comprises:
and carrying out vector representation on the registration information of the user by adopting a one-hot code, carrying out vector representation on the behavior information of the user by adopting bag-of-words, and splicing the registration information of the user and the behavior information of the user which are represented by the vector to obtain a representation vector of the user portrait.
Preferably, the data of the food shop comprises comment texts, cuisine, dishes and tastes of the food shop;
the implementation of the representation vector of the gourmet store comprises:
and carrying out vector representation on the comment text of the food store by using bag-of-words, carrying out vector representation on the cuisine, dishes and taste of the food store by using one-hot, and splicing the two vector representations to obtain the representation vector of the food store.
Preferably, the calculating of the similarity between the expression vector of the user portrait and the expression vector of each gourmet shop is performed by using a cosine similarity calculation formula, where the cosine similarity calculation formula is:
wherein, similarity (user, item) represents a cosine similarity function between the user image user and the food store item, and n represents the dimension of the vector.
Preferably, the training self-encoder neural network model comprises:
weighting and summing the calculated similarity and the obtained user history scoring data to obtain a summation vector which is recorded as sumui,sumuiIs calculated as follows:
sumui=ω1ui+ω2simi
wherein, ω is1Scoring data u for user historyiThe weighted value of (1); omega2For the calculated similarity simiThe weighted value of (1);
will su muiAs an input of the self-encoder neural network model, the layers of the self-encoder neural network model are:
wherein h is1、h2Is a two-layer hidden layer from the encoder,W1、b1、W2、b2are the training parameters involved in the two hidden layers,is the output layer from the encoder and,W3、b3training parameters of an output layer;
the self-encoder neural network model is in a stack structure, is trained in a gradient descent mode, and the trained parameters comprise the weighting factors of the similarity and the parameters of each layer of the self-encoder neural network model.
Preferably, the obtaining of the food stores recommended for the user according to the ranking of the scores of the food stores comprises:
and setting an extraction threshold, and extracting the food stores with the set extraction threshold number from the food stores according to the grading sequence of the food stores as the food stores recommended by the user.
A system for recommending a gourmet store, comprising: an acquisition module of user portrait, an acquisition module of a gourmet shop, a similarity calculation module, a model module and a recommendation module, wherein,
the user portrait acquisition module is used for acquiring data of a user portrait;
the device comprises an acquisition module of the food stores, a storage module and a display module, wherein the acquisition module is used for acquiring data of a plurality of food stores;
the similarity calculation module is used for calculating the similarity between the representation vector of the user portrait and the representation vector of each food store respectively based on the acquired data of the user portrait and the acquired data of the food stores;
the model module is used for inputting the calculated similarity and the obtained historical scoring data of the user into a trained self-encoder neural network model for scoring prediction to obtain the score of the gourmet shop;
and the recommending module is used for obtaining the food stores recommended for the user according to the grading sequence of the food stores.
Preferably, the user portrait acquisition module is further configured to obtain user portrait data including registration information and behavior information of a user, where the user portrait data is directly acquired through a background;
the similarity calculation module is further configured to implement a representation vector for the user representation including:
one-ho vector representation is carried out on the registration information of the user, bag-of-words vector representation is carried out on the behavior information of the user, and the registration information of the user and the behavior information of the user which are represented by the vectors are spliced to obtain a representation vector of the user portrait;
the acquisition module of the food shop is also used for acquiring data of the food shop, wherein the data comprises comment texts, cuisine, dishes and flavors of the food shop;
the similarity calculation module is further used for realizing the representation vector of the gourmet shop and comprises the following steps: and carrying out vector representation on the comment text of the food store by using bag-of-words, carrying out vector representation on the cuisine, dishes and taste of the food store by using one-hot, and splicing the two vector representations to obtain the representation vector of the food store.
Preferably, the similarity calculation module is further configured to calculate the similarity between the representation vector of the user portrait and the representation vector of each gourmet store by using a cosine similarity calculation formula.
Preferably, the model module is further configured to train the self-encoder neural network model:
weighting and summing the calculated similarity and the obtained user history scoring data to obtain a summation vector which is recorded as sumui,sumuiIs calculated as follows:
sumui=ω1ui+ω2simi
wherein, ω is1Scoring data u for user historyiThe weighted value of (1); omega2For the calculated similarity simiThe weighted value of (1);
will su muiAs input to the self-encoder neural network model, self-encodingThe layers of the encoder neural network model are:
the self-encoder neural network model is in a stack structure, is trained in a gradient descent mode, and the trained parameters comprise the weighting factors of the similarity and the parameters of each layer of the self-encoder neural network model.
As can be seen from the above, the embodiment of the present invention respectively calculates the similarity between the representation vector of the user portrait and the representation vector of each gourmet store based on the acquired data of a user portrait and the acquired data of a plurality of gourmet stores; inputting the calculated similarity and the obtained historical scoring data of the user into a trained self-encoder neural network model for scoring prediction to obtain a score of the gourmet shop; and obtaining the food stores recommended for the user according to the grading sequence of the food stores. According to the embodiment of the invention, because the data of the user image is not only the comment of the user, but also the multi-dimensional user characteristic data, the embodiment of the invention can accurately recommend the food shop based on the multi-dimensional user characteristic data.
Drawings
FIG. 1 is a flowchart of a method for recommending a gourmet store according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a neural network model of a self-encoder according to an embodiment of the present invention
Fig. 3 is a schematic structural diagram of a system for recommending a gourmet shop according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and examples.
In order to accurately recommend the food stores for the user and improve the user experience degree, the similarity between the representation vector of the user portrait and the representation vector of each food store is respectively calculated based on the acquired data of the user portrait and the acquired data of the food stores; inputting the calculated similarity and the obtained historical scoring data of the user into a trained self-encoder neural network model for scoring prediction to obtain a score of the gourmet shop; and obtaining the food stores recommended for the user according to the grading sequence of the food stores.
In this way, the embodiment of the invention can accurately recommend the food shop based on the multi-dimensional user characteristic data because the data drawn by the user is not only the comment of the user but also the multi-dimensional user characteristic data.
Specifically, when the embodiment of the invention recommends the food shop for the user, on one hand, multi-dimensional user characteristic data is considered, so that more personalized food shop recommendation can be obtained; on the other hand, a part of data of the user portrait can be given by the user registration without being acquired through comment data of the user, so that the embodiment of the invention can better face the cold start problem.
According to the embodiment of the invention, stronger personalized recommendation can be presented by calculating the similarity between the representation vector of the user portrait and the representation vector of each gourmet shop; according to the embodiment of the invention, the similarity between the representation vector of the user portrait and the representation vector of each gourmet shop is used as the input of the model, so that the problem that in the background technology, only the user scoring data is used as the user characteristic data of the input model is single is solved.
Fig. 1 is a flowchart of a method for recommending a gourmet shop according to an embodiment of the present invention, which includes the following specific steps:
102, respectively calculating the similarity between a representation vector of the user portrait and a representation vector of each food store based on the acquired data of the user portrait and the acquired data of the food stores;
and step 104, obtaining the food stores recommended for the user according to the grading sequence of the food stores.
In the method, the data of the user portrait comprises registration information and behavior information of the user, and the acquisition mode is directly acquired through a background, wherein the registration information of the user comprises favorite cuisine, favorite dishes and preference taste; the behavior information of the user comprises a user comment text, dishes liked by the user, the user shop scores and the like.
In the method, the implementation of the vector of representations of the user representation includes:
and carrying out vector representation on the registration information of the user by adopting one-hot codes (one-hot), carrying out vector representation on the behavior information of the user by adopting bag-of-words (bag-of-words), and splicing the registration information of the user and the behavior information of the user, which are represented by the vector, so as to obtain a representation vector of the user portrait.
Specifically, a favorite dish series, favorite dishes and favorite tastes are vector-represented by one-hot, a user comment text is vector-represented by bag-of-words, and the two vector representations are spliced to form a user image vector representation. Therefore, multi-dimensional user feature data can be mined, and the user portrait can be converted into multi-dimensional feature representation.
In the method, the data of the food shop comprises comment texts, cuisine, dishes and tastes of the food shop.
In the method, the implementation of the representation vector of the gourmet store comprises:
and carrying out vector representation on the comment text of the food store by using bag-of-words, carrying out vector representation on the cuisine, dishes and taste of the food store by using one-hot, and splicing the two vector representations to obtain the representation vector of the food store. Therefore, the food store characteristics of the characteristic dimension corresponding to the data of the user portrait can be mined, and the data of the food store can be converted into multi-dimensional characteristic representation.
In the method, the similarity between the expression vector of the user portrait and the expression vector of each gourmet shop is calculated by adopting a cosine similarity calculation formula, wherein the cosine similarity calculation formula is as follows:
wherein, similarity (user, item) represents a cosine similarity function between the user image (user) and the food store (item). n denotes the dimension of the vector.
In the method, the historical scoring data of the user is acquired through a background and is the historical scoring data of the user on the corresponding gourmet shop.
In the method, the self-encoder neural network model needs to be trained before being used, and the process of training the self-encoder neural network model comprises the following steps:
taking the calculated similarity and the obtained user history scoring data as input, weighting and summing the two to obtain a summation vector which is recorded as sumui,sumuiIs calculated as follows:
sumui=ω1ui+ω2simi
wherein, ω is1Scoring data u for user historyiThe weighted value of (1); omega2For the calculated similarity simiWeighted value of (1), sumuiAs an input of the self-encoder neural network model, the layers of the self-encoder neural network model are:
wherein h is1、h2Is a two-layer hidden layer from the encoder,W1、b1、W2、b2are the training parameters involved in the two hidden layers,is the output layer from the encoder and,W3、b3is the training parameter of the output layer.
The self-encoder neural network model is in a stack structure, is trained in a gradient descent mode, and the trained parameters comprise the weighting factors of the similarity and the parameters of each layer of the self-encoder neural network model.
Fig. 2 is a schematic structural diagram of a self-encoder neural network model according to an embodiment of the present invention.
In the method, the obtaining of the food stores recommended for the user according to the ranking of the scores of the food stores comprises:
and setting an extraction threshold, and extracting the food stores with the set extraction threshold number from the food stores according to the grading sequence of the food stores as the food stores recommended by the user.
Fig. 3 is a schematic structural diagram of a system for recommending a gourmet shop according to an embodiment of the present invention, including: an acquisition module of user portrait, an acquisition module of a gourmet shop, a similarity calculation module, a model module and a recommendation module, wherein,
the user portrait acquisition module is used for acquiring data of a user portrait;
the device comprises an acquisition module of the food stores, a storage module and a display module, wherein the acquisition module is used for acquiring data of a plurality of food stores;
the similarity calculation module is used for calculating the similarity between the representation vector of the user portrait and the representation vector of each food store respectively based on the acquired data of the user portrait and the acquired data of the food stores;
the model module is used for inputting the calculated similarity and the obtained historical scoring data of the user into a trained self-encoder neural network model for scoring prediction to obtain the score of the gourmet shop;
and the recommending module is used for obtaining the food stores recommended for the user according to the grading sequence of the food stores.
In the system, the user portrait acquisition module is also used for acquiring user portrait data including registration information and behavior information of a user, and the acquisition mode is directly acquired through a background;
a similarity calculation module, further for implementation of the representation vector of the user representation, comprising:
and carrying out vector representation on the registration information of the user by adopting one-ho, carrying out vector representation on the behavior information of the user by adopting bag-of-words, and splicing the registration information of the user and the behavior information of the user which are represented by the vector to obtain a representation vector of the user portrait.
In the system, the acquisition module of the food shop is also used for acquiring the data of the food shop, wherein the data comprises comment texts, cuisine, dishes and flavors of the food shop;
the similarity calculation module is also used for realizing the representation vector of the gourmet shop and comprises the following steps: and carrying out vector representation on the comment text of the food store by using bag-of-words, carrying out vector representation on the cuisine, dishes and taste of the food store by using one-hot, and splicing the two vector representations to obtain the representation vector of the food store.
In the system, the similarity calculation module is further configured to calculate the similarity between the expression vector of the user portrait and the expression vector of each gourmet shop by using a cosine similarity calculation formula.
The model module is further configured to train the self-encoder neural network model:
taking the calculated similarity and the obtained user history scoring data as input, and weighting the calculated similarity and the obtained user history scoring data to obtain the similarityAnd, obtaining a sum vector, denoted sumui,sumuiIs calculated as follows:
sumui=ω1ui+ω2simi
wherein, ω is1Scoring data u for user historyiThe weighted value of (1); omega2For the calculated similarity simiWeighted value of (1), sumuiAs an input of the self-encoder neural network model, the layers of the self-encoder neural network model are:
the self-encoder neural network model is in a stack structure, is trained in a gradient descent mode, and the trained parameters comprise the weighting factors of the similarity and the parameters of each layer of the self-encoder neural network model.
In the system, the model module is further used for setting an extraction threshold value, and extracting the food stores with the set extraction threshold value number from the food stores according to the grading ranking of the food stores as the food stores recommended by the user.
The embodiment of the invention realizes the recommendation of the gourmet shop, better solves the cold start problem, considers the characteristics of the favorite taste, the cuisine, the comments and the like of the user besides the historical scoring data of the user, and realizes the recommendation of the gourmet shop with stronger individuation.
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 made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A method of recommending a gourmet store, the method comprising:
respectively calculating the similarity between the representation vector of the user portrait and the representation vector of each food store based on the acquired data of the user portrait and the acquired data of the food stores;
inputting the calculated similarity and the obtained historical scoring data of the user into a trained self-encoder neural network model for scoring prediction to obtain a score of the gourmet shop;
and obtaining the food stores recommended for the user according to the grading sequence of the food stores.
2. The method of claim 1, wherein the user representation data includes registration information and behavior information of the user, the acquisition being by the background;
the implementation of the user representation vector comprises:
and carrying out vector representation on the registration information of the user by adopting a one-hot code, carrying out vector representation on the behavior information of the user by adopting bag-of-words, and splicing the registration information of the user and the behavior information of the user which are represented by the vector to obtain a representation vector of the user portrait.
3. The method of claim 1, wherein the food store data includes food store review text, cuisine, dishes, and flavors;
the implementation of the representation vector of the gourmet store comprises:
and carrying out vector representation on the comment text of the food store by using bag-of-words, carrying out vector representation on the cuisine, dishes and taste of the food store by using one-hot, and splicing the two vector representations to obtain the representation vector of the food store.
4. The method of claim 1, wherein the calculating the similarity between the vector representation of the user representation and the vector representation of each gourmet store is performed using a cosine similarity calculation formula, wherein the cosine similarity calculation formula is:
wherein, similarity (user, item) represents a cosine similarity function between the user image user and the food store item, and n represents the dimension of the vector.
5. The method of claim 1, wherein the training the self-encoder neural network model comprises:
weighting and summing the calculated similarity and the obtained user history scoring data to obtain a summation vector which is recorded as sumui,sumuiIs calculated as follows:
sumui=ω1ui+ω2simi
wherein, ω is1Scoring data u for user historyiThe weighted value of (1); omega2For the calculated similarity simiThe weighted value of (1);
will su muiAs an input of the self-encoder neural network model, the layers of the self-encoder neural network model are:
wherein h is1、h2Is a two-layer hidden layer from the encoder,W1、b1、W2、b2are the training parameters involved in the two hidden layers,is the output layer from the encoder and,W3、b3training parameters of an output layer;
the self-encoder neural network model is in a stack structure, is trained in a gradient descent mode, and the trained parameters comprise the weighting factors of the similarity and the parameters of each layer of the self-encoder neural network model.
6. The method of claim 1, wherein the deriving the food stores recommended for the user according to the ranking of the food stores' scores comprises:
and setting an extraction threshold, and extracting the food stores with the set extraction threshold number from the food stores according to the grading sequence of the food stores as the food stores recommended by the user.
7. A system for recommending a gourmet store, comprising: an acquisition module of user portrait, an acquisition module of a gourmet shop, a similarity calculation module, a model module and a recommendation module, wherein,
the user portrait acquisition module is used for acquiring data of a user portrait;
the device comprises an acquisition module of the food stores, a storage module and a display module, wherein the acquisition module is used for acquiring data of a plurality of food stores;
the similarity calculation module is used for calculating the similarity between the representation vector of the user portrait and the representation vector of each food store respectively based on the acquired data of the user portrait and the acquired data of the food stores;
the model module is used for inputting the calculated similarity and the obtained historical scoring data of the user into a trained self-encoder neural network model for scoring prediction to obtain the score of the gourmet shop;
and the recommending module is used for obtaining the food stores recommended for the user according to the grading sequence of the food stores.
8. The system of claim 7, wherein the user representation acquisition module is further configured to acquire user representation data including registration information and behavior information of the user directly from the background;
the similarity calculation module is further configured to implement a representation vector for the user representation including:
one-ho vector representation is carried out on the registration information of the user, bag-of-words vector representation is carried out on the behavior information of the user, and the registration information of the user and the behavior information of the user which are represented by the vectors are spliced to obtain a representation vector of the user portrait;
the acquisition module of the food shop is also used for acquiring data of the food shop, wherein the data comprises comment texts, cuisine, dishes and flavors of the food shop;
the similarity calculation module is further used for realizing the representation vector of the gourmet shop and comprises the following steps: and carrying out vector representation on the comment text of the food store by using bag-of-words, carrying out vector representation on the cuisine, dishes and taste of the food store by using one-hot, and splicing the two vector representations to obtain the representation vector of the food store.
9. The system of claim 7, wherein the similarity calculation module is further configured to calculate the similarity between the vector representation of the user representation and the vector representation of each gourmet store using a cosine similarity calculation formula.
10. The system of claim 7, wherein the model module is further to train a self-coder neural network model to:
weighting and summing the calculated similarity and the obtained user history scoring data to obtain a summation vector which is recorded as sumui,sumuiIs calculated as follows:
sumui=ω1ui+ω2simi
wherein, ω is1Scoring data u for user historyiThe weighted value of (1); omega2For the calculated similarity simiThe weighted value of (1);
will su muiAs an input of the self-encoder neural network model, the layers of the self-encoder neural network model are:
the self-encoder neural network model is in a stack structure, is trained in a gradient descent mode, and the trained parameters comprise the weighting factors of the similarity and the parameters of each layer of the self-encoder neural network model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010841508.XA CN112052388A (en) | 2020-08-20 | 2020-08-20 | Method and system for recommending gourmet stores |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010841508.XA CN112052388A (en) | 2020-08-20 | 2020-08-20 | Method and system for recommending gourmet stores |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112052388A true CN112052388A (en) | 2020-12-08 |
Family
ID=73599738
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010841508.XA Pending CN112052388A (en) | 2020-08-20 | 2020-08-20 | Method and system for recommending gourmet stores |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112052388A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112667714A (en) * | 2021-03-17 | 2021-04-16 | 腾讯科技(深圳)有限公司 | User portrait optimization method and device based on deep learning and storage medium |
CN113570432A (en) * | 2021-07-28 | 2021-10-29 | 北京达佳互联信息技术有限公司 | Resource recommendation determining method, device, equipment and storage medium |
Citations (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012073966A (en) * | 2010-09-29 | 2012-04-12 | Dainippon Printing Co Ltd | Data generation device, data generation method and data generation program |
CN103345699A (en) * | 2013-07-10 | 2013-10-09 | 湖南大学 | Personalized food recommendation method based on commodity forest system |
CN104732422A (en) * | 2015-02-27 | 2015-06-24 | 湖南大学 | Mobile user tag based personalized food recommendation method |
KR101707753B1 (en) * | 2015-12-21 | 2017-02-16 | 비씨카드(주) | Method for personalized recommending affiliated store and affiliated store recommendation server |
CN106934498A (en) * | 2017-03-14 | 2017-07-07 | 携程旅游网络技术(上海)有限公司 | The recommendation method and system of hotel's house type in OTA websites |
CN107220852A (en) * | 2017-05-26 | 2017-09-29 | 北京小度信息科技有限公司 | Method, device and server for determining target recommended user |
CN107256494A (en) * | 2017-05-18 | 2017-10-17 | 深圳大学 | A kind of item recommendation method and device |
WO2018032790A1 (en) * | 2016-08-16 | 2018-02-22 | 武汉斗鱼网络科技有限公司 | Weighted k-nearest-neighbor scoring-based live broadcast room recommendation method and system |
CN109064278A (en) * | 2018-07-26 | 2018-12-21 | 北京三快在线科技有限公司 | Target object recommended method and device, electronic equipment, storage medium |
CN109284413A (en) * | 2018-09-07 | 2019-01-29 | 平安科技(深圳)有限公司 | Method of Commodity Recommendation, device, equipment and storage medium based on recognition of face |
KR20190031001A (en) * | 2017-09-15 | 2019-03-25 | 조선대학교산학협력단 | User preferred restaurant recommendation method and computer program |
CN109766491A (en) * | 2018-12-18 | 2019-05-17 | 深圳壹账通智能科技有限公司 | Product search method, device, computer equipment and storage medium |
CN110020186A (en) * | 2018-05-08 | 2019-07-16 | 美味不用等(上海)信息科技股份有限公司 | A kind of dining room recommended method and system |
WO2019242144A1 (en) * | 2018-06-19 | 2019-12-26 | 平安科技(深圳)有限公司 | Electronic device, preference tendency prediction method and computer readable storage medium |
CN110619076A (en) * | 2018-12-25 | 2019-12-27 | 北京时光荏苒科技有限公司 | Search term recommendation method and device, computer and storage medium |
CN111008336A (en) * | 2019-12-23 | 2020-04-14 | 腾讯科技(深圳)有限公司 | Content recommendation method, device and equipment and readable storage medium |
CN111429204A (en) * | 2020-03-10 | 2020-07-17 | 携程计算机技术(上海)有限公司 | Hotel recommendation method, system, electronic equipment and storage medium |
CN111444428A (en) * | 2020-03-27 | 2020-07-24 | 腾讯科技(深圳)有限公司 | Information recommendation method and device based on artificial intelligence, electronic equipment and storage medium |
-
2020
- 2020-08-20 CN CN202010841508.XA patent/CN112052388A/en active Pending
Patent Citations (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012073966A (en) * | 2010-09-29 | 2012-04-12 | Dainippon Printing Co Ltd | Data generation device, data generation method and data generation program |
CN103345699A (en) * | 2013-07-10 | 2013-10-09 | 湖南大学 | Personalized food recommendation method based on commodity forest system |
CN104732422A (en) * | 2015-02-27 | 2015-06-24 | 湖南大学 | Mobile user tag based personalized food recommendation method |
KR101707753B1 (en) * | 2015-12-21 | 2017-02-16 | 비씨카드(주) | Method for personalized recommending affiliated store and affiliated store recommendation server |
WO2018032790A1 (en) * | 2016-08-16 | 2018-02-22 | 武汉斗鱼网络科技有限公司 | Weighted k-nearest-neighbor scoring-based live broadcast room recommendation method and system |
CN106934498A (en) * | 2017-03-14 | 2017-07-07 | 携程旅游网络技术(上海)有限公司 | The recommendation method and system of hotel's house type in OTA websites |
CN107256494A (en) * | 2017-05-18 | 2017-10-17 | 深圳大学 | A kind of item recommendation method and device |
CN107220852A (en) * | 2017-05-26 | 2017-09-29 | 北京小度信息科技有限公司 | Method, device and server for determining target recommended user |
KR20190031001A (en) * | 2017-09-15 | 2019-03-25 | 조선대학교산학협력단 | User preferred restaurant recommendation method and computer program |
CN110020186A (en) * | 2018-05-08 | 2019-07-16 | 美味不用等(上海)信息科技股份有限公司 | A kind of dining room recommended method and system |
WO2019242144A1 (en) * | 2018-06-19 | 2019-12-26 | 平安科技(深圳)有限公司 | Electronic device, preference tendency prediction method and computer readable storage medium |
CN109064278A (en) * | 2018-07-26 | 2018-12-21 | 北京三快在线科技有限公司 | Target object recommended method and device, electronic equipment, storage medium |
CN109284413A (en) * | 2018-09-07 | 2019-01-29 | 平安科技(深圳)有限公司 | Method of Commodity Recommendation, device, equipment and storage medium based on recognition of face |
CN109766491A (en) * | 2018-12-18 | 2019-05-17 | 深圳壹账通智能科技有限公司 | Product search method, device, computer equipment and storage medium |
CN110619076A (en) * | 2018-12-25 | 2019-12-27 | 北京时光荏苒科技有限公司 | Search term recommendation method and device, computer and storage medium |
CN111008336A (en) * | 2019-12-23 | 2020-04-14 | 腾讯科技(深圳)有限公司 | Content recommendation method, device and equipment and readable storage medium |
CN111429204A (en) * | 2020-03-10 | 2020-07-17 | 携程计算机技术(上海)有限公司 | Hotel recommendation method, system, electronic equipment and storage medium |
CN111444428A (en) * | 2020-03-27 | 2020-07-24 | 腾讯科技(深圳)有限公司 | Information recommendation method and device based on artificial intelligence, electronic equipment and storage medium |
Non-Patent Citations (10)
Title |
---|
LEI ZHOU 等: "Personalized Recommendation Algorithm Based on User Preference and User Profile", pages 1202, Retrieved from the Internet <URL:https://link.springer.com/chapter/10.1007/978-3-030-44038-1_110> * |
ZHICHAO CHANG 等: "Hotel Recommendation Based on Surrounding Environments", 2013 SECOND IIAI INTERNATIONAL CONFERENCE ON ADVANCED APPLIED INFORMATICS, 15 October 2013 (2013-10-15), pages 330 - 336 * |
何溢曼 等: "基于Android的美食推荐系统设计与实现", 数码世界, no. 9, pages 24 * |
宋丹丹;: "基于微信小程序的美食点餐推荐系统的设计与实现", 信息通信, no. 12, 15 December 2017 (2017-12-15), pages 89 - 90 * |
山间一棵松: "基于大数据的餐饮推荐系统总结", pages 1, Retrieved from the Internet <URL:https://www.cnblogs.com/share23/articles/9589239.html> * |
岳子静 等: "用户饮食偏好挖掘及应用研究", 图书情报知识, no. 03, 10 May 2019 (2019-05-10), pages 81 - 90 * |
岳子静 等: "用户饮食偏好挖掘及应用研究", 图书情报知识, no. 03, pages 81 - 90 * |
杨晓蕾 等: "基于张量分解的多维信息融合兴趣点推荐算法", 小型微型计算机系统, no. 05, 15 May 2020 (2020-05-15), pages 902 - 907 * |
范顺忠: "基于用户兴趣感知的个性化美食推荐算法研究与应用", 中国优秀硕士学位论文全文数据库 信息科技辑, no. 7, 15 July 2018 (2018-07-15), pages 138 - 1977 * |
郑尔财: "基于情感分析的个性化推荐算法研究", 中国优秀硕士学位论文全文数据库 信息科技辑, no. 1, 15 January 2019 (2019-01-15), pages 138 - 5295 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112667714A (en) * | 2021-03-17 | 2021-04-16 | 腾讯科技(深圳)有限公司 | User portrait optimization method and device based on deep learning and storage medium |
CN113570432A (en) * | 2021-07-28 | 2021-10-29 | 北京达佳互联信息技术有限公司 | Resource recommendation determining method, device, equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108829822B (en) | Media content recommendation method and device, storage medium and electronic device | |
CN112214685B (en) | Knowledge graph-based personalized recommendation method | |
CN111061946B (en) | Method, device, electronic equipment and storage medium for recommending scenerized content | |
CN111061945B (en) | Recommendation method, recommendation device, electronic equipment and storage medium | |
US20220171760A1 (en) | Data processing method and apparatus, computer-readable storage medium, and electronic device | |
CN111061962A (en) | Recommendation method based on user score analysis | |
CN112052326A (en) | Intelligent question and answer method and system based on long and short text matching | |
CN110413888B (en) | Book recommendation method and device | |
CN103198089A (en) | Information processing apparatus, information processing method, and program | |
CN110795571A (en) | Cultural tourism resource recommendation method based on deep learning and knowledge graph | |
CN112966091B (en) | Knowledge map recommendation system fusing entity information and heat | |
Liu et al. | Using collaborative filtering algorithms combined with Doc2Vec for movie recommendation | |
CN115131698B (en) | Video attribute determining method, device, equipment and storage medium | |
CN110019650A (en) | Method, apparatus, storage medium and the electronic equipment of search associational word are provided | |
CN112052388A (en) | Method and system for recommending gourmet stores | |
CN115964560B (en) | Information recommendation method and equipment based on multi-mode pre-training model | |
CN110688565A (en) | Next item recommendation method based on multidimensional Hox process and attention mechanism | |
CN110110218A (en) | A kind of Identity Association method and terminal | |
CN116910375B (en) | Cross-domain recommendation method and system based on user preference diversity | |
CN116956183A (en) | Multimedia resource recommendation method, model training method, device and storage medium | |
CN112989177A (en) | Information processing method, information processing device, electronic equipment and computer storage medium | |
CN116957128A (en) | Service index prediction method, device, equipment and storage medium | |
CN116257618A (en) | Multi-source intelligent travel recommendation method based on fine granularity emotion analysis | |
Joseph et al. | A Comparative Study of Collaborative Movie Recommendation System | |
CN113849518A (en) | Assisted search query |
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 |