CN116911926A - Advertisement marketing recommendation method based on data analysis - Google Patents
Advertisement marketing recommendation method based on data analysis Download PDFInfo
- Publication number
- CN116911926A CN116911926A CN202310757497.0A CN202310757497A CN116911926A CN 116911926 A CN116911926 A CN 116911926A CN 202310757497 A CN202310757497 A CN 202310757497A CN 116911926 A CN116911926 A CN 116911926A
- Authority
- CN
- China
- Prior art keywords
- commodity
- updated
- type
- types
- recommendation
- 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 42
- 238000007405 data analysis Methods 0.000 title claims abstract description 20
- 230000000630 rising effect Effects 0.000 claims abstract description 36
- 238000012545 processing Methods 0.000 claims abstract description 4
- 238000013528 artificial neural network Methods 0.000 claims description 19
- 230000001174 ascending effect Effects 0.000 claims description 15
- 238000013527 convolutional neural network Methods 0.000 claims description 15
- 230000008859 change Effects 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 230000004927 fusion Effects 0.000 claims description 2
- 230000001105 regulatory effect Effects 0.000 abstract 1
- 238000004458 analytical method Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000006399 behavior Effects 0.000 description 4
- 239000002537 cosmetic Substances 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005034 decoration Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0269—Targeted advertisements based on user profile or attribute
- G06Q30/0271—Personalized advertisement
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2323—Non-hierarchical techniques based on graph theory, e.g. minimum spanning trees [MST] or graph cuts
-
- 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
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Finance (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Accounting & Taxation (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Discrete Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of electronic digital data processing, in particular to an advertising marketing recommendation method based on data analysis, which comprises the following steps: the commodity in the consumption data is classified by acquiring the consumption data of a user in real time, leading parameters are obtained according to the number and the price of the commodity in the commodity category, a chain tree structure is built again, difference characteristics are obtained by calculating and judging the parameters, rising amplitude and rising speed are respectively obtained according to the difference characteristics and the consumption data, and finally the recommended importance degree of the commodity category is obtained according to the leading parameters, the rising amplitude and the rising speed. According to the invention, the quantity of different commodity types in the browsing page is regulated according to the recommendation importance degree, the accuracy of commodity advertisement marketing is improved, the marketing recommendation result can be changed in real time according to the requirement of a user, the real-time performance of advertisement recommendation is enhanced, the user is helped to better select required commodities, and the personalized recommendation of intelligent advertisement marketing is realized.
Description
Technical Field
The invention relates to the technical field of electronic digital data processing, in particular to an advertising marketing recommendation method based on data analysis.
Background
With the high-speed development of the Internet, the consumption modes of people are changed over the sky, the network consumption is more and more convenient, and more new consumption modes are continuously emerging. In order to open the popularity of the product, more products are recommended to the user, and it is generally necessary to recommend the product and deliver the advertisement to the user. When advertisement delivery and commodity recommendation are carried out, consumption data of a user are generally acquired, then the consumption data of the user are analyzed, shopping preference and consumption habit of the user are obtained, and then related products are pushed to the user. Because when pushing commodities according to consumption data of users, if personalized analysis is carried out on each user and then commodity pushing is carried out, the data volume required to be processed is very large, and in order to reduce the quantity required to be processed and carry out accurate commodity pushing personalized analysis on the commodities, the commodity pushing method and device take the commodities as guide and then analyze purchasing behaviors of the users to update the types of pushed commodities.
In the prior art, a big data technology is currently adopted to collect consumption information of a consumption user, and then consumption behavior habits of the user and preferences of the user are analyzed, so that related commodities are recommended to the user. The clustering algorithm is well applied to advertisement delivery and commodity pushing, and can be used for recommending commodities of a user according to shopping preferences and consumption habits of the user, but because the shopping types and habits of the user can change, if commodity recommendation and advertisement delivery are directly carried out according to historical consumption records, the commodity information can be updated untimely, the recommended commodities are not needed recently by the user, and advertisement delivery resources can be wasted.
The invention provides an advertisement marketing recommendation method based on data analysis, which takes commodities purchased by a user as a guide, updates commodity advertisement recommendation results of the user through purchase records updated by the user in real time, accurately recommends commodity information to the user, and realizes personalized recommendation of intelligent advertisement marketing.
Disclosure of Invention
The invention provides an advertising marketing recommendation method based on data analysis, which aims to solve the existing problems.
The advertisement marketing recommendation method based on data analysis adopts the following technical scheme:
the invention provides an advertising marketing recommendation method based on data analysis, which comprises the following steps:
updating in real time to obtain consumption data of a user, including: commodity name, consumption time, and commodity price;
acquiring commodity types of all commodities in the consumption data by using a neural network; obtaining dominant parameters of commodity types according to the quantity of commodities contained in each commodity type and commodity price; clustering all commodity types according to the quantity of the commodity contained in each commodity type and the distance between the average prices of the contained commodity to obtain a plurality of cluster clusters, constructing a tree structure, taking each cluster as each node of the tree structure according to the size and average value of the dominant parameters of the commodity types in the cluster clusters in sequence to obtain a chain-shaped tree structure, marking the node of the chain-shaped tree structure as a layer, and marking the commodity type corresponding to the maximum dominant parameter in any layer as a main type;
when the quantity of the commodities contained in the commodity category is changed and updated, the corresponding commodity category is marked as the updated commodity category, and corresponding dominant parameters are obtained; obtaining judgment parameters according to the difference between the type of the updated commodity and the dominant parameters of any main type, and obtaining the updated layer number of the type of the updated commodity according to the size of the judgment parameters; obtaining difference characteristics according to the commodity price and the quantity of the commodities contained in the main type; when the consumption data of the updated commodity type appears for the first time, the layer number in the chain tree structure is marked as an initial layer number, the result of adjusting the difference between the updated layer number and the initial layer number by the dominant parameter of the updated commodity type is marked as the rising amplitude of the updated commodity type, and the fusion result of the consumption time, the dominant parameter of the main type and the difference characteristic is marked as the rising speed of the updated commodity type;
and obtaining the recommendation importance degree of the updated commodity type according to the dominant parameters, the ascending amplitude and the ascending speed, and realizing intelligent advertisement marketing recommendation according to the recommendation importance degree.
Further, the method for obtaining the commodity category of each commodity in the consumption data by using the neural network comprises the following specific steps:
firstly, acquiring commodity pictures, acquiring commodity types of all commodities according to the division of the commodity types of an online mall, taking the commodity types and commodity names as manual labels of the corresponding commodity pictures, taking a commodity image with the manual labels as one sample, and forming a data set for training a CNN neural network by a large number of samples;
then, dividing the data set into a training set, a testing set and a verification set according to a preset proportion, inputting the training set, the testing set and the verification set into a CNN neural network, and training the CNN neural network by combining a cross entropy loss function to obtain a trained CNN neural network;
and finally, classifying all the commodities in the consumption data of the user by using the trained CNN neural network, and outputting commodity types corresponding to all the commodities to obtain a plurality of commodity types.
Further, the method for obtaining the dominant parameters of the commodity types according to the quantity of the commodities contained in each commodity type and the commodity price comprises the following specific steps:
wherein G is i N representing the dominant parameter of the ith commodity type in the consumption data i Representing the number of goods contained in the ith goods category in the consumption data, T ij Represents the commodity price of the j-th commodity in the i-th commodity category in the consumption data, m represents the number of commodity categories in the consumption data, n i Indicating the number of items included in the ith item category.
Further, the chain tree structure is obtained by the following steps:
step (1), presetting a K value to be 5 according to experience, and clustering commodities in consumption data by using a K-means algorithm according to the quantity of commodities contained in each commodity type in the consumption data and the distance between average prices of the contained commodities to obtain a plurality of cluster clusters;
step (2), firstly, constructing a chain-shaped tree structure from top to bottom, wherein the tree structure has 5 nodes;
then, the cluster where the commodity category corresponding to the maximum value of the dominant parameter is located is used as the node at the uppermost layer in the tree structure, namely the 1 st node of the chain-shaped tree structure;
and finally, calculating the average value of all commodity types corresponding to the dominant parameters contained in the residual cluster, respectively serving as the 2 nd to 5 th nodes of the chain tree structure in the sequence from big to small, and sequentially arranging the obtained chain tree structure serving as a commodity recommendation model.
Further, the method for obtaining the judgment parameters according to the difference between the type of the updated commodity and the dominant parameters of any main type, and obtaining the updated layer number of the type of the updated commodity according to the size of the judgment parameters comprises the following specific steps:
firstly, recording a difference value between a dominant parameter of the updated commodity type and a dominant parameter of a corresponding main type of an arbitrary layer as a judging parameter of the updated commodity type;
then, starting from the 5 th layer of the chain tree structure, sequentially reducing and acquiring corresponding judgment parameters, and if the judgment parameters are larger than 0, classifying the updated commodity types into the layer number; if the judgment parameter is smaller than 0, the layer number is iterated upwards, and the judgment is continued until the layer number is 1, so that the updated layer number of the updated commodity type is obtained.
Further, the rising amplitude is obtained by the following steps:
the absolute value of the difference between the number of layers after updating and the number of initial layers in the tree structure of the updated commodity type is recorded as the layer change characteristic;
and (3) marking the product result of the dominant parameter of the updated commodity type and the layer number change characteristic as the rising amplitude of the updated commodity type.
Further, the rising speed is obtained by the following steps:
firstly, obtaining the difference characteristic between main types corresponding to the initial layer number when the consumption data of the updated commodity types appears for the first time, and marking the difference characteristic as D'; recording dominant parameters of the main type corresponding to the number of updated layers where the updated commodity types are located as Gg; recording dominant parameters of the main type corresponding to the initial layer number where the updated commodity type is located as Gc; when the consumption data change is updated, the corresponding consumption time of the updated commodity type is recorded as updated time; when corresponding consumption data appears for the first time, the corresponding consumption time of the updated commodity type is recorded as initial time;
then, the rising speed of any updated commodity kind is obtained:
wherein v represents the rising speed of the updated commodity type in the consumption data, and D' represents the difference characteristic between the updated commodity type and the main type corresponding to the updated layer number; d' represents a difference feature between the updated commodity type and the primary type corresponding to the initial layer number; gg represents the dominant parameter of the main type corresponding to the number of layers after updating where the updated commodity type is located, gc represents the dominant parameter of the main type corresponding to the initial number of layers where the updated commodity type is located, t' represents the time after updating, and t represents the initial time; the absolute value is acquired.
Further, the method obtains the recommendation importance degree of the updated commodity type according to the dominant parameter, the ascending amplitude and the ascending speed, and realizes the intelligent advertisement marketing recommendation according to the recommendation importance degree, comprising the following specific steps:
firstly, recording the ascending amplitude, ascending speed and accumulated result of the updated commodity type as the recommended importance degree of the updated commodity type, and obtaining the recommended importance degree of all the updated commodity types;
and then, carrying out normalization processing on the recommendation importance degrees of all the updated commodity types by using a linear normalization method, taking the normalization result as the recommendation probability of the same commodity type when a user browses on an e-commerce platform, and carrying out marketing recommendation on commodity advertisements corresponding to the commodity types in the platform according to the recommendation probability, wherein the larger the recommendation probability is, the more commodity numbers corresponding to the commodity types are on a browsing page, so as to realize intelligent advertisement marketing recommendation.
The technical scheme of the invention has the beneficial effects that: the method has the advantages that the dominant parameters of various commodity types are obtained by analyzing the consumption data of the user, the number of layers of the built chain tree structure is combined with the consumption data which are updated and changed in real time, the number of layers of the commodity types are changed in real time, the recommendation importance degree of the commodity types is obtained, the number of different commodity types in a browsing page is adjusted according to the recommendation importance degree, the accuracy of commodity advertisement marketing is improved, meanwhile, the marketing recommendation result can be changed in real time according to the requirement of the user, the instantaneity of advertisement recommendation is enhanced, the user is helped to better select required commodities, and personalized recommendation of advertisement intelligent marketing is realized.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart illustrating the steps of the advertising marketing recommendation method based on data analysis according to the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description is given below of the advertisement marketing recommendation method based on data analysis according to the invention, which is provided by combining the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the advertisement marketing recommendation method based on data analysis provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, a flowchart illustrating steps of a data analysis-based advertising marketing recommendation method according to an embodiment of the present invention is shown, the method includes the steps of:
and S001, acquiring data, and acquiring consumption data of a user.
When advertisement delivery and commodity recommendation are carried out, consumption data of a user are generally acquired, then the consumption data of the user are analyzed, shopping preference and consumption habit of the user are obtained, and then related products are pushed to the user.
Because when pushing the commodity according to the consumption data of the user, if personalized analysis is performed on each user and then the commodity is pushed, the amount of data to be processed is very large, and in order to reduce the amount to be processed and accurately push the commodity for personalized analysis, the commodity is used as a guide in the embodiment, and then the purchasing behavior of the user is analyzed to update the type of the pushed commodity.
Therefore, the consumption data of the user needs to be collected, when the consumption data of the user is collected, the consumption data of the user is called through a background server of the electronic commerce platform, the consumption data comprises the contents of commodity names, consumption time, commodity prices and the like, and then the obtained consumption data is arranged according to the order of the consumption time.
And step S002, analyzing the commodity types in the consumption data to obtain dominant parameters of the commodity types, and constructing a chain tree structure according to the dominant parameters.
The main purpose of this embodiment is to accurately deliver advertisements and push goods to users according to consumption data, because under normal conditions, when the shopping platform pushes goods to users, the shopping platform firstly analyzes shopping behaviors and shopping preferences of users, and then pushes goods according to shopping preferences and consumption level of users;
however, the shopping preferences may change stepwise when the user makes a purchase, for example: in the historical consumption data of a certain woman, shopping is more prone to clothes, cosmetic and household goods, but because the house is decorated recently, a lot of decoration-related materials, furniture and other goods need to be purchased, if the historical consumption data is directly analyzed, the clothes, cosmetic and household goods are definitely dominant, and when the goods are recommended according to shopping preferences, more clothes, cosmetic and household goods are recommended, so that the recommended goods and advertisements are put in much lower in efficiency, and the advertisement resources are wasted.
Therefore, the embodiment takes the commodities purchased by the user as the guide, analyzes the shopping preferences of the user, clusters the commodities, performs incremental clustering according to the real-time shopping records of the user, updates the data set of the commodities purchased by the user in real time, obtains the shopping preferences of the current user according to the updated data set, and further performs advertisement delivery and commodity recommendation according to the shopping preferences of the user.
Step (1), when describing the commodity purchased by the user, firstly acquiring the type of the commodity purchased by the user and classifying, wherein the system classification method is used in the classifying process, namely classifying the commodity according to the classifying rule of the online mall, and the specific process of classifying the type of the commodity in the consumption data of the user is as follows:
firstly, acquiring commodity pictures, acquiring commodity types of all commodities according to the division of the commodity types of an online mall, taking the commodity types and commodity names as manual labels of the corresponding commodity pictures, taking a commodity image with the manual labels as one sample, and forming a data set for training a CNN neural network by a large number of samples;
then, dividing the data set into a training set, a testing set and a verification set according to a preset proportion, inputting the training set, the testing set and the verification set into a CNN neural network, and training the CNN neural network by combining a cross entropy loss function to obtain a trained CNN neural network;
and finally, classifying all the commodities in the consumption data of the user by using the trained CNN neural network, and outputting commodity types corresponding to all the commodities to obtain a plurality of commodity types.
And (2) obtaining the main category of the commodity purchased by the user according to the obtained commodity category, wherein the main category is calculated according to the number and price of the purchased commodity, the commodity number represents the purchase preference of the user, and the commodity price represents the consumption level of the user, and the shopping preference of the user and the consumption level of the user are required to be considered when the commodity recommendation is carried out.
The main category of the structured commodity comprises the following specific acquisition methods:
firstly, according to the quantity of commodities contained in commodity types and commodity prices in consumption data of any user, dominant parameters of the commodity types are obtained, and the specific calculation method is as follows:
wherein G is i N representing the dominant parameter of the ith commodity type in the consumption data i Representing the number of goods contained in the ith goods category in the consumption data, T ij Represents the commodity price of the j-th commodity in the i-th commodity category in the consumption data, m represents the number of commodity categories in the consumption data, n i Indicating the number of items included in the i-th item type;
indicating the total number of goods contained in all the goods types in the consumption data; />Indicating that the ith commodity type in the consumption data comprises the ratio of the commodity quantity to the corresponding quantity of all commodities, wherein the larger the ratio is, the more the shopping preference of the user is biased to the corresponding commodity type; />The number of items of the i-th item type purchased by the user is represented, and the ratio of the number to the corresponding number of all items is calculated.
And obtaining dominant parameters corresponding to all commodity types, and marking the commodity type corresponding to the maximum value of the dominant parameters as a main type.
And (3) presetting a K value to be 5 according to experience, and clustering the commodities in the consumption data by using a K-means algorithm according to the quantity of the commodities contained in each commodity type in the consumption data and the distance between average prices of the contained commodities to obtain a plurality of clustering clusters.
It should be noted that, when recommending the commodity, not only one type of commodity is recommended, so after clustering is performed in this embodiment, each cluster contains a plurality of commodity types; in addition, in the clustering process, iterative updating is performed according to the changes of different kinds of commodities in each cluster.
It should be noted that the purchase quantity and price of the commodities in the same cluster are relatively similar, and the values of the dominant parameters of different clusters are different;
in order to update the recommended merchandise according to the merchandise currently purchased by the user, the embodiment describes the purchase intention of the user on the merchandise in the consumption data by constructing a tree structure, which comprises the following specific steps:
firstly, constructing a chain-shaped tree structure from top to bottom, wherein the tree structure has 5 nodes in total;
then, the cluster where the commodity category corresponding to the maximum value of the dominant parameter is located is used as the node at the uppermost layer in the tree structure, namely the 1 st node of the chain-shaped tree structure;
finally, calculating the average value of all commodity types corresponding to dominant parameters contained in the residual cluster, respectively serving as the 2 nd to 5 th nodes of the chain tree structure in sequence from big to small, and sequentially arranging the obtained chain tree structure serving as a commodity recommendation model;
it should be noted that the number of nodes in the chain tree structure is an empirical value, and can be adjusted according to actual conditions;
it should be noted that, because new consumption data will appear continuously during the consumption process of the e-commerce platform, the real-time update and analysis are performed subsequently according to the collected consumption data of the user, and the gradient rising method is adopted to update the commodity recommendation model during the update process.
So far, the chain tree structure corresponding to the consumption data is obtained.
Step S003, obtaining dominant parameters of the updated commodity types according to consumption data updated by the user in real time, and obtaining the ascending amplitude and ascending speed of the updated commodity types according to the dominant parameters of the updated commodity types combined with the chain tree structure.
It should be noted that, the consumer data of the user is updated in real time, so that the chain tree structure is utilized to update the consumer data of the user acquired in real time, and all commodity types contained in the online mall are utilized to cluster when the commodity types in the consumer data of the user are clustered, wherein the commodity types with zero shopping quantity are contained;
in addition, when more commodity types with zero shopping quantity appear in the consumption data, the position of the cluster where the commodity types are located in the chain tree structure moves upwards, so that the chain tree structure is updated iteratively according to the change of corresponding commodities in each commodity type.
The method comprises the steps of (1) marking the 1 st node to the 5 th node from top to bottom in a chain tree structure as a first layer to a fifth layer respectively, wherein each layer corresponds to one cluster, each cluster comprises a plurality of commodity types, each commodity type corresponds to one dominant parameter, and the commodity type corresponding to the largest dominant parameter in each layer is marked as the main type of the corresponding layer number;
firstly, in the consumption data acquired in real time, the quantity of commodities and commodity price contained in each commodity type are changed and updated, so that the commodity type with the quantity of the commodities and commodity price changed and updated in the consumption data acquired in real time is marked as an updated commodity type, and a dominant parameter of the updated commodity type is acquired by a dominant parameter acquisition method and is marked as G';
then, according to the dominant parameters, obtaining judgment parameters for randomly updating commodity types, wherein the calculation method comprises the following steps:
p＝G′ Y ―G′
wherein p represents the judgment parameter of the type of the updated commodity, G' Y The dominant parameters of the Y-th layer corresponding to the main type are represented, and G' represents the dominant parameters of the updated commodity type.
Finally, if p >0, classifying the updated commodity type into a Y layer, and if p <0, iterating upwards the layer number, and continuing to judge until Y=1;
in addition, when the consumption data of the updated commodity type appears for the first time, the layer number in the chain tree structure is recorded as an initial layer number and is recorded as Yc; after the change of the updated commodity type is updated, the number of layers corresponding to the judgment parameter is obtained, and the number of layers corresponding to the chain tree structure is recorded as the number of updated layers and is recorded as Yg.
When the judgment parameters of the updated commodity types are obtained, calculation is performed from the lowest layer, and calculation is performed upwards until the threshold condition is satisfied.
The chain tree structure is updated according to the consumption data updated by the user in real time, and the commodity types corresponding to the consumption data are obtained when the user has new shopping demands in a period of time, but initially, the number of commodities corresponding to the commodity types is smaller under the corresponding shopping demands, so that the dominant parameters can be divided into lower tree layers, and the dominant parameters corresponding to the commodity types can be changed to a larger extent only when the user generates multiple purchasing records corresponding to the commodity types in a short time, and then the rising amplitude and the rising speed of the dominant parameters of the updated commodity types in the chain tree structure are obtained according to the consumption data of the user.
Describing shopping preference of a user according to consumption data acquired in real time, wherein each layer in the chain tree structure is a cluster corresponding to a large number of commodities with similar average prices in the consumption data of the user and comprises a plurality of commodity types, so that differences between the commodity types in each layer and corresponding main types are acquired, and the difference characteristics between the main types and the commodity types in any layer in the chain tree structure are obtained, and the specific calculation method comprises the following steps:
wherein D (i, Y) represents a difference characteristic between the corresponding main type and the ith commodity type in the Y-th layer of the chain tree structure; n is n i Representing the number of items included in the i-th item type; t (T) ij Representing the commodity price of the j-th commodity in the i-th commodity category; n is n Y Representing the number of commodities contained in the main type corresponding to the Y layer; t (T) Yj Representing the commodity price of the j commodity in the main type corresponding to the Y layer;
in addition, by utilizing an acquisition method of the difference characteristics between the main types and the commodity types in any layer in the chain tree structure, the difference characteristics between the corresponding main types of any updated commodity type in the updated layer number are acquired, and are marked as D', and the difference characteristics between the updated commodity type and the corresponding main types of the updated layer number are expressed;
step (3), according to the difference characteristics between each commodity type and the corresponding main type in each layer in the chain tree structure, the ascending amplitude and the ascending speed of the commodity type in the chain tree structure are obtained, and the obtaining method comprises the following steps:
first, the ascending amplitude of any updated commodity category is acquired:
f＝|Yg―Yc|×G′
wherein f represents the rising amplitude of the updated commodity type in the consumption data; yg represents the number of layers after updating in the tree structure of the updated commodity type; yc represents the initial number of layers in the tree structure for updating the commodity category; g' represents a dominant parameter for updating the commodity type; the absolute value is acquired.
The number of layers change feature |Y' — Y| indicates that the number of layers where the updated commodity type is in the chain tree structure changes, and the larger the number of layers is, the more the number of layers is increased; the commodity type representation rising in the chain tree structure is biased to the consumption preference of the user, so that the change of the position of the commodity type in the chain tree structure reflects the purchase intention of the user;
in addition, the updated dominant parameters of the commodity types indicate that the number of the included commodities and the commodity price are changed, and the updated dominant parameters corresponding to the commodity types, namely the updated dominant parameters of the commodity types, have larger values, which indicate that the purchasing preference of the user for the commodity types is changed greatly, so that the rising amplitude is increased.
Then, when consumption data of the updated commodity type appears for the first time, the difference characteristic between the updated commodity type and the main type corresponding to the initial layer number is marked as D', and the difference characteristic between the updated commodity type and the main type corresponding to the initial layer number is represented; recording dominant parameters of the main type corresponding to the number of updated layers where the updated commodity types are located as Gg; recording dominant parameters of the main type corresponding to the initial layer number where the updated commodity type is located as Gc; meanwhile, when the consumption data change is updated, the corresponding consumption time of the updated commodity type is recorded as updated time t'; when corresponding consumption data appears for the first time in the updated commodity type, the corresponding consumption time is recorded as initial time t;
the rising speed of any updated commodity category is obtained:
wherein v represents the rising speed of the updated commodity type in the consumption data, and D' represents the difference characteristic between the updated commodity type and the main type corresponding to the updated layer number; d' represents a difference feature between the updated commodity type and the primary type corresponding to the initial layer number; gg represents the dominant parameter of the main type corresponding to the number of layers after updating where the updated commodity type is located, gc represents the dominant parameter of the main type corresponding to the number of initial layers where the updated commodity type is located, t' represents the time after updating, t represents the initial time, and I represents the absolute value obtained.
Dividing the change of the layer number in the chain tree structure according to the updated commodity type by the time, namely representing the rising speed in the chain tree structure; the larger the rising speed is, the larger the user's recent consumption will for the corresponding commodity category is, and the corresponding commodity category should be used as an important reference factor when recommending commodities.
And S004, obtaining the recommendation importance degree corresponding to the updated commodity type according to the rising amplitude, the rising speed and the dominant parameters of the updated commodity type, and further realizing intelligent marketing recommendation of advertisements.
Firstly, according to the rising amplitude, rising speed and dominant parameters of the updated commodity types, the recommended importance degree of the corresponding updated commodity types is obtained, and the specific acquisition method is as follows:
ω＝G′×f×v
wherein ω represents a recommended importance degree of the updated commodity category; g' represents a dominant parameter for updating the commodity type; f represents the rising amplitude of the updated commodity type in the consumption data; v represents the rising speed of the updated commodity type in the consumption data;
then, recommending the commodity according to the recommending importance degree of the updated commodity type, and because the updated commodity type with faster rising speed and rising amplitude shows that the commodity belonging to the commodity type corresponding to the updated commodity type is recently needed by a user and has large demand degree and large purchasing quantity, the commodity of the type needs to be used as a main reference factor when recommending the commodity;
therefore, the recommendation importance degrees of all the updated commodity types in the consumption data of the user are obtained, the recommendation importance degrees of all the updated commodity types are normalized by using a linear normalization method, the normalization result is used as the recommendation probability of the same commodity type when the user browses on an electronic commerce platform, the larger the recommendation probability is, the more commodity numbers of the corresponding commodity types are on a browsing page, and intelligent advertisement marketing recommendation is realized.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (8)
1. An advertising marketing recommendation method based on data analysis, which is characterized by comprising the following steps:
updating in real time to obtain consumption data of a user, including: commodity name, consumption time, and commodity price;
acquiring commodity types of all commodities in the consumption data by using a neural network; obtaining dominant parameters of commodity types according to the quantity of commodities contained in each commodity type and commodity price; clustering all commodity types according to the quantity of the commodity contained in each commodity type and the distance between the average prices of the contained commodity to obtain a plurality of cluster clusters, constructing a tree structure, taking each cluster as each node of the tree structure according to the size and average value of the dominant parameters of the commodity types in the cluster clusters in sequence to obtain a chain-shaped tree structure, marking the node of the chain-shaped tree structure as a layer, and marking the commodity type corresponding to the maximum dominant parameter in any layer as a main type;
when the quantity of the commodities contained in the commodity category is changed and updated, the corresponding commodity category is marked as the updated commodity category, and corresponding dominant parameters are obtained; obtaining judgment parameters according to the difference between the type of the updated commodity and the dominant parameters of any main type, and obtaining the updated layer number of the type of the updated commodity according to the size of the judgment parameters; obtaining difference characteristics according to the commodity price and the quantity of the commodities contained in the main type; when the consumption data of the updated commodity type appears for the first time, the layer number in the chain tree structure is marked as an initial layer number, the result of adjusting the difference between the updated layer number and the initial layer number by the dominant parameter of the updated commodity type is marked as the rising amplitude of the updated commodity type, and the fusion result of the consumption time, the dominant parameter of the main type and the difference characteristic is marked as the rising speed of the updated commodity type;
and obtaining the recommendation importance degree of the updated commodity type according to the dominant parameters, the ascending amplitude and the ascending speed, and realizing intelligent advertisement marketing recommendation according to the recommendation importance degree.
2. The advertising marketing recommendation method based on data analysis according to claim 1, wherein the obtaining the commodity category to which each commodity belongs in the consumption data by using the neural network comprises the following specific steps:
firstly, acquiring commodity pictures, acquiring commodity types of all commodities according to the division of the commodity types of an online mall, taking the commodity types and commodity names as manual labels of the corresponding commodity pictures, taking a commodity image with the manual labels as one sample, and forming a data set for training a CNN neural network by a large number of samples;
then, dividing the data set into a training set, a testing set and a verification set according to a preset proportion, inputting the training set, the testing set and the verification set into a CNN neural network, and training the CNN neural network by combining a cross entropy loss function to obtain a trained CNN neural network;
and finally, classifying all the commodities in the consumption data of the user by using the trained CNN neural network, and outputting commodity types corresponding to all the commodities to obtain a plurality of commodity types.
3. The advertising marketing recommendation method based on data analysis according to claim 1, wherein the obtaining the dominant parameters of the commodity category according to the number of commodities included in each commodity category and the commodity price comprises the following specific steps:
wherein G is i Representing the ith commodity species in the consumption dataDominant parameters of class, N i Representing the number of goods contained in the ith goods category in the consumption data, T ij Represents the commodity price of the j-th commodity in the i-th commodity category in the consumption data, m represents the number of commodity categories in the consumption data, n i Indicating the number of items included in the ith item category.
4. The data analysis-based advertising marketing recommendation method of claim 1, wherein the chain tree knots
The acquisition method is as follows:
step (1), presetting a K value to be 5 according to experience, and clustering commodities in consumption data by using a K-means algorithm according to the quantity of commodities contained in each commodity type in the consumption data and the distance between average prices of the contained commodities to obtain a plurality of cluster clusters;
step (2), firstly, constructing a chain-shaped tree structure from top to bottom, wherein the tree structure has 5 nodes;
then, the cluster where the commodity category corresponding to the maximum value of the dominant parameter is located is used as the node at the uppermost layer in the tree structure, namely the 1 st node of the chain-shaped tree structure;
and finally, calculating the average value of all commodity types corresponding to the dominant parameters contained in the residual cluster, respectively serving as the 2 nd to 5 th nodes of the chain tree structure in the sequence from big to small, and sequentially arranging the obtained chain tree structure serving as a commodity recommendation model.
5. The data analysis-based advertising marketing recommendation method of claim 1, wherein the data is updated based on the update
Obtaining judgment parameters according to the difference between commodity types and dominant parameters of any main types and obtaining the judgment parameters according to the sizes of the judgment parameters
The number of updated layers for updating the commodity category comprises the following specific steps:
firstly, recording a difference value between a dominant parameter of the updated commodity type and a dominant parameter of a corresponding main type of an arbitrary layer as a judging parameter of the updated commodity type;
then, starting from the 5 th layer of the chain tree structure, sequentially reducing and acquiring corresponding judgment parameters, and if the judgment parameters are larger than 0, classifying the updated commodity types into the layer number; if the judgment parameter is smaller than 0, the layer number is iterated upwards, and the judgment is continued until the layer number is 1, so that the updated layer number of the updated commodity type is obtained.
6. The advertising marketing recommendation method based on data analysis of claim 1, wherein the rise amplitude is obtained by:
the absolute value of the difference between the number of layers after updating and the number of initial layers in the tree structure of the updated commodity type is recorded as the layer change characteristic;
and (3) marking the product result of the dominant parameter of the updated commodity type and the layer number change characteristic as the rising amplitude of the updated commodity type.
7. The advertising marketing recommendation method based on data analysis of claim 1, wherein the rising speed is obtained by the following method:
firstly, obtaining the difference characteristic between main types corresponding to the initial layer number when the consumption data of the updated commodity types appears for the first time, and marking the difference characteristic as D'; recording dominant parameters of the main type corresponding to the number of updated layers where the updated commodity types are located as Gg; recording dominant parameters of the main type corresponding to the initial layer number where the updated commodity type is located as Gc; when the consumption data change is updated, the corresponding consumption time of the updated commodity type is recorded as updated time; when corresponding consumption data appears for the first time, the corresponding consumption time of the updated commodity type is recorded as initial time;
then, the rising speed of any updated commodity kind is obtained:
wherein v represents the rise of the updated commodity type in the consumption dataSpeed, D ′ Representing the difference characteristics between the updated commodity types and the main types corresponding to the updated layer numbers; d' represents a difference feature between the updated commodity type and the primary type corresponding to the initial layer number; gg represents the dominant parameter of the main type corresponding to the number of layers after updating the type of the updated commodity, gc represents the dominant parameter of the main type corresponding to the number of initial layers of the type of the updated commodity, t ′ Representing updated time, t representing initial time; the absolute value is acquired.
8. The advertising marketing recommendation method based on data analysis according to claim 1, wherein the obtaining the recommendation importance degree of the updated commodity category according to the dominant parameter, the rising amplitude and the rising speed, and the implementing the advertising intelligent marketing recommendation according to the recommendation importance degree, comprises the following specific steps:
firstly, recording the ascending amplitude, ascending speed and accumulated result of the updated commodity type as the recommended importance degree of the updated commodity type, and obtaining the recommended importance degree of all the updated commodity types;
and then, carrying out normalization processing on the recommendation importance degrees of all the updated commodity types by using a linear normalization method, taking the normalization result as the recommendation probability of the same commodity type when a user browses on an e-commerce platform, and carrying out marketing recommendation on commodity advertisements corresponding to the commodity types in the platform according to the recommendation probability, wherein the larger the recommendation probability is, the more commodity numbers corresponding to the commodity types are on a browsing page, so as to realize intelligent advertisement marketing recommendation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310757497.0A CN116911926A (en) | 2023-06-26 | 2023-06-26 | Advertisement marketing recommendation method based on data analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310757497.0A CN116911926A (en) | 2023-06-26 | 2023-06-26 | Advertisement marketing recommendation method based on data analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116911926A true CN116911926A (en) | 2023-10-20 |
Family
ID=88365915
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310757497.0A Pending CN116911926A (en) | 2023-06-26 | 2023-06-26 | Advertisement marketing recommendation method based on data analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116911926A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117436957A (en) * | 2023-12-20 | 2024-01-23 | 锦诚实业科技(深圳)有限公司 | Game software advertisement accurate delivery method |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103984749A (en) * | 2014-05-27 | 2014-08-13 | 电子科技大学 | Focused crawler method based on link analysis |
| CN109918563A (en) * | 2019-01-24 | 2019-06-21 | 暨南大学 | A method of the book recommendation based on public data |
| CN112070577A (en) * | 2020-08-31 | 2020-12-11 | 深圳市卡牛科技有限公司 | Commodity recommendation method, system, equipment and medium |
| US20210256403A1 (en) * | 2018-11-09 | 2021-08-19 | Huawei Technologies Co., Ltd. | Recommendation method and apparatus |
| CN113724042A (en) * | 2021-08-23 | 2021-11-30 | 中国建设银行股份有限公司 | Commodity recommendation method, commodity recommendation device, commodity recommendation medium and commodity recommendation equipment |
| CN113763027A (en) * | 2021-05-10 | 2021-12-07 | 腾讯科技（深圳）有限公司 | Recommendation information processing method, recommendation information generating method and device |
| CN114462894A (en) * | 2022-04-11 | 2022-05-10 | 国能(北京)商务网络有限公司 | Data analysis-based e-commerce order material replacement assistant decision method |
| CN116205675A (en) * | 2023-04-28 | 2023-06-02 | 华南师范大学 | Data acquisition method and device based on thread division |
-
2023
- 2023-06-26 CN CN202310757497.0A patent/CN116911926A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103984749A (en) * | 2014-05-27 | 2014-08-13 | 电子科技大学 | Focused crawler method based on link analysis |
| US20210256403A1 (en) * | 2018-11-09 | 2021-08-19 | Huawei Technologies Co., Ltd. | Recommendation method and apparatus |
| CN109918563A (en) * | 2019-01-24 | 2019-06-21 | 暨南大学 | A method of the book recommendation based on public data |
| CN112070577A (en) * | 2020-08-31 | 2020-12-11 | 深圳市卡牛科技有限公司 | Commodity recommendation method, system, equipment and medium |
| CN113763027A (en) * | 2021-05-10 | 2021-12-07 | 腾讯科技（深圳）有限公司 | Recommendation information processing method, recommendation information generating method and device |
| CN113724042A (en) * | 2021-08-23 | 2021-11-30 | 中国建设银行股份有限公司 | Commodity recommendation method, commodity recommendation device, commodity recommendation medium and commodity recommendation equipment |
| CN114462894A (en) * | 2022-04-11 | 2022-05-10 | 国能(北京)商务网络有限公司 | Data analysis-based e-commerce order material replacement assistant decision method |
| CN116205675A (en) * | 2023-04-28 | 2023-06-02 | 华南师范大学 | Data acquisition method and device based on thread division |
Non-Patent Citations (2)
| Title |
|---|
| 凌坤: "基于网络表示的个性化推荐算法研究与应用", 中国优秀硕士论文电子期刊, 15 November 2022 (2022-11-15) * |
| 李瑞远;洪亮;曾承;: "多层聚簇中基于协同过滤的跨类推荐算法", 小型微型计算机系统, no. 04, 15 April 2017 (2017-04-15) * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117436957A (en) * | 2023-12-20 | 2024-01-23 | 锦诚实业科技(深圳)有限公司 | Game software advertisement accurate delivery method |
| CN117436957B (en) * | 2023-12-20 | 2024-03-29 | 深圳市艾森互动科技有限公司 | Game software advertisement accurate delivery method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Al-Halah et al. | Fashion forward: Forecasting visual style in fashion | |
| CN108629665B (en) | Personalized commodity recommendation method and system | |
| CN110909176B (en) | Data recommendation method and device, computer equipment and storage medium | |
| CN109903117B (en) | Knowledge graph processing method and device for commodity recommendation | |
| CN103714130B (en) | Video recommendation system and method | |
| CN109359244A (en) | A kind of recommendation method for personalized information and device | |
| CN108520450B (en) | Recommendation method and system for local low-rank matrix approximation based on implicit feedback information | |
| CN105718184A (en) | Data processing method and apparatus | |
| EP4198864A1 (en) | User feature-based page display method and apparatus, medium, and electronic device | |
| CN101482884A (en) | Cooperation recommending system based on user predilection grade distribution | |
| CN104679771A (en) | Individual data searching method and device | |
| CN111242729A (en) | Serialization recommendation method based on long-term and short-term interests | |
| CN112487199B (en) | User characteristic prediction method based on user purchasing behavior | |
| CN111737418B (en) | Method, apparatus and storage medium for predicting relevance of search term and commodity | |
| CN111724235A (en) | Online commodity recommendation method based on user novelty | |
| CN111291191B (en) | Broadcast television knowledge graph construction method and device | |
| CN111949887A (en) | Item recommendation method and device and computer-readable storage medium | |
| CN102103641A (en) | Method for adding banner advertisement into user-browsed network image | |
| CN116911926A (en) | Advertisement marketing recommendation method based on data analysis | |
| CN113946754A (en) | User portrait based rights and interests recommendation method, device, equipment and storage medium | |
| CN110334278A (en) | A kind of web services recommended method based on improvement deep learning | |
| CN114971747A (en) | Data analysis method and system based on big data commodity accurate marketing | |
| CN111310035A (en) | Recommendation method and device based on psychological and behavioral characteristics | |
| CN116089745A (en) | Information recommendation method, device, electronic equipment and computer readable storage medium | |
| CN113327132A (en) | Multimedia recommendation method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |