CN112070577A - Commodity recommendation method, system, equipment and medium - Google Patents
Commodity recommendation method, system, equipment and medium Download PDFInfo
- Publication number
- CN112070577A CN112070577A CN202010898544.XA CN202010898544A CN112070577A CN 112070577 A CN112070577 A CN 112070577A CN 202010898544 A CN202010898544 A CN 202010898544A CN 112070577 A CN112070577 A CN 112070577A
- Authority
- CN
- China
- Prior art keywords
- user
- clustering
- users
- preference
- cluster
- 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 45
- 239000013598 vector Substances 0.000 claims abstract description 116
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 238000012545 processing Methods 0.000 claims description 34
- 238000004590 computer program Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 abstract description 15
- 230000000694 effects Effects 0.000 abstract description 9
- 238000010586 diagram Methods 0.000 description 8
- 235000013194 Lyophyllum decastes Nutrition 0.000 description 7
- 240000005856 Lyophyllum decastes Species 0.000 description 7
- 235000015220 hamburgers Nutrition 0.000 description 7
- 235000012020 french fries Nutrition 0.000 description 5
- 235000015243 ice cream Nutrition 0.000 description 5
- 238000003064 k means clustering Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000008450 motivation Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—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
Abstract
The embodiment of the invention discloses a commodity recommendation method, a commodity recommendation system, commodity recommendation equipment and a commodity recommendation medium. The method comprises the following steps: acquiring historical data of a user, and preprocessing the historical data to obtain user characteristic data; coding and splicing the user characteristic data to obtain a first user characteristic vector; converting the first user characteristic vector into a second user characteristic vector with a preset length after passing through two full-connection layers; performing incremental clustering on the second user characteristic vector to divide users corresponding to the second user characteristic vector into different clusters; acquiring preference sets of all users of each cluster; and recommending the preference set to all users in the corresponding cluster. The user characteristics are analyzed by acquiring the historical data of the user, the user characteristics are processed into the characteristic vectors with preset lengths to carry out incremental clustering, the preference sets of similar users are obtained and recommended to related users, the problem of low conversion rate of recommended commodities is solved, and the effects of improving the accuracy and conversion rate of the recommended commodities are achieved.
Description
Technical Field
The embodiment of the invention relates to a computer application technology, in particular to a commodity recommendation method, a commodity recommendation system, commodity recommendation equipment and a commodity recommendation medium.
Background
With the development of internet technology, the internet has become an indispensable part in people's life, and consumption, entertainment, learning, going on a journey, financing and the like in people's life can not leave the internet.
Meanwhile, the types of goods of e-commerce are more and more abundant, the promotion activities are more and more endless, the amount of generated data information is huge, and when people need to acquire certain types of data, the wanted products or services can be obtained through active searching or recommendation of related application programs.
The user hopes to obtain commodity recommendation meeting the preference of the user from a large amount of information, and if the recommendation of the application program cannot be well combined with the potential needs of the user, people need to spend a large amount of time on searching for information meeting the requirements of the user facing excessive information, the conversion rate of recommended commodities is reduced, and the viscosity of the user is reduced; inaccurate recommended commodities often cannot obtain potential preference commodities which are not purchased by the user due to inaccurate user characteristic analysis and commodity recommendation only based on commodities which are purchased by the user, and cannot perform accurate commodity recommendation on the new user due to the fact that the new user does not purchase commodities.
Disclosure of Invention
The invention provides a commodity recommendation method, a commodity recommendation system, commodity recommendation equipment and a commodity recommendation medium, which are used for realizing commodity recommendation according to user preference and improving the accuracy and conversion rate of recommended commodities.
In a first aspect, an embodiment of the present invention provides a commodity recommendation method, including:
acquiring historical data of a user, and preprocessing the historical data to obtain user characteristic data;
encoding the user characteristic data;
performing feature splicing processing on the coded data to obtain a first user feature vector;
converting the first user characteristic vector into a second user characteristic vector with a preset length after passing through two full-connection layers;
performing incremental clustering on the second user characteristic vector to divide users corresponding to the second user characteristic vector into different clusters;
acquiring preference sets of all users of each cluster;
and recommending the preference set to all users in the corresponding cluster.
Optionally, the performing incremental clustering on the second user feature vector, and dividing the users corresponding to the second user feature vector into different clusters includes:
pre-clustering the second user characteristic vector by adopting a clustering algorithm;
saving the clustering centers and class labels of the pre-clusters;
acquiring a newly added feature vector of a user, and calculating the distance from the feature vector to each clustering center;
dividing the user corresponding to the newly added feature vector to the nearest clustering center;
the cluster center for each class is updated.
Optionally, the obtaining the preference set of the user of each cluster includes:
acquiring commodity information generated by each user of one cluster to form a preference pre-set;
judging whether the quantity of commodities in the preference pre-set is larger than a preset numerical value or not;
if the number of the commodities is larger than the preset number, sorting the commodities in the preference pre-set, and sequentially selecting the commodities with the preset number to form a preference set;
and if the number of the commodities is not more than the preset number, determining the preference preset set as the preference set.
Optionally, the user feature data includes a dense user feature and a sparse user feature, and when the user feature data is encoded, the method includes:
the user dense features are processed by adopting One-Hot coding; the user sparse feature is coded by using Embedding.
Optionally, the user's historical data includes, but is not limited to, device information, geographic location information, and demographic information, as well as browsing, clicking, collecting, and purchasing information for the goods.
In a second aspect, an embodiment of the present invention further provides a product recommendation system, including:
the user data acquisition module is used for acquiring historical data of a user and obtaining user characteristic data through preprocessing;
the coding processing module is used for carrying out coding processing and feature splicing processing on the user feature data to obtain a first user feature vector;
the full-connection layer processing module is used for converting the first user characteristic vector into a second user characteristic vector with a preset length after passing through two full-connection layers;
the clustering processing module is used for carrying out incremental clustering on the second user characteristic vector and dividing users corresponding to the second user characteristic vector into different clusters;
the preference acquisition module is used for acquiring a preference set of users of each cluster;
and the commodity recommending module is used for recommending the preference set to all users in the corresponding cluster.
Optionally, the clustering processing module includes:
the pre-clustering sub-module is used for pre-clustering the second user characteristic vector by adopting a clustering algorithm;
the storage submodule is used for storing the clustering center and the class label of the pre-clustering;
the computing submodule is used for acquiring a newly added feature vector of a user and computing the distance from each clustering center;
the clustering submodule is used for dividing the user corresponding to the newly added feature vector to a clustering center with the nearest distance;
and the updating submodule is used for updating the clustering center of each class.
In a third aspect, an embodiment of the present invention further provides a product recommendation device, where the product recommendation device includes:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the item recommendation method of any one of the first aspects.
In a fourth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the item recommendation method according to any one of the first aspect.
According to the embodiment of the invention, the user characteristics are analyzed by acquiring the historical data of the user, the user characteristics are processed into the characteristic vectors with preset lengths to carry out incremental clustering, the preference set of similar users is obtained and recommended to the related users, the problems that the recommended commodities are not matched with the expected consumption preferences of the users or the recommended commodities are single, so that the conversion rate of the recommended commodities is low and the commodities cannot be accurately recommended to new users are solved, and the effects of improving the accuracy and conversion rate of the recommended commodities and accurately recommending the new users who do not purchase the commodities are realized.
Drawings
Fig. 1 is a schematic flowchart of a commodity recommendation method according to an embodiment of the present invention;
fig. 2 is a schematic sub-flow chart of a commodity recommendation method according to a second embodiment of the present invention;
fig. 3 is a schematic sub-flow chart of another commodity recommendation method according to a second embodiment of the present invention;
fig. 4A is a schematic structural diagram of a commodity recommendation system according to a third embodiment of the present invention;
fig. 4B is a schematic structural diagram of a clustering module of a commodity recommendation system according to a third embodiment of the present invention;
fig. 5 is a schematic structural diagram of a commodity recommendation device according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a product recommendation method according to an embodiment of the present invention, where this embodiment is applicable to a case where a product preference consumed by a user is obtained through user history data and a product recommendation is performed, and the method may be executed by a product recommendation system, and specifically includes the following steps:
and step 110, acquiring historical data of the user, and preprocessing the historical data to obtain user characteristic data.
And acquiring historical data of the user through the data buried point, wherein the historical data of the user comprises but is not limited to equipment information, geographical location information and demographic information, and browsing, clicking, collecting and purchasing information of commodities. The device information comprises information of mobile terminals such as mobile phones, tablets and the like which generate shopping behaviors by users, such as models and the like; the in-ground position is obtained through a mobile terminal used by a user, and comprises work and living places of the user; demographic information includes user gender, age, occupation, academic calendar, etc.; the commodity information includes commodity names, commodity classifications, commodity prices, and the like; the user characteristic data can be obtained by preprocessing the information.
And 120, coding and splicing the user characteristic data to obtain a first user characteristic vector.
And after the user characteristic data is obtained, encoding processing is carried out to obtain the encoding characters which can be processed by the system. The user feature data comprise user dense features and user sparse features; the dense features of the user comprise traditional features such as identity information of the user and the like, and the dimensionality is not particularly high; the user sparse features comprise commodity information and the like generated by a user, for example, the user clicks a browsed commodity list, and the features are high in dimensionality and large in quantity.
The encoding method of the user characteristic data includes but is not limited to the following three modes:
firstly, after word embedding processing is carried out on the user sparse features, each user sparse feature is mapped into a vector, cross features are obtained through cross operation among different user sparse features, the cross features and the user dense features are spliced to obtain a first user feature vector, and the first user feature vector is used as input of a full connection layer.
And secondly, firstly, discretizing the user dense features into discrete features, then respectively carrying out word embedding processing on the discrete user dense features and the discrete user sparse features, and then carrying out feature crossing and splicing on the user dense features and the user sparse features after the word embedding processing to obtain a first user feature vector which is used as the input of the full connection layer.
And thirdly, performing dot multiplication on each user dense feature and the word embedding matrix, performing word embedding processing, multiplying the word embedding matrix by a weight vector to obtain a feature vector, namely mapping a dense feature into a feature vector, performing word embedding processing on the user sparse feature, and splicing the user sparse feature with the feature vector to obtain a first user feature vector which is used as the input of the full connection layer.
Optionally, when encoding the user feature data, the method includes: the dense features of the users are processed by adopting One-Hot coding; the sparse characteristics of the user are coded by Embedding; then a feature splicing operation.
One-Hot encoding, also known as One-bit-efficient encoding, mainly uses an N-bit state register to encode N states, each state being represented by an independent register bit and having only One bit available at any time. One-Hot encoding is the representation of classification variables as binary vectors. This first requires mapping the classification values to integer values. Each integer value is then represented as a binary vector, which is a zero value, except for the index of the integer, which is marked as 1. The discrete features are coded by One-Hot, so that the distance between the features can be calculated more reasonably.
The essence of Embedding coding is mapping from semantic space to vector space, and at the same time, the relationship of the original sample in the semantic space is kept in the vector space as much as possible, for example, the positions of two words with similar semantics in the vector space are also relatively close. And the Embedding coding is adopted to reduce the dimension of a large amount of user sparse features to a low-dimensional space.
And step 130, converting the first user characteristic vector into a second user characteristic vector with a preset length after passing through two full connection layers.
The first user characteristic vector obtained by splicing is provided for a deep network model for learning, and is converted into a second user characteristic vector with a preset length after passing through two full-connection layers, wherein the number of layers of the network can be set according to the effect without limitation, and experiments show that generally two full-connection layers can obtain good effect.
And 140, performing incremental clustering on the second user characteristic vector to divide the users corresponding to the second user characteristic vector into different clusters.
The process of dividing a collection of physical or abstract objects into classes composed of similar objects is called clustering, and a cluster generated by clustering is a collection of data objects that are similar to objects in the same cluster and different from objects in other clusters. Incremental clustering is primarily the view of incremental data as time series data or data in a particular order.
And performing incremental clustering on the second user characteristic vector, specifically clustering users corresponding to the second user characteristic vector, and dividing the users into different clusters, wherein when a new characteristic vector is added, the clustering center is kept unchanged, the cluster to which the user belongs can be changed, namely the relation between the user and the clustering center can be changed.
The new user does not purchase any commodity, and cannot classify the commodity according to the commodity purchasing preference obtained by the commodity browsing, clicking, collecting and purchasing information, but can obtain clusters to which other users similar to the conditions of occupation, income, life or work place and the like of the new user belong according to the equipment information, the geographic position information and the demographic information, and the new user is classified into corresponding clusters.
Acquiring preference sets of all users in each cluster, and collecting information of all commodities clicked or purchased by all users to form a preference set; illustratively, there are three users A, B, C in a cluster, and the items purchased by A are hamburger A1 and French fries A2; b, the purchased commodities are ice cream B1 and fried chicken B2, and the clicked and browsed commodity is a coupon B3; c purchased goods, hamburger a1, fried chicken chop B2, and full barrel C1. Then the set of commercial preferences in this category are hamburger a1, french fries a2, ice cream B1, fried chicken B2, coupon B3, and full barrel C1.
The users in the cluster have the same or similar preferences and have an incentive to purchase commodities purchased by other users in the cluster, so that the preference set 'hamburger A1, French fries A2, ice cream B1, fried chicken B2, coupons B3 and whole barrel C1' acquired by the method is recommended to all users A, B, C in the cluster corresponding to the preference set, and the conversion rate of the recommended commodities can be improved.
According to the embodiment of the invention, the user characteristics are analyzed by acquiring the historical data of the user, the user characteristics are processed into the characteristic vectors with preset lengths to carry out incremental clustering, the preference set of similar users is obtained and recommended to the related users, the problems that the recommended commodities are not matched with the expected consumption preferences of the users or the recommended commodities are single, so that the conversion rate of the recommended commodities is low and the commodities cannot be accurately recommended to new users are solved, and the effects of improving the accuracy and conversion rate of the recommended commodities and accurately recommending the new users who do not purchase the commodities are realized.
Example two
Fig. 2 is a sub-flow diagram of a product recommendation method according to a second embodiment of the present invention, and fig. 3 is a sub-flow diagram of another product recommendation method according to the second embodiment of the present invention.
On the basis of the above embodiment, the incremental clustering is performed on the second user feature vector, and the users corresponding to the second user feature vector are classified into different clusters, which specifically includes the following steps:
and step 141, pre-clustering the second user feature vector by adopting a clustering algorithm.
And (3) processing the second user characteristic vector by adopting a clustering algorithm, such as a K-Means clustering algorithm, specifically, the K-Means clustering algorithm is an iterative solution clustering analysis algorithm, dividing data into K groups in advance, randomly selecting K user characteristics as initial clustering centers, calculating the distance between each user characteristic and each seed clustering center, and allocating the user corresponding to each user characteristic to the nearest clustering center. The cluster centers and the objects assigned to them represent a cluster. The cluster center of the cluster is recalculated for each sample assigned based on the existing user characteristics in the cluster. This process will be repeated until some termination condition is met. The termination condition may be that no (or a minimum number) of user features are reassigned to different clusters, no (or a minimum number) cluster centers are changed again, and the sum of squared errors is locally minimal.
And step 142, storing the clustering centers and the class labels of the pre-clusters.
And storing the clustering center and the class label obtained by pre-clustering, keeping the clustering center unchanged, keeping the class label as the same mark of the same cluster, and keeping the meaning of the class label unchanged for a long time, so that the system can learn the characteristic better.
And 143, acquiring the newly added feature vectors of the users, and calculating the distance from each cluster center.
At present, incremental clustering is mainly studied to regard incremental data as time series data or data in a specific sequence, and the technical scheme of the embodiment obtains a new feature vector of a user each time by using a last clustering result, and calculates the distance between the new feature vector of the user and each clustering center.
And 144, dividing the user corresponding to the newly added feature vector to the nearest clustering center.
According to the calculated clustering of the newly added feature vector of the user and each clustering center, the newly added feature vector of one user is divided into the existing clusters every time, namely the newly added feature vector of the user is divided into the cluster with the center nearest to the newly added feature vector of the user and the center is moved to the newly added feature vector, namely the newly added feature vector of the user does not influence the original division, and the method has the advantages that all data do not need to be re-clustered every time.
And step 145, updating the clustering center of each class.
And during the service low peak period, re-clustering the current user feature vector data, and updating the clustering center of each class in a full amount so as to eliminate the local deviation possibly introduced by incremental clustering and improve the accuracy of the system.
On the basis of the above embodiment, acquiring the preference set of the user of each cluster specifically includes the following steps:
and 151, acquiring commodity information generated by each user of one cluster to form a preference pre-set.
And acquiring commodity information generated by each user in a certain cluster, wherein the commodity information comprises browsing, clicking, collecting and purchasing information of commodities. The method comprises the steps of obtaining purchasing information, browsing, clicking, collecting and other information, analyzing purchasing requirements and expected consumption of users in a cluster, and forming a commodity preference pre-set of the users in the cluster.
And 152, judging whether the quantity of the commodities in the preference pre-set is larger than a preset quantity.
When more users exist in the cluster or the commodity information of the users exists, the obtained commodity preference pre-set is too much, and the commodity recommendation conversion rate is reduced due to the fact that all users are recommended; judging whether the number of the commodities in the preference pre-set is greater than a preset value, if so, executing step 153; if the quantity of the goods is not greater than the predetermined quantity, step 154 is executed.
And 153, sorting the commodities in the preference pre-set, and sequentially selecting a preset number of commodities to form a preference set.
When the number of the commodities in the commodity preference pre-set is larger than the preset number, the commodities in the preference pre-set are sorted in a descending order according to the clicking and purchasing times, the commodities with the preset number are sequentially selected to form the preference set, and the commodity recommendation conversion rate of the obtained preference set is high.
And step 154, determining the preference pre-set as a preference set.
When the number of the commodities in the commodity preference pre-set is not more than the preset number, the obtained preference pre-set is determined as a preference set, the number of the commodities in the preference set is small at the moment, the user preference in the cluster is concentrated, and therefore the commodity recommendation conversion rate is high according to the preference set.
EXAMPLE III
Fig. 4A is a schematic structural diagram of a product recommendation system provided in a third embodiment of the present invention, and fig. 4B is a schematic structural diagram of a clustering module of the product recommendation system provided in the third embodiment of the present invention.
The commodity recommendation system provided by the embodiment of the invention can execute the commodity recommendation method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
As shown in fig. 4A, a commodity recommendation system includes:
a user data obtaining module 410, configured to obtain historical data of a user, and obtain user characteristic data through preprocessing;
historical data of a user is obtained, wherein the historical data of the user comprises but is not limited to equipment information, geographical location information and demographic information, and browsing, clicking, collecting and purchasing information of commodities. The device information comprises information of mobile terminals such as mobile phones, tablets and the like which generate shopping behaviors by users, such as models and the like; the in-ground position is obtained through a mobile terminal used by a user, and comprises work and living places of the user; demographic information includes user gender, age, occupation, academic calendar, etc.; the commodity information includes commodity names, commodity classifications, commodity prices, and the like; the user characteristic data can be obtained by preprocessing the information.
And the encoding processing module 420 is configured to perform encoding processing and feature splicing processing on the user feature data to obtain a first user feature vector.
And after the user characteristic data is obtained, encoding processing is carried out to obtain the encoding characters which can be processed by the system. The user feature data comprises user dense features and user sparse features.
Wherein, the encoding processing module 420 includes:
and an encoding sub-module (not shown) for performing encoding processing on the user dense features and the user sparse features respectively.
When the user characteristic data is coded, the method comprises the following steps: the dense features of the users are processed by adopting One-Hot coding; the sparse characteristics of the user are coded by Embedding; then a feature splicing operation.
And a splicing submodule (not shown) for performing feature splicing processing on the encoded data to obtain a first user feature vector.
And a full connection layer processing module 430, configured to convert the first user feature vector into a second user feature vector with a preset length after passing through two full connection layers.
The first user characteristic vector obtained by splicing is provided for a deep network model for learning, and is converted into a second user characteristic vector with a preset length after passing through two full-connection layers, wherein the number of layers of the network can be set according to the effect without limitation, and experiments show that generally two full-connection layers can obtain good effect.
And the clustering processing module 440 is configured to perform incremental clustering on the second user feature vector, and divide users corresponding to the second user feature vector into different clusters.
And performing incremental clustering on the second user characteristic vector, specifically clustering users corresponding to the second user characteristic vector, and dividing the users into different clusters, wherein when a new characteristic vector is added, the clustering center is kept unchanged, and the cluster to which the user belongs can be changed, namely the relationship between the user and the clustering center can be changed.
The new user does not purchase any commodity, and cannot classify the commodity according to the commodity purchasing preference obtained by the commodity browsing, clicking, collecting and purchasing information, but can obtain clusters to which other users similar to the conditions of occupation, income, life or work place and the like of the new user belong according to the equipment information, the geographic position information and the demographic information, and the new user is classified into corresponding clusters.
A preference obtaining module 450, configured to obtain a preference set of users of each cluster;
acquiring preference sets of all users in each cluster, and collecting information of all commodities clicked or purchased by all users to form a preference set; illustratively, there are three users A, B, C in a cluster, and the items purchased by A are hamburger A1 and French fries A2; b, the purchased commodities are ice cream B1 and fried chicken B2, and the clicked and browsed commodity is a coupon B3; c purchased goods, hamburger a1, fried chicken chop B2, and full barrel C1. Then the set of commercial preferences in this category are hamburger a1, french fries a2, ice cream B1, fried chicken B2, coupon B3, and full barrel C1.
And the commodity recommending module 460 is used for recommending the preference set to all the users in the corresponding cluster.
The users in the cluster have the same or similar preferences and have the motivation of purchasing commodities purchased by other users in the cluster, so that the preference set acquired by the method is recommended to all the users in the cluster corresponding to the preference set, and the conversion rate of the recommended commodities can be improved.
According to the embodiment of the invention, the user characteristics are analyzed by acquiring the historical data of the user, the user characteristics are processed into the characteristic vectors with preset lengths to carry out incremental clustering, the preference set of similar users is obtained and recommended to the related users, the problems that the recommended commodities are not matched with the expected consumption preferences of the users or the recommended commodities are single, so that the conversion rate of the recommended commodities is low and the commodities cannot be accurately recommended to new users are solved, and the effects of improving the accuracy and conversion rate of the recommended commodities and accurately recommending the new users who do not purchase the commodities are realized.
On the basis of the foregoing embodiment, as shown in fig. 4B, optionally, the clustering module includes:
and the pre-clustering submodule 441 is configured to pre-cluster the second user feature vector by using a clustering algorithm.
And (3) processing the second user characteristic vector by adopting a clustering algorithm, such as a K-Means clustering algorithm, specifically, the K-Means clustering algorithm is an iterative solution clustering analysis algorithm, dividing data into K groups in advance, randomly selecting K user characteristics as initial clustering centers, calculating the distance between each user characteristic and each seed clustering center, and allocating the user corresponding to each user characteristic to the nearest clustering center. The cluster centers and the objects assigned to them represent a cluster. The cluster center of the cluster is recalculated for each sample assigned based on the existing user characteristics in the cluster. This process will be repeated until some termination condition is met. The termination condition may be that no (or a minimum number) of user features are reassigned to different clusters, no (or a minimum number) cluster centers are changed again, and the sum of squared errors is locally minimal.
The saving sub-module 442 is configured to save the pre-clustered cluster centers and the class labels.
And storing the clustering center and the class label obtained by pre-clustering, keeping the clustering center unchanged, keeping the class label as the same mark of the same cluster, and keeping the meaning of the class label unchanged for a long time, so that the system can learn the characteristic better.
And the calculating submodule 443 is configured to obtain a feature vector newly added by the user, and calculate a distance to each cluster center.
At present, incremental clustering is mainly studied to regard incremental data as time series data or data in a specific sequence, and the technical scheme of the embodiment obtains a new feature vector of a user each time by using a last clustering result, and calculates the distance between the new feature vector of the user and each clustering center.
And the clustering submodule 444 is used for dividing the users corresponding to the newly added feature vectors into the nearest clustering centers.
According to the calculated clustering of the newly added feature vector of the user and each clustering center, the newly added feature vector of one user is divided into the existing clusters every time, namely the newly added feature vector of the user is divided into the cluster with the center nearest to the newly added feature vector of the user and the center is moved to the newly added feature vector, namely the newly added feature vector of the user does not influence the original division, and the method has the advantages that all data do not need to be re-clustered every time.
The cluster center is kept unchanged for a period of time, and the relationship between the user and the cluster center can be changed according to the cluster to which the user feature vector belongs, for example, one user may belong to the cluster today and belong to another cluster tomorrow. The advantage of this is that the meaning of the same cluster label remains unchanged for a long time, which facilitates the better learning of the feature by the ranking layer.
And an update sub-module 445 for updating the cluster center of each class.
And during the service low peak period, re-clustering the current user feature vector data, and updating the clustering center of each class in a full amount so as to eliminate the local deviation possibly introduced by incremental clustering and improve the accuracy of the system.
Example four
Fig. 5 is a schematic structural diagram of a merchandise recommendation apparatus according to a fourth embodiment of the present invention, as shown in fig. 5, the apparatus includes a processor 50, a memory 51, an input device 52, and an output device 53; the number of processors 50 in the device may be one or more, and one processor 50 is taken as an example in fig. 5; the processor 50, the memory 51, the input device 52 and the output device 53 in the apparatus may be connected by a bus or other means, which is exemplified in fig. 5.
The memory 51 is used as a computer-readable storage medium and can be used for storing software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the product recommendation method in the embodiment of the present invention (for example, the user data acquisition module 410, the encoding processing module 420, the full connection layer processing module 430, the clustering processing module 440, the preference acquisition module 450, and the product recommendation module 460 in the subject product recommendation system). The processor 50 executes various functional applications and data processing of the device by executing software programs, instructions and modules stored in the memory 51, that is, implements the above-described commodity recommendation method.
The memory 51 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal, and the like. Further, the memory 51 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, the memory 51 may further include memory located remotely from the processor 50, which may be connected to the device over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input device 52 is operable to receive input numeric or character information and to generate key signal inputs relating to user settings and function controls of the apparatus. The output device 53 may include a display device such as a display screen.
EXAMPLE five
An embodiment of the present invention further provides a storage medium containing computer-executable instructions, which when executed by a computer processor, perform a search method, the method including:
acquiring historical data of a user, and preprocessing the historical data to obtain user characteristic data;
coding the user characteristic data and splicing the characteristics to obtain a first user characteristic vector;
converting the first user characteristic vector into a second user characteristic vector with a preset length after passing through two full-connection layers;
performing incremental clustering on the second user characteristic vector to divide users corresponding to the second user characteristic vector into different clusters;
acquiring preference sets of all users of each cluster;
and recommending the preference set to all users in the corresponding cluster.
Of course, the storage medium provided by the embodiment of the present invention contains computer-executable instructions, and the computer-executable instructions are not limited to the method operations described above, and may also perform related operations in the product recommendation method provided by any embodiment of the present invention.
From the above description of the embodiments, it is obvious for those skilled in the art that the present invention can be implemented by software and necessary general hardware, and certainly, can also be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute the methods according to the embodiments of the present invention.
It should be noted that, in the embodiment of the above search apparatus, each included unit and module are merely divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be implemented; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (9)
1. A method for recommending an article, comprising:
acquiring historical data of a user, and preprocessing the historical data to obtain user characteristic data;
coding and splicing the user characteristic data to obtain a first user characteristic vector;
converting the first user characteristic vector into a second user characteristic vector with a preset length after passing through two full-connection layers;
performing incremental clustering on the second user characteristic vector to divide users corresponding to the second user characteristic vector into different clusters;
acquiring preference sets of all users of each cluster;
and recommending the preference set to all users in the corresponding cluster.
2. The commodity recommendation method according to claim 1, wherein the incrementally clustering the second user feature vector and the classifying users corresponding to the second user feature vector into different clusters comprises:
pre-clustering the second user characteristic vector by adopting a clustering algorithm;
saving the clustering centers and class labels of the pre-clusters;
acquiring a newly added feature vector of a user, and calculating the distance from the feature vector to each clustering center;
dividing the user corresponding to the newly added feature vector to the nearest clustering center;
the cluster center for each class is updated.
3. The item recommendation method according to claim 1, wherein said obtaining a set of preferences of the user of each cluster comprises:
acquiring commodity information generated by each user of one cluster to form a preference pre-set;
judging whether the number of commodities in the preference pre-set is larger than a preset number or not;
if the number of the commodities is larger than the preset number, sorting the commodities in the preference pre-set, and sequentially selecting the commodities with the preset number to form a preference set;
and if the number of the commodities is not more than the preset number, determining the preference preset set as the preference set.
4. The product recommendation method according to claim 1, wherein the user feature data includes a dense user feature and a sparse user feature, and when the encoding processing is performed on the user feature data, the method includes:
the user dense features are processed by adopting One-Hot coding; the user sparse feature is coded by using Embedding.
5. The item recommendation method according to claim 1, wherein the user's history data includes but is not limited to device information, geographical location information and demographic information, and browsing, clicking, collecting and purchasing information of items.
6. An article recommendation system, comprising:
the user data acquisition module is used for acquiring historical data of a user and obtaining user characteristic data through preprocessing;
the coding processing module is used for carrying out coding processing and feature splicing processing on the user feature data to obtain a first user feature vector;
the full-connection layer processing module is used for converting the first user characteristic vector into a second user characteristic vector with a preset length after passing through two full-connection layers;
the clustering processing module is used for carrying out incremental clustering on the second user characteristic vector and dividing users corresponding to the second user characteristic vector into different clusters;
the preference acquisition module is used for acquiring a preference set of users of each cluster;
and the commodity recommending module is used for recommending the preference set to all users in the corresponding cluster.
7. The item recommendation system of claim 6, wherein the clustering module comprises:
the pre-clustering sub-module is used for pre-clustering the second user characteristic vector by adopting a clustering algorithm;
the storage submodule is used for storing the clustering center and the class label of the pre-clustering;
the computing submodule is used for acquiring a newly added feature vector of a user and computing the distance from each clustering center;
the clustering submodule is used for dividing the user corresponding to the newly added feature vector to a clustering center with the nearest distance;
and the updating submodule is used for updating the clustering center of each class.
8. An article recommendation apparatus characterized by comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the item recommendation method of any one of claims 1-5.
9. A computer-readable storage medium on which a computer program is stored, the program, when being executed by a processor, implementing the item recommendation method according to any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010898544.XA CN112070577A (en) | 2020-08-31 | 2020-08-31 | Commodity recommendation method, system, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010898544.XA CN112070577A (en) | 2020-08-31 | 2020-08-31 | Commodity recommendation method, system, equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112070577A true CN112070577A (en) | 2020-12-11 |
Family
ID=73666146
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010898544.XA Pending CN112070577A (en) | 2020-08-31 | 2020-08-31 | Commodity recommendation method, system, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112070577A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112669128A (en) * | 2021-01-28 | 2021-04-16 | 天元大数据信用管理有限公司 | Real-time clustering recommendation method, device and medium |
| CN112820379A (en) * | 2021-01-26 | 2021-05-18 | 吾征智能技术(北京)有限公司 | Intelligent diet recommendation method and system integrating user images |
| CN113379454A (en) * | 2021-06-09 | 2021-09-10 | 北京房江湖科技有限公司 | Data processing method and device, electronic equipment and storage medium |
| CN113392359A (en) * | 2021-08-18 | 2021-09-14 | 腾讯科技（深圳）有限公司 | Multi-target prediction method, device, equipment and storage medium |
| CN113674063A (en) * | 2021-08-27 | 2021-11-19 | 卓尔智联(武汉)研究院有限公司 | Shopping recommendation method, shopping recommendation device and electronic equipment |
| CN116911926A (en) * | 2023-06-26 | 2023-10-20 | 杭州火奴数据科技有限公司 | Advertisement marketing recommendation method based on data analysis |
| CN117422528A (en) * | 2023-11-24 | 2024-01-19 | 久睦本(海南)科技股份有限公司 | Commodity sales management method and system based on small program |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102750647A (en) * | 2012-06-29 | 2012-10-24 | 南京大学 | Merchant recommendation method based on transaction network |
| CN106651546A (en) * | 2017-01-03 | 2017-05-10 | 重庆邮电大学 | Intelligent community oriented electronic commerce information recommendation method |
| CN110458627A (en) * | 2019-08-19 | 2019-11-15 | 华南师范大学 | A kind of commodity sequence personalized recommendation method of user oriented preference of dynamic |
| CN110598111A (en) * | 2019-09-17 | 2019-12-20 | 山东爱城市网信息技术有限公司 | Personalized recommendation system and method based on block chain |
| CN111597428A (en) * | 2019-02-20 | 2020-08-28 | 中国科学院声学研究所 | Recommendation method for splicing user and article with q-separation k sparsity |
-
2020
- 2020-08-31 CN CN202010898544.XA patent/CN112070577A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102750647A (en) * | 2012-06-29 | 2012-10-24 | 南京大学 | Merchant recommendation method based on transaction network |
| CN106651546A (en) * | 2017-01-03 | 2017-05-10 | 重庆邮电大学 | Intelligent community oriented electronic commerce information recommendation method |
| CN111597428A (en) * | 2019-02-20 | 2020-08-28 | 中国科学院声学研究所 | Recommendation method for splicing user and article with q-separation k sparsity |
| CN110458627A (en) * | 2019-08-19 | 2019-11-15 | 华南师范大学 | A kind of commodity sequence personalized recommendation method of user oriented preference of dynamic |
| CN110598111A (en) * | 2019-09-17 | 2019-12-20 | 山东爱城市网信息技术有限公司 | Personalized recommendation system and method based on block chain |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112820379A (en) * | 2021-01-26 | 2021-05-18 | 吾征智能技术(北京)有限公司 | Intelligent diet recommendation method and system integrating user images |
| CN112820379B (en) * | 2021-01-26 | 2024-02-02 | 吾征智能技术(北京)有限公司 | Intelligent diet recommendation method and system integrating user images |
| CN112669128A (en) * | 2021-01-28 | 2021-04-16 | 天元大数据信用管理有限公司 | Real-time clustering recommendation method, device and medium |
| CN113379454A (en) * | 2021-06-09 | 2021-09-10 | 北京房江湖科技有限公司 | Data processing method and device, electronic equipment and storage medium |
| CN113392359A (en) * | 2021-08-18 | 2021-09-14 | 腾讯科技（深圳）有限公司 | Multi-target prediction method, device, equipment and storage medium |
| CN113674063A (en) * | 2021-08-27 | 2021-11-19 | 卓尔智联(武汉)研究院有限公司 | Shopping recommendation method, shopping recommendation device and electronic equipment |
| CN113674063B (en) * | 2021-08-27 | 2024-01-12 | 卓尔智联(武汉)研究院有限公司 | Shopping recommendation method, shopping recommendation device and electronic equipment |
| CN116911926A (en) * | 2023-06-26 | 2023-10-20 | 杭州火奴数据科技有限公司 | Advertisement marketing recommendation method based on data analysis |
| CN117422528A (en) * | 2023-11-24 | 2024-01-19 | 久睦本(海南)科技股份有限公司 | Commodity sales management method and system based on small program |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112070577A (en) | Commodity recommendation method, system, equipment and medium | |
| CN111523976B (en) | Commodity recommendation method and device, electronic equipment and storage medium | |
| CN108363821A (en) | A kind of information-pushing method, device, terminal device and storage medium | |
| Miao et al. | Context‐based dynamic pricing with online clustering | |
| CN108073667B (en) | Method for generating user browsing attributes, and non-transitory computer readable medium | |
| CN112115363A (en) | Recommendation method, computing device and storage medium | |
| CN116431931B (en) | Real-time incremental data statistical analysis method | |
| CN110852785A (en) | User grading method, device and computer readable storage medium | |
| CN114240555A (en) | Click rate prediction model training method and device and click rate prediction method and device | |
| CN114861050A (en) | Feature fusion recommendation method and system based on neural network | |
| CN111340522A (en) | Resource recommendation method, device, server and storage medium | |
| US20210090105A1 (en) | Technology opportunity mapping | |
| CN113326432A (en) | Model optimization method based on decision tree and recommendation method | |
| Bhargavi et al. | Comparative study of consumer purchasing and decision pattern analysis using pincer search based data mining method | |
| Park | MLP modeling for search advertising price prediction | |
| CN114429384B (en) | Intelligent product recommendation method and system based on e-commerce platform | |
| CN107943943B (en) | User similarity determination method and device, electronic equipment and storage medium | |
| CN112632275B (en) | Crowd clustering data processing method, device and equipment based on personal text information | |
| CN112527851B (en) | User characteristic data screening method and device and electronic equipment | |
| CN112732891A (en) | Office course recommendation method and device, electronic equipment and medium | |
| CN113887613A (en) | Deep learning method, device and equipment based on attention mechanism and storage medium | |
| Paul et al. | User-item-based hybrid recommendation system by employing mahout framework | |
| CN113792952A (en) | Method and apparatus for generating a model | |
| CN111460300A (en) | Network content pushing method and device and storage medium | |
| CN113793161A (en) | Advertisement delivery method, advertisement delivery device, readable storage medium and electronic device |
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 |