CN116012115A - Crowd interest dynamic labeling method based on commodity map and supply chain - Google Patents
Crowd interest dynamic labeling method based on commodity map and supply chain Download PDFInfo
- Publication number
- CN116012115A CN116012115A CN202310100205.6A CN202310100205A CN116012115A CN 116012115 A CN116012115 A CN 116012115A CN 202310100205 A CN202310100205 A CN 202310100205A CN 116012115 A CN116012115 A CN 116012115A
- Authority
- CN
- China
- Prior art keywords
- commodity
- similarity
- data
- influence
- interest
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a crowd interest dynamic labeling method based on a commodity map and a supply chain, which comprises the following steps of S1, judging whether browsing data and collection data of a user can be obtained, if yes, turning to S2, if not, obtaining age data and interest commodity type data of the user, and configuring a first label for the user; s2, acquiring browsing data and collection data; step S3, calculating a first similarity among all browsed commodities and a second similarity among all collected commodities, and calculating to obtain historical interestingness according to the first similarity, browsing duration and the second similarity; s3, constructing a knowledge graph model according to the historical interestingness and the commodity model second label; s4, retraining the commodity spectrum model according to the commodity influence index and the supply chain influence index to obtain an optimized commodity spectrum model; and S5, optimizing the commodity map model to configure a second label for the user according to the interestingness. The method and the device improve the accuracy of marking the label according to the user interest.
Description
Technical Field
The invention relates to the technical field of content pushing, in particular to a crowd interest dynamic labeling method based on a commodity map and a supply chain.
Background
The Knowledge map (knowledgegraph), called Knowledge domain visualization or Knowledge domain mapping map in book condition report, is a series of various graphs showing Knowledge development process and structural relationship, and uses visualization technology to describe Knowledge resources and their carriers, and excavate, analyze, construct, draw and display Knowledge and their interrelationships. The knowledge graph for electronic commerce is called a commodity graph. The commodity atlas is generally applied to a shopping website or a commodity recommendation system of shopping software. The Supply chain (Supply chain) refers to the entire chain of products from the merchant to the consumer during production and distribution, involving the network of businesses upstream and downstream of the end user's activities to which the products or services are provided. The elements in the supply chain may affect the user's interest in the product, and when the supply chain changes, the user's interest level in the same product may also change. At present, the existing commodity recommendation system generally analyzes browsing records and collection records of users, recommends approximate commodities to the users according to the browsing records and the collection records, and performs interest tags on the users according to the types and the quantity of the commodities in the browsing records and the collection records. However, the interest tag mode is static and fixed, and the user can simultaneously change the demands and interests of the commodity along with the change of various influence indexes such as actual demands, characters and the like, so that the existing user interest tag mode cannot meet the dynamic change demands of the user on the interests of the commodity.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a crowd interest dynamic labeling method based on commodity graphs and supply chains, which is used for dynamically adjusting a knowledge graph model and improving the accuracy of labeling according to the user interests.
In order to achieve the above purpose, the present invention provides the following technical solutions: a crowd interest dynamic labeling method based on commodity atlas and supply chain comprises the following steps:
step S1, judging whether browsing data and collection data of a user can be obtained, if yes, turning to step S2, if not, obtaining age data and interest commodity type data of the user, and configuring a first label corresponding to a commodity model for the user according to the age data and the interest commodity type data;
s2, acquiring the browsing data and the collection data, wherein the browsing data comprises a plurality of pieces of browsing commodity information and corresponding browsing duration, and the collection data comprises a plurality of pieces of collection commodity information;
step S3, respectively calculating a first similarity between browsed commodities corresponding to the browsed commodity information and a second similarity between collected commodities corresponding to the collected commodity information, and calculating to obtain historical interestingness of a user on each commodity according to the first similarity, the browsed duration and the second similarity;
s4, constructing a knowledge graph model according to the historical interestingness, the corresponding commodity model and the second label;
s5, acquiring commodity influence indexes and supply chain influence indexes in real time, adjusting historical interestingness according to the commodity influence indexes and the supply chain influence indexes to obtain real-time interestingness, and retraining the knowledge graph model according to the real-time interestingness to obtain an optimized commodity graph model;
and S6, configuring the second label for the user according to the commodity model by the optimized commodity map model.
Further, the step S1 includes:
step S11, the association relationship among the user age data, the interest commodity data and the corresponding first label is pre-stored in a storage module;
step S12, acquiring the age data of the user and the type data of the interesting commodity in real time;
and step S13, matching the first label corresponding to the user age data and the interest commodity type data in the storage module according to the association relation, and outputting the first label.
Further, the step S3 includes:
step S31, obtaining first commodity characteristic information of each browsed commodity and second commodity characteristic information of each collected commodity;
step S32, calculating the similarity between the first commodity feature information to obtain the first similarity, and calculating the similarity between the second commodity feature information to obtain the second similarity;
step S33, allocating the first similarity, the second similarity and the browsing duration to corresponding coefficients respectively, and inputting the coefficients into a preset interest calculation formula to obtain the historical interest level.
Further, the first merchandise feature information and the second merchandise feature information each include merchandise brands, merchandise prices, and merchandise configuration data, and the step S32 specifically includes:
and respectively calculating the brand similarity of the commodity brands, the price similarity of the commodity prices and the configuration similarity of the commodity configuration, and inputting the brand similarity, the price similarity and the configuration similarity into a preset similarity calculation formula to obtain the first similarity or the second similarity.
Further, the similarity calculation formula is configured to:
wherein S is i For representing the first similarity or the second similarity;
B r for representing the brand similarity;
P r for representing the price similarity;
D i for representing the configuration similarity;
k 1 、k 2 and k 3 The method is used for respectively representing a preset first similarity coefficient, a preset second similarity coefficient and a preset third similarity coefficient, wherein the first similarity coefficient, the second similarity coefficient and the preset third similarity coefficient are positive numbers.
Further, the interest calculation formula is configured to:
ρ 3 >ρ 2 >ρ 1 >0；
wherein I is n For representing the interestingness;
T B representing the browsing duration;
S i1 for representing the first similarity;
S i2 for representing the second similarity;
ρ 1 、ρ 2 and ρ 3 Respectively used for representing a preset first configuration coefficient, a preset second configuration coefficient and a preset third configuration coefficient.
Further, the step S5 includes:
step S51, processing according to the brand influence index, the price influence index and the configuration influence index to obtain a commodity influence index;
step S52, processing the influence indexes of the raw material suppliers, the influence indexes of the manufacturers, the influence indexes of the intermediate suppliers and the influence indexes of the express companies to obtain a supply chain influence index;
step S53, inputting the commodity influence index and the supply chain influence index into a preset state influence equation to obtain a comprehensive influence index, and multiplying the comprehensive influence index by the historical interestingness to obtain the real-time interestingness;
and S54, re-training the knowledge graph model by taking the real-time interestingness as input quantity to obtain the optimized knowledge graph model.
Further, the state impact equation is configured to:
f(x)＝(B r i,P r i,D i i) Τ ；
g(x)＝(R a i,M a i,M i i,E x i) Τ ；
wherein S is y For representing the integrated impact index;
B r i is used to denote theBrand impact index;
P r i is used to represent the price impact index;
D i i is used to represent the configuration impact index;
f (x) is used to represent the commodity impact index;
R a i is used to represent the raw material supplier impact index;
M a i is used to represent the manufacturer impact index;
M i i is used to represent the intermediate quotient influencing exponent;
E x and i is used for representing the express company influence index.
Further, the step S2 further includes filtering and cleaning the invalid data, the abnormal data and the unsteady data in the browsing data and the collecting data, so as to obtain the browsing data and the collecting data after cleaning.
The invention has the beneficial effects that:
according to the method, when browsing data and collecting data of a user cannot be obtained, age data and interest commodity type data of the user are obtained, a first label corresponding to a commodity model is configured for the user according to the age data and the interest commodity type data, meanwhile, when the browsing data and the collecting data can be obtained, first similarity between browsing commodities and second similarity between collecting commodities are calculated, historical interest of the commodities is calculated according to the first similarity, the second similarity and browsing time, and then a knowledge graph model is built according to the historical interest; and finally, the knowledge graph model is retrained through the commodity influence index and the supply chain influence index which are acquired in real time to obtain an optimized commodity graph model, so that the dynamic adjustment of the knowledge graph model is realized, the high timeliness and the prediction accuracy of model prediction are improved, and the accuracy of labeling labels according to the user interests is further improved.
Drawings
FIG. 1 is a flow chart of the steps of the method for dynamic tagging of group interests in the present invention;
FIG. 2 is a sub-flowchart of step S1 of the present invention;
FIG. 3 is a sub-flowchart of step S3 of the present invention;
fig. 4 is a sub-flowchart of step S4 in the present invention.
Detailed Description
The invention will now be described in further detail with reference to the drawings and examples. Wherein like parts are designated by like reference numerals. It should be noted that the words "front", "back", "left", "right", "upper" and "lower" used in the following description refer to directions in the drawings, and the words "bottom" and "top", "inner" and "outer" refer to directions toward or away from, respectively, the geometric center of a particular component.
As shown in fig. 1, the method for dynamically labeling crowd interests based on a commodity map and a supply chain according to the present embodiment includes:
step S1, judging whether browsing data and collection data of a user can be obtained, if yes, turning to step S2, if not, obtaining age data and interest commodity type data of the user, and configuring a first label corresponding to the commodity model for the user according to the age data and the interest commodity type data;
step S2, browsing data and collection data are obtained, wherein the browsing data comprise a plurality of pieces of browsing commodity information and corresponding browsing duration, and the collection data comprise a plurality of pieces of collection commodity information;
step S3, respectively calculating a first similarity between browsed commodities corresponding to the browsed commodity information and a second similarity between corresponding collected commodities in the collected commodity information, and calculating to obtain historical interest degrees of users on the commodities according to the first similarity, the browsing duration and the second similarity;
s4, constructing a knowledge graph model according to the historical interestingness, the corresponding commodity model and the second label;
s5, acquiring commodity influence indexes and supply chain influence indexes in real time, adjusting historical interestingness according to the commodity influence indexes and the supply chain influence indexes to obtain real-time interestingness, and retraining a knowledge graph model according to the real-time interestingness to obtain an optimized commodity graph model;
and S6, optimizing the commodity map model to configure a second label for the user according to the commodity model.
Specifically, in the present embodiment, it is necessary to determine whether the user is a new user or a guest who is not logged in by determining whether browsing data and collection data of the user can be acquired. When the browsing data and the collecting data of the user can be acquired, the current user is indicated to be an old user after login, and when the browsing data or the collecting data of the user can not be acquired, the current user is indicated to be a new user which is not logged in. When the current user is a new user or a visitor is not logged in, the user is required to select an age group or input an age to acquire age data of the user, the user is required to check the age data to acquire interest commodity type data, the age data and the interest commodity type data are input into a preset commodity recommendation model, a predicted interest commodity containing a commodity model is output, and a corresponding first label is configured for the user according to the commodity model. And when the browsing data and the collection data of the current user can be acquired, indicating that the current user is an old user. The historical interest degree of the old user on each commodity is needed to be calculated according to the browsing data and the collection data of the old user: firstly, calculating a first similarity among browsing commodities corresponding to information of all browsing commodities contained in browsing data and a second similarity among all collecting commodities contained in collecting data, and calculating historical interestingness of a user on all commodities according to the first similarity, the second similarity and browsing duration when models of the browsing commodities and the collecting commodities are the same. And constructing a knowledge graph model according to the historical interestingness, the corresponding commodity model and the second label, introducing the obtained commodity influence index and supply chain influence index, adjusting the historical interestingness by utilizing the commodity influence index and the supply chain influence index to obtain the real-time interestingness, retraining the knowledge graph model according to the real-time interestingness to obtain an optimized commodity graph model, so that the optimized commodity graph model configures the second label corresponding to the commodity model for a user, and realizing quick and accurate marking of the user interests. According to the method and the device, the knowledge graph model is dynamically adjusted, so that the high timeliness and the prediction accuracy of model prediction are improved, and the accuracy of labeling according to the user interests is further improved.
Preferably, as shown in fig. 2, step S1 includes:
step S11, the association relationship among the age data of each user, the data of each interest commodity and the corresponding first label is pre-stored in the storage module;
step S12, acquiring user age data and interest commodity type data in real time;
and S13, matching in a storage module according to the association relationship to obtain a first label corresponding to the user age data and the interest commodity type data, and outputting the first label.
Specifically, in this embodiment, after the age data and the interest commodity type data are obtained, the corresponding first tag may be obtained by matching in the storage module, so that the interest tag configuration for the user is conveniently and rapidly implemented.
Preferably, as shown in fig. 3, step S3 includes:
step S31, acquiring first commodity characteristic information of each browsed commodity and second commodity characteristic information of each collected commodity;
step S32, calculating the similarity between the first commodity characteristic information to obtain first similarity, and calculating the similarity between the second commodity characteristic information to obtain second similarity;
step S33, the first similarity, the second similarity and the browsing duration are respectively distributed with corresponding coefficients and input into a preset interest calculation formula, and the historical interest is obtained.
Preferably, the first merchandise feature information and the second merchandise feature information each include a merchandise brand, a merchandise price, and merchandise configuration data, and step S32 specifically includes:
and respectively calculating the brand similarity of the commodity brands, the price similarity of the commodity prices and the configuration similarity of the commodity configuration, and inputting the brand similarity, the price similarity and the configuration similarity into a preset similarity calculation formula to obtain the first similarity or the second similarity.
Preferably, the similarity calculation formula is configured to:
wherein S is i For representing the first similarity or the second similarity;
B r for representing brand similarity;
P r for representing price similarity;
D i for representing configuration similarity;
k 1 、k 2 and k 3 The method is used for respectively representing a preset first similarity coefficient, a preset second similarity coefficient and a preset third similarity coefficient, wherein the first similarity coefficient, the second similarity coefficient and the preset third similarity coefficient are positive numbers.
Preferably, the interest calculation formula is configured to:
ρ 3 >ρ 2 >ρ 1 >0；
wherein I is n For representing the interest level;
T B representing browsing duration;
S i1 for representing a first degree of similarity;
S i2 for representing a second degree of similarity;
ρ 1 、ρ 2 and ρ 3 Respectively used for representing a preset first configuration coefficient, a preset second configuration coefficient and a preset third configuration coefficient.
Specifically, in this embodiment, if the types of the browsed commodity and the collected commodity are different, the interestingness needs to be calculated respectively, and the calculation formula of the interestingness for the browsed commodity is I n ＝ρ 1 S i1 +ρ 2 T B The method comprises the steps of carrying out a first treatment on the surface of the The interest degree calculation formula for the collected commodity is as follows:
preferably, as shown in fig. 4, step S5 includes:
step S51, processing according to the brand influence index, the price influence index and the configuration influence index to obtain a commodity influence index;
step S52, processing the influence indexes of the raw material suppliers, the influence indexes of the manufacturers, the influence indexes of the intermediate suppliers and the influence indexes of the express companies to obtain a supply chain influence index;
step S53, inputting the commodity influence index and the supply chain influence index into a preset state influence equation to obtain a comprehensive influence index, and multiplying the comprehensive influence index by the historical interestingness to obtain a real-time interestingness;
and S54, retraining the knowledge graph model by taking the real-time interestingness as input quantity to obtain an optimized knowledge graph model.
Specifically, in this embodiment, the interest level of a user in a commodity changes in real time with the attribute of the commodity and the change of the supply chain. The commodity influence index is used for representing the influence degree of the attribute of the commodity on the user interest, and is influenced by the brand influence index, the price influence index and the configuration influence index, and the commodity influence index is obtained through processing according to the brand influence index, the price influence index and the configuration influence index; the supply chain influence index is used for representing the influence degree of the supply chain on the user interest, and is influenced by the raw material supplier influence index, the manufacturer influence index, the intermediate influence index and the express company influence index, and the supply chain influence factors are obtained through processing according to the raw material supplier influence index, the manufacturer influence index, the intermediate influence index and the express company influence index. After the commodity influence index and the supply chain influence factor are obtained through processing, the comprehensive influence index for reflecting the comprehensive influence degree of the commodity influence index and the supply chain influence factor on the commodity interest degree can be calculated through inputting the state influence equation. The real-time interestingness is obtained by multiplying the comprehensive influence index by the historical interestingness, and finally the knowledge graph model is retrained by replacing the historical interestingness with the real-time interestingness to obtain an optimized commodity graph model, so that the dynamic adjustment of the knowledge graph model is realized, the high timeliness and the prediction accuracy of model prediction are improved, and the accuracy of labeling labels according to the user interests is further improved.
Preferably, the state influence equation is configured to:
f(x)＝(B r i,P r i,D i i) Τ ；
g(x)＝(R a i,M a i,M i i,E x i) Τ ；
wherein S is y For representing the integrated impact index;
B r i is used to represent a brand impact index;
P r i is used to represent a price impact index;
D i i is used to represent a configuration impact index;
f (x) is used to represent a commodity impact index;
R a i is used to represent the raw material supplier impact index;
M a i is used to represent the manufacturer impact index;
M i i is used to represent the intermediate quotient influencing exponent;
E x i is used to represent the express company impact index.
Preferably, step S2 further includes filtering and cleaning invalid data, abnormal data and unsteady data in the browsing data and the collecting data, so as to obtain the browsing data and the collecting data after cleaning.
Specifically, in this embodiment, after the browsing data and the collecting data are obtained, invalid data rejection, abnormal data rejection and unsteady state data filtering are required to be performed on the browsing data and the collecting data, where the invalid data rejection and the abnormal data rejection are implemented through preset invalid data filtering rules and abnormal data filtering criteria. And the unsteady state data in the browsing data and the collecting data are filtered and cleaned according to the six sigma rule, so that the effective screening of the browsing data and the collecting data is realized, and the subsequent operation quantity is reduced.
The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above examples, and all technical solutions belonging to the concept of the present invention belong to the protection scope of the present invention. It should be noted that modifications and adaptations to the present invention may occur to one skilled in the art without departing from the principles of the present invention and are intended to be within the scope of the present invention.
Claims (9)
1. The utility model provides a crowd interest dynamic labeling method based on commodity atlas and supply chain, which is characterized by comprising the following steps:
step S1, judging whether browsing data and collection data of a user can be obtained, if yes, turning to step S2, if not, obtaining age data and interest commodity type data of the user, and configuring a first label corresponding to a commodity model for the user according to the age data and the interest commodity type data;
s2, acquiring the browsing data and the collection data, wherein the browsing data comprises a plurality of pieces of browsing commodity information and corresponding browsing duration, and the collection data comprises a plurality of pieces of collection commodity information;
step S3, respectively calculating a first similarity between browsed commodities corresponding to the browsed commodity information and a second similarity between collected commodities corresponding to the collected commodity information, and calculating to obtain historical interestingness of a user on each commodity according to the first similarity, the browsed duration and the second similarity;
s4, constructing a knowledge graph model according to the historical interestingness, the corresponding commodity model and the second label;
s5, acquiring commodity influence indexes and supply chain influence indexes in real time, adjusting historical interestingness according to the commodity influence indexes and the supply chain influence indexes to obtain real-time interestingness, and retraining the knowledge graph model according to the real-time interestingness to obtain an optimized commodity graph model;
and S6, configuring the second label for the user according to the commodity model by the optimized commodity map model.
2. The crowd interest dynamic tagging method of claim 1, wherein: the step S1 includes:
step S11, the association relationship among a plurality of user age data, a plurality of interest commodity category data and a plurality of commodity price interval data and corresponding first labels is pre-stored in a storage module;
step S12, acquiring the user age data, the interest commodity kind data and the commodity price interval data in real time;
and step S13, matching the user age data, the interest commodity type data and the first label of the commodity corresponding to the commodity price interval data in the storage module according to the association relation.
3. The crowd interest dynamic tagging method of claim 1, wherein: the step S3 includes:
step S31, obtaining first commodity characteristic information of each browsed commodity and second commodity characteristic information of each collected commodity;
step S32, calculating the similarity between the first commodity feature information to obtain the first similarity, and calculating the similarity between the second commodity feature information to obtain the second similarity;
step S33, allocating the first similarity, the second similarity and the browsing duration to corresponding coefficients respectively, and inputting the coefficients into a preset interest calculation formula to obtain the historical interest level.
4. A crowd interest dynamic tagging method as claimed in claim 3, wherein: the step S32 specifically includes:
and respectively calculating the brand similarity of the commodity brands, the price similarity of the commodity prices and the configuration similarity of the commodity configuration, and inputting the brand similarity, the price similarity and the configuration similarity into a preset similarity calculation formula to obtain the first similarity or the second similarity.
5. The crowd interest dynamic tagging method of claim 4, wherein: the similarity calculation formula is configured as follows:
wherein S is i For representing the first similarity or the second similarity;
B r for representing the brand similarity;
P r for representing the price similarity;
D i for representing the configuration similarity;
k 1 、k 2 and k 3 The method is used for respectively representing a preset first similarity coefficient, a preset second similarity coefficient and a preset third similarity coefficient, wherein the first similarity coefficient, the second similarity coefficient and the preset third similarity coefficient are positive numbers.
6. A crowd interest dynamic tagging method as claimed in claim 3, wherein: the interest calculation formula is configured to:
ρ 3 >ρ 2 >ρ 1 >0；
wherein I is n For representing the interestingness;
T B representing the browsing duration;
S i1 for use inRepresenting the first similarity;
S i2 for representing the second similarity;
ρ 1 、ρ 2 and ρ 3 Respectively used for representing a preset first configuration coefficient, a preset second configuration coefficient and a preset third configuration coefficient.
7. The crowd interest dynamic tagging method of claim 1, wherein: the step S5 includes:
step S51, processing according to the brand influence index, the price influence index and the configuration influence index to obtain a commodity influence index;
step S52, processing the influence indexes of the raw material suppliers, the influence indexes of the manufacturers, the influence indexes of the intermediate suppliers and the influence indexes of the express companies to obtain a supply chain influence index;
step S53, inputting the commodity influence index and the supply chain influence index into a preset state influence equation to obtain a comprehensive influence index, and multiplying the comprehensive influence index by the historical interestingness to obtain the real-time interestingness;
and S54, re-training the knowledge graph model by taking the real-time interestingness as input quantity to obtain the optimized knowledge graph model.
8. The crowd interest dynamic tagging method of claim 7, wherein: the state impact equation is configured to:
f(x)＝(B r i,P r i,D i i) Τ ；
g(x)＝(R a i,M a i,M i i,E x i) Τ ；
wherein S is y For representing the integrated impact index;
B r i is used to represent the brand impact index;
P r i is used to represent the price impact index;
D i i is used to represent the configuration impact index;
f (x) is used to represent the commodity impact index;
R a i is used to represent the raw material supplier impact index;
M a i is used to represent the manufacturer impact index;
M i i is used to represent the intermediate quotient influencing exponent;
E x i is used for representing the express company influence index;
f 0 for representing a preset commodity influence constant;
g 0 for representing a preset supply chain impact constant.
9. The crowd interest dynamic tagging method of claim 1, wherein: and step S2 further comprises filtering and cleaning invalid data, abnormal data and unsteady state data in the browsing data and the collecting data to obtain the browsing data and the collecting data after cleaning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310100205.6A CN116012115A (en) | 2023-02-06 | 2023-02-06 | Crowd interest dynamic labeling method based on commodity map and supply chain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310100205.6A CN116012115A (en) | 2023-02-06 | 2023-02-06 | Crowd interest dynamic labeling method based on commodity map and supply chain |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116012115A true CN116012115A (en) | 2023-04-25 |
Family
ID=86021117
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310100205.6A Pending CN116012115A (en) | 2023-02-06 | 2023-02-06 | Crowd interest dynamic labeling method based on commodity map and supply chain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116012115A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116861077A (en) * | 2023-06-25 | 2023-10-10 | 北京信大融金教育科技有限公司 | Product recommendation method, device, equipment and storage medium based on supply chain system |
-
2023
- 2023-02-06 CN CN202310100205.6A patent/CN116012115A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116861077A (en) * | 2023-06-25 | 2023-10-10 | 北京信大融金教育科技有限公司 | Product recommendation method, device, equipment and storage medium based on supply chain system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110222272B (en) | Potential customer mining and recommending method | |
| CN106097076B (en) | Commodity sales management system | |
| CN111401777B (en) | Enterprise risk assessment method, enterprise risk assessment device, terminal equipment and storage medium | |
| CN112381623A (en) | Electronic commerce platform commodity intelligent recommendation method based on big data analysis | |
| CN107301592A (en) | The method and device excavated for commodity substitute | |
| CN111047412A (en) | Big data electricity merchant operation platform | |
| CN113538012A (en) | Buyer user intelligent management method based on internet e-commerce platform data analysis | |
| BenMark et al. | How retailers can drive profitable growth through dynamic pricing | |
| EP3400571A1 (en) | Consumer decision tree generation system | |
| CN116012115A (en) | Crowd interest dynamic labeling method based on commodity map and supply chain | |
| CN111666424A (en) | Steam part inquiry and quotation optimization method based on knowledge representation learning | |
| CN111861679A (en) | Commodity recommendation method based on artificial intelligence | |
| CN113781106B (en) | Commodity operation data analysis method, device, equipment and computer readable medium | |
| Apichottanakul et al. | Customer behaviour analysis based on buying-data sparsity for multi-category products in pork industry: A hybrid approach | |
| CN116452299A (en) | Intelligent recommendation system and method for electronic commerce | |
| CN106920108A (en) | A kind of method and system of commodity typing | |
| CN114266594A (en) | Big data analysis method based on southeast Asia cross-border e-commerce platform | |
| Huang et al. | Sales forecast for O2O services-based on incremental random forest method | |
| EP2933771A1 (en) | Method and system for identifying defective goods based on user purchasing behaviour | |
| CN114049141A (en) | Enterprise financial operation digital management optimization equipment based on data analysis | |
| CN117333269B (en) | Fresh flower order management method and system | |
| Yang et al. | Agri-food Products Live Streaming: Fad or A Rising Marketing Channel? | |
| CN116308466B (en) | Data information acquisition and intelligent analysis method, system, equipment and storage medium | |
| CN116151933B (en) | International trade information data digital supervision system and method based on big data | |
| CN112884072A (en) | Commodity data classification processing and comparison technical method and system |
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 |