CN110852785B - User grading method, device and computer readable storage medium - Google Patents
User grading method, device and computer readable storage medium Download PDFInfo
- Publication number
- CN110852785B CN110852785B CN201910976136.9A CN201910976136A CN110852785B CN 110852785 B CN110852785 B CN 110852785B CN 201910976136 A CN201910976136 A CN 201910976136A CN 110852785 B CN110852785 B CN 110852785B
- Authority
- CN
- China
- Prior art keywords
- user
- product
- grading
- users
- behavior record
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000006399 behavior Effects 0.000 claims description 118
- 230000006870 function Effects 0.000 claims description 18
- 210000002569 neuron Anatomy 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 16
- 230000004913 activation Effects 0.000 claims description 12
- 238000011176 pooling Methods 0.000 claims description 12
- 238000004378 air conditioning Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 abstract description 7
- 238000013145 classification model Methods 0.000 abstract description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 abstract description 2
- 238000010606 normalization Methods 0.000 description 8
- 230000008901 benefit Effects 0.000 description 6
- 238000006243 chemical reaction Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000004590 computer program Methods 0.000 description 1
- 230000000007 visual effect 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
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
-
- 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
Abstract
The invention relates to an artificial intelligence technology, and discloses a user grading method, which comprises the following steps: obtaining an original user tag set, an original user behavior record set and a product classification set, processing the original user tag set, the original user behavior record set and the product classification set to obtain a primary user tag set and a primary user behavior record set, performing classification operation on the primary user tag set to obtain a common user tag set and a core user tag set, preprocessing the primary user behavior record set to obtain a training set, inputting the training set, the common user tag set and the core user tag set into a classification model for training, receiving a user behavior record, inputting the behavior record into the classification model to obtain a user classification set, establishing a user product corresponding model according to the user classification set and the product classification set, and recommending a product corresponding to a user according to the user product corresponding model. The invention also provides a user grading device and a computer readable storage medium. The invention can realize accurate and efficient user grading function.
Description
Technical Field
The present invention relates to the field of artificial intelligence technology, and in particular, to a method, an apparatus, and a computer readable storage medium for classifying users according to user behaviors.
Background
User classification is an operation means for classifying users according to their characteristics, such as how to classify clients according to client information for life insurance companies, thereby providing suitable products more precisely and automatically. At present, based on user classification, a naive Bayes or a support vector machine is mostly used as a basis, for example, the naive Bayes establish a conditional probability model according to users, the user classification is solved according to the conditional probability model, the support vector machine classifies the users by constructing a hyperplane, and the method can achieve the purpose of user classification, but lacks of combining analysis between the users and products, and is difficult to achieve the purpose of accurate user classification.
Disclosure of Invention
The invention provides a user grading method, a user grading device and a computer readable storage medium, and mainly aims to provide a method capable of combining analysis between a user and a product to achieve the purpose of accurate user grading.
In order to achieve the above object, the present invention provides a user grading method, including:
Acquiring an original user tag set, an original user behavior record set and a product classification set, and performing exception removal processing on the original user tag set and the original user behavior record set to obtain a primary user tag set and a primary user behavior record set which have corresponding relations;
performing grading operation on the primary user tag set according to the user attribute to obtain a common user tag set and a core user tag set;
the primary user behavior record set is preprocessed to obtain a training set, the common user tag set and the core user tag set are input into a pre-built grading model to be trained to obtain a loss value, the relation between the loss value and a preset loss value is judged, if the loss value is larger than the preset loss value, the grading model continues training internal parameters, and if the loss value is smaller than or equal to the preset loss value, the grading model completes training;
receiving a behavior record of a user, inputting the behavior record of the user into the grading model which is trained, and grading the user to obtain a user grading set;
and establishing a user-product corresponding model according to the user grading set and the product grading set, and recommending the product corresponding to the user according to the user-product corresponding model.
Optionally, the hierarchical model includes an input layer, a convolution layer, a pooling layer, a full connection layer, and an output layer; a kind of electronic device with high-pressure air-conditioning system
The training comprises:
receiving the primary user behavior record set by using the input layer, training by using the convolution layer, the pooling layer and the full-connection layer in combination with an activation function according to the primary user behavior record set to obtain a training value and outputting a predicted value through an output layer;
and carrying out loss calculation on the predicted value and the common user tag set as well as the core user tag set to obtain a loss value.
Optionally, the activation function includes:
wherein O is j Representing the output value of the j-th neuron of the full connection layer, I j An input value representing a j-th neuron of the output layer, t representing the total amount of neurons of the output layer, e being an infinite non-circulating fraction;
the loss calculation includes:
wherein s is j For the predicted value, k is the number of primary user tag sets, y i For the common user tag set, y i ′ A tag set for the core user.
Optionally, the establishing the user-product correspondence model includes:
calculating the preference similarity of different users m and n to the ith grade product in the product grading set according to the user grading set and the product grading set;
Calculating the similarity of user behavior habits in the user grading set;
and constructing the user-product corresponding model based on the preference similarity and the user behavior habit similarity.
Optionally, the preference similarity includes:
wherein ps (m, n) is the bias of different users m, n to the ith grade product in the product grading setGood similarity, wherein m, n represent different user level data in the user grading set, m is user level data of core users in the user grading set, n is user level data of common users in the user grading set, ps represents the product grading set, i is different grading products, r mi Representing the number of users, r, of core users at the ith level in the product hierarchy ni Representing the number of users of the common user at the ith level in the product classification set, r m Representing the total number of users, r, of the core users in the user hierarchy n And representing the total number of users of the common users in the user grading set, wherein a and b are correlation coefficients of the preference similarity.
In addition, in order to achieve the above object, the present invention provides a user grading apparatus, including a memory and a processor, wherein the memory stores a user grading program that can be executed on the processor, and the user grading program when executed by the processor implements the steps of:
Acquiring an original user tag set, an original user behavior record set and a product classification set, and performing exception removal processing on the original user tag set and the original user behavior record set to obtain a primary user tag set and a primary user behavior record set which have corresponding relations;
performing grading operation on the primary user tag set according to the user attribute to obtain a common user tag set and a core user tag set;
the primary user behavior record set is preprocessed to obtain a training set, the common user tag set and the core user tag set are input into a pre-built grading model to be trained to obtain a loss value, the relation between the loss value and a preset loss value is judged, if the loss value is larger than the preset loss value, the grading model continues training internal parameters, and if the loss value is smaller than or equal to the preset loss value, the grading model completes training;
receiving a behavior record of a user, inputting the behavior record of the user into the grading model which is trained, and grading the user to obtain a user grading set;
and establishing a user-product corresponding model according to the user grading set and the product grading set, and recommending the product corresponding to the user according to the user-product corresponding model.
Optionally, the hierarchical model includes an input layer, a convolution layer, a pooling layer, a full connection layer, and an output layer; a kind of electronic device with high-pressure air-conditioning system
The training comprises:
receiving the primary user behavior record set by using the input layer, training by using the convolution layer, the pooling layer and the full-connection layer in combination with an activation function according to the primary user behavior record set to obtain a training value and outputting a predicted value through an output layer;
and carrying out loss calculation on the predicted value and the common user tag set as well as the core user tag set to obtain a loss value.
Optionally, the activation function includes:
wherein O is j Representing the output value of the j-th neuron of the full connection layer, I j An input value representing a j-th neuron of the output layer, t representing the total amount of neurons of the output layer, e being an infinite non-circulating fraction;
the loss calculation includes:
wherein s is j For the predicted value, k is the number of primary user tag sets, y i For the common user tag set, y i ′ A tag set for the core user.
Optionally, the establishing the user-product correspondence model includes:
calculating the preference similarity of different users m and n to the ith grade product in the product grading set according to the user grading set and the product grading set;
Calculating the similarity of user behavior habits in the user grading set;
and constructing the user-product corresponding model based on the preference similarity and the user behavior habit similarity.
In addition, to achieve the above object, the present invention also provides a computer-readable storage medium having stored thereon a user grading program executable by one or more processors to implement the steps of the user grading method as described above.
The invention carries out exception removal processing on the original user tag set, the original user behavior record set and the product classification set, obtains the common user tag set and the core user tag set through classification operation, can divide users through first classification operation, trains a classification model constructed in advance, judges user behaviors by utilizing the classification model which is already trained, establishes a user-product corresponding model according to the user classification set and the product classification set, achieves final user product recommendation, and can effectively improve classification efficiency by dividing and recommending three steps in total. Therefore, the user grading method, the user grading device and the computer readable storage medium can realize accurate and efficient user grading.
Drawings
FIG. 1 is a flow chart of a user classification method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating an internal structure of a user classification apparatus according to an embodiment of the present invention;
fig. 3 is a schematic block diagram of a user classification procedure in the user classification apparatus according to an embodiment of the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The invention provides a user grading method. Referring to fig. 1, a flow chart of a user classification method according to an embodiment of the invention is shown. The method may be performed by an apparatus, which may be implemented in software and/or hardware.
In this embodiment, the user grading method includes:
s1, acquiring an original user tag set, an original user behavior record set and a product classification set, and performing exception removal processing on the original user tag set and the original user behavior record set to obtain a primary user tag set and a primary user behavior record set which have corresponding relations.
Preferably, the user grading method of the present invention performs grading operation on the user according to the preset attribute. If the telecom company divides the user into light telephone charge consumption and heavy telephone charge consumption according to the monthly telephone charge consumption record of the user, the monthly times of the user, the working types of the user and other attributes; the insurance company divides the users into non-insurance users, light insurance users, heavy insurance users and the like according to the attributes of the insurance cost, the academic history of the users, the number of family members of the users and the like of the users on the insurance every year, so when a new user is received, the invention ranks the users according to the preset attributes of the new user.
In the preferred embodiment of the present invention, the original user behavior record set is the preset attribute. Such as a life support company, the original user behavior record set generally comprises: the total value of the life insurance products purchased by the user in the life insurance company and the total times of the purchased life insurance products, the types of the life insurance products collected by the user on the webpage of the life insurance company, the times of the life insurance products browsed by the user on the webpage of the life insurance company and the like.
Further, the original user tag set is the number of different users, such as the user a001, the user a002, the user B001, etc. in a digital and alphabetical manner.
Preferably, in the preferred embodiment of the present invention, the original product is classified according to the attributes of the product to obtain the product classification set, and generally, different products may be classified according to different product attributes. For example, life insurance companies have child insurance, health insurance, endowment insurance, security insurance, two-full insurance, accident insurance, additional insurance, red insurance and the like, so that the life insurance can be divided into popular products, long-tail products, benefit-base products and other product classification sets according to the sales condition of each dangerous species.
Preferably, the exception removing processing is to perform one-to-one correspondence between the relationship between the original user tag set and the original user behavior record set, and if the one-to-one correspondence is not satisfied, reject redundant data in the original user tag set and the original user behavior record set, or add missing data in the original user tag set and the original user behavior record set.
Preferably, the exception removal processing according to the present invention includes, for example: the life insurance company performs grading division on users, records a user A001 in the original user tag set, records 13000 yuan of total value of life insurance products purchased by the user A001 in the life insurance company in the original user behavior record set, and 9 times of total times of purchasing life insurance products, wherein the types of life insurance products collected by the user A001 on the webpage of the life insurance company are 18, the times of browsing life insurance products on the webpage of the life insurance company are 621 times, and the like, so that the one-to-one correspondence relationship is satisfied; if the original user tag set records the user B001, and the original user behavior record set does not record any behavior record of the user B001, removing the user B001 from the original user tag set; similarly, if the original user behavior record set records a user C001 behavior record, but the original user tag set does not present the user C001, the user C001 is added to the original user tag set.
In summary, the exception handling can effectively prevent the phenomenon of low classification accuracy caused by incomplete data or unequal data in the earlier stage.
And S2, grading the primary user tag set according to the user attribute to obtain a common user tag set and a core user tag set.
Preferably, the user is classified by the life insurance company according to the classification, the primary user tag set may be classified into a common user tag set and a core user tag set according to whether the user purchases the common product and the core product of the life insurance company, the common user tag set is the common product of the life insurance company, and the core user tag set is the core product, and whether the common product is purchased is not required to be concerned. Therefore, when the invention receives the prediction of the new user, whether the new user is a core user or a common user is predicted.
S3, preprocessing the primary user behavior record set to obtain a training set, inputting the training set, the common user tag set and the core user tag set into a pre-constructed grading model to train to obtain a loss value, judging the magnitude relation between the loss value and a preset loss value, if the loss value is larger than the preset loss value, continuing training internal parameters by the grading model, and if the loss value is smaller than or equal to the preset loss value, completing training by the grading model.
Preferably, the preprocessing operation comprises qualitative and quantitative conversion, digital extraction, digital normalization and the like.
The qualitative and quantitative conversion is to convert non-numeric attributes in the primary user behavior record set to numeric attributes. If the online store A performs consumption classification on the consumption capability of the user, and if the user collects the product a, the product b, the product e, the product s and the like of the online store A, the qualitative and quantitative conversion is to count all the collected products and output a result.
The digital extraction is to extract digital attributes such as product collection number, browsing number and the like.
The number normalization is to map the numbers in the primary user behavior record set to the range of 0,1, and is generally necessary to reduce the gap due to the magnitude, since the magnitude of each number is generally very different, e.g., the total amount purchased for life insurance must be much greater than the number of purchases. Preferably, the normalization method is as follows:
wherein X is normal For the normalized number, X is the number before normalization, X max The primary user behavior record set has the largest number, X min The number with the smallest value in the set is recorded for the primary user behavior.
In a preferred embodiment of the present invention, the hierarchical model includes an input layer, a convolution layer, a pooling layer, a full connection layer, and an output layer.
Preferably, the training process is specifically: and receiving the primary user behavior record set by using the input layer, obtaining a training value by using the convolution layer, the pooling layer and the full-connection layer in combination with an activation function, outputting a predicted value set by using an output layer, carrying out loss calculation on the predicted value, the common user tag set and the core user tag set to obtain a loss value, judging the magnitude of the loss value and a preset loss value, continuing training if the loss value is larger than the preset loss value, and exiting training if the loss value is smaller than the preset loss value.
In a preferred embodiment of the present invention, the activation function comprises a Softmax function. The Softmax function is:
wherein O is j Representing the output value of the j-th neuron of the full connection layer, I j An input value representing a j-th neuron of the output layer, t representing the total amount of neurons of the output layer, e being an infinite non-circulating fraction;
the loss calculation includes:
wherein s is j For the predicted value, k is the number of primary user tag sets, y i For the common user tag set, y i ′ A tag set for the core user.
S4, receiving a behavior record of the user, inputting the behavior record of the user into the classification model which is trained, and classifying the user to obtain a user classification set.
Preferably, for example, the life insurance client grading model is used for grading and dividing the users, and when the life insurance client grading model receives all the behavior records of the users A to Z, the life insurance client grading model grades the life insurance client grading model into a core user, a common user and the like according to all the behavior records of the users A to Z.
S5, establishing a user-product corresponding model according to the user grading set and the product grading set, and recommending the product corresponding to the user according to the user-product corresponding model.
As described above, the product classification set may include popular products, long tail products, base products, and the like. Wherein the popular products are the most popular products; the long tail product is a product with no strong demand or poor sales; the interest-based product is a product which is generally sold, but has unique benefits which are obviously different from other products, and can be accepted by users.
Preferably, the purpose of the user-product correspondence model is to correspond users with different levels in the user grading set to products with different product attributes in the product grading set, so that the purpose of performing associated recommendation on the users with different levels and the products with different product attributes is achieved, and further, the purpose of more accurate intelligent recommendation is achieved.
Further, the user-product correspondence model comprises the steps of calculating the preference similarity of different users m and n to the i-th level product in the product grading set, calculating the user behavior habit similarity, and constructing the user-product correspondence model based on the preference similarity and the user behavior habit similarity.
Preferably, the preference similarity is:
wherein ps (m, n) is the preference similarity of different users m, n to the i-th level of products in the product classification set, wherein m, n represents different levels of users in the user classification set, such as m may represent core users, n may represent common users, ps represents the product classification set, i is different classified products, such as the popular product, long tail product, and benefit base product, and thus when i e ps, i=1, represents one of the levels of the product classification set, such as the popular product, r mi Representing the number of users, r, of core users at the ith level in the product hierarchy ni Representing the number of users of the common user at the ith level in the product classification set, r m Representing the total number of users, r, of the core users in the user hierarchy n And representing the total number of users of the common users in the user grading set, wherein a and b are correlation coefficients of the preference similarity.
Further, as described in the above step S1, the user behavior habit similarity includes the total value of the life insurance products purchased by the user in the life insurance company and the total number of times of the life insurance products purchased, the types of the life insurance products collected by the user on the web page of the life insurance company, the number of times of the life insurance products browsed by the user on the web page of the life insurance company, and the like, so that the user behavior habit similarity among different users can be obtained according to the behavior records of the users at different levels in the user classification set.
Preferably, the user-product correspondence model is finally:
sim(x，[m，n])＝α*ps(x，[m，n])+β*hs(x，[m，n])
and when the unknown user x is received, the user-product corresponding model sequentially carries out similarity calculation on the unknown user x and the users [ m, n ] with different levels, sequentially traverses the user with the highest similarity with the unknown user x, and therefore, the corresponding products with the same level are recommended to the unknown user x according to the level of the user with the highest similarity, so that the purpose of product recommendation is achieved.
The invention also provides a user grading device. Referring to fig. 2, an internal structure of a user grading device according to an embodiment of the present invention is shown.
In this embodiment, the user grading apparatus 1 may be a PC (Personal Computer ), or a terminal device such as a smart phone, a tablet computer, a portable computer, or a server. The user grading device 1 comprises at least a memory 11, a processor 12, a communication bus 13, and a network interface 14.
The memory 11 includes at least one type of readable storage medium including flash memory, a hard disk, a multimedia card, a card memory (e.g., SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. The memory 11 may in some embodiments be an internal storage unit of the user grading device 1, such as a hard disk of the user grading device 1. The memory 11 may also be an external storage device of the user grading apparatus 1 in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) or the like, which are provided on the user grading apparatus 1. Further, the memory 11 may also include both an internal storage unit and an external storage device of the user grading apparatus 1. The memory 11 may be used not only for storing application software installed in the user gradation apparatus 1 and various types of data, for example, codes of the user gradation program 01 and the like, but also for temporarily storing data that has been output or is to be output.
The processor 12 may in some embodiments be a central processing unit (Central Processing Unit, CPU), controller, microcontroller, microprocessor or other data processing chip for executing program code or processing data stored in the memory 11, e.g. for executing a user-grading program 01 or the like.
The communication bus 13 is used to enable connection communication between these components.
The network interface 14 may optionally comprise a standard wired interface, a wireless interface (e.g. WI-FI interface), typically used to establish a communication connection between the apparatus 1 and other electronic devices.
Optionally, the device 1 may further comprise a user interface, which may comprise a Display (Display), an input unit such as a Keyboard (Keyboard), and a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, or the like. The display may also be referred to as a display screen or display unit, as appropriate, for displaying information processed in the user grading device 1 and for displaying a visual user interface.
Fig. 2 shows only a user grading device 1 with components 11-14 and a user grading program 01, it being understood by a person skilled in the art that the structure shown in fig. 1 does not constitute a limitation of the user grading device 1, and may comprise fewer or more components than shown, or may combine certain components, or a different arrangement of components.
In the embodiment of the apparatus 1 shown in fig. 2, the memory 11 has stored therein a user rating program 01; the processor 12 performs the following steps when executing the user rating program 01 stored in the memory 11:
step one, acquiring an original user tag set, an original user behavior record set and a product classification set, and performing exception removal processing on the original user tag set and the original user behavior record set to obtain a primary user tag set and a primary user behavior record set with corresponding relations.
Preferably, the user grading method of the present invention performs grading operation on the user according to the preset attribute. If the telecom company divides the user into light telephone charge consumption and heavy telephone charge consumption according to the monthly telephone charge consumption record of the user, the monthly times of the user, the working types of the user and other attributes; the insurance company divides the users into non-insurance users, light insurance users, heavy insurance users and the like according to the attributes of the insurance cost, the academic history of the users, the number of family members of the users and the like of the users on the insurance every year, so when a new user is received, the invention ranks the users according to the preset attributes of the new user.
In the preferred embodiment of the present invention, the original user behavior record set is the preset attribute. Such as a life support company, the original user behavior record set generally comprises: the total value of the life insurance products purchased by the user in the life insurance company and the total times of the purchased life insurance products, the types of the life insurance products collected by the user on the webpage of the life insurance company, the times of the life insurance products browsed by the user on the webpage of the life insurance company and the like.
Further, the original user tag set is the number of different users, such as the user a001, the user a002, the user B001, etc. in a digital and alphabetical manner.
Preferably, in the preferred embodiment of the present invention, the original product is classified according to the attributes of the product to obtain the product classification set, and generally, different products may be classified according to different product attributes. For example, life insurance companies have child insurance, health insurance, endowment insurance, security insurance, two-full insurance, accident insurance, additional insurance, red insurance and the like, so that the life insurance can be divided into popular products, long-tail products, benefit-base products and other product classification sets according to the sales condition of each dangerous species.
Preferably, the exception removing processing is to perform one-to-one correspondence between the relationship between the original user tag set and the original user behavior record set, and if the one-to-one correspondence is not satisfied, reject redundant data in the original user tag set and the original user behavior record set, or add missing data in the original user tag set and the original user behavior record set.
Preferably, the exception removal processing according to the present invention includes, for example: the life insurance company performs grading division on users, records a user A001 in the original user tag set, records 13000 yuan of total value of life insurance products purchased by the user A001 in the life insurance company in the original user behavior record set, and 9 times of total times of purchasing life insurance products, wherein the types of life insurance products collected by the user A001 on the webpage of the life insurance company are 18, the times of browsing life insurance products on the webpage of the life insurance company are 621 times, and the like, so that the one-to-one correspondence relationship is satisfied; if the original user tag set records the user B001, and the original user behavior record set does not record any behavior record of the user B001, removing the user B001 from the original user tag set; similarly, if the original user behavior record set records a user C001 behavior record, but the original user tag set does not present the user C001, the user C001 is added to the original user tag set.
In summary, the exception handling can effectively prevent the phenomenon of low classification accuracy caused by incomplete data or unequal data in the earlier stage.
Step two, grading operation is carried out on the primary user tag set according to the user attribute, and a common user tag set and a core user tag set are obtained.
Preferably, the user is classified by the life insurance company according to the classification, the primary user tag set may be classified into a common user tag set and a core user tag set according to whether the user purchases the common product and the core product of the life insurance company, the common user tag set is the common product of the life insurance company, and the core user tag set is the core product, and whether the common product is purchased is not required to be concerned. Therefore, when the invention receives the prediction of the new user, whether the new user is a core user or a common user is predicted.
Step three, preprocessing the primary user behavior record set to obtain a training set, inputting the training set, the common user tag set and the core user tag set into a pre-constructed grading model to train to obtain a loss value, judging the magnitude relation between the loss value and a preset loss value, if the loss value is larger than the preset loss value, continuing training internal parameters by the grading model, and if the loss value is smaller than or equal to the preset loss value, completing training by the grading model.
Preferably, the preprocessing operation comprises qualitative and quantitative conversion, digital extraction, digital normalization and the like.
The qualitative and quantitative conversion is to convert non-numeric attributes in the primary user behavior record set to numeric attributes. If the online store A performs consumption classification on the consumption capability of the user, and if the user collects the product a, the product b, the product e, the product s and the like of the online store A, the qualitative and quantitative conversion is to count all the collected products and output a result.
The digital extraction is to extract digital attributes such as product collection number, browsing number and the like.
The number normalization is to map the numbers in the primary user behavior record set to the range of 0,1, and is generally necessary to reduce the gap due to the magnitude, since the magnitude of each number is generally very different, e.g., the total amount purchased for life insurance must be much greater than the number of purchases. Preferably, the normalization method is as follows:
wherein X is normal For the normalized number, X is the number before normalization, X max The primary user behavior record set has the largest number, X min The number with the smallest value in the set is recorded for the primary user behavior.
In a preferred embodiment of the present invention, the hierarchical model includes an input layer, a convolution layer, a pooling layer, a full connection layer, and an output layer.
Preferably, the training process is specifically: and receiving the primary user behavior record set by using the input layer, obtaining a training value by using the convolution layer, the pooling layer and the full-connection layer in combination with an activation function, outputting a predicted value set by using an output layer, carrying out loss calculation on the predicted value, the common user tag set and the core user tag set to obtain a loss value, judging the magnitude of the loss value and a preset loss value, continuing training if the loss value is larger than the preset loss value, and exiting training if the loss value is smaller than the preset loss value.
In a preferred embodiment of the present invention, the activation function comprises a Softmax function. The Softmax function is:
wherein O is j Representing the output value of the j-th neuron of the full connection layer, I j An input value representing a j-th neuron of the output layer, t representing the total amount of neurons of the output layer, e being an infinite non-circulating fraction;
the loss calculation includes:
wherein s is j For the predicted value, k is the number of primary user tag sets, y i For the common user tag set, y i ′ A tag set for the core user.
And step four, receiving a behavior record of the user, inputting the behavior record of the user into the classification model which is trained, and classifying the user to obtain a user classification set.
Preferably, for example, the life insurance client grading model is used for grading and dividing the users, and when the life insurance client grading model receives all the behavior records of the users A to Z, the life insurance client grading model grades the life insurance client grading model into a core user, a common user and the like according to all the behavior records of the users A to Z.
And fifthly, establishing a user-product corresponding model according to the user grading set and the product grading set, and recommending the product corresponding to the user according to the user-product corresponding model.
As described above, the product classification set may include popular products, long tail products, base products, and the like. Wherein the popular products are the most popular products; the long tail product is a product with no strong demand or poor sales; the interest-based product is a product which is generally sold, but has unique benefits which are obviously different from other products, and can be accepted by users.
Preferably, the purpose of the user-product correspondence model is to correspond users with different levels in the user grading set to products with different product attributes in the product grading set, so that the purpose of performing associated recommendation on the users with different levels and the products with different product attributes is achieved, and further, the purpose of more accurate intelligent recommendation is achieved.
Further, the user-product correspondence model comprises the steps of calculating the preference similarity of different users m and n to the i-th level product in the product grading set, calculating the user behavior habit similarity, and constructing the user-product correspondence model based on the preference similarity and the user behavior habit similarity.
Preferably, the preference similarity is:
wherein ps (m, n) is the preference similarity of different users m, n to the i-th level of products in the product classification set, wherein m, n represents different levels of users in the user classification set, such as m may represent core users, n may represent common users, ps represents the product classification set, i is different classified products, such as the popular product, long tail product, and benefit base product, and thus when i e ps, i=1, represents one of the levels of the product classification set, such as the popular product, r mi Representing the number of users, r, of core users at the ith level in the product hierarchy ni Representing the number of users of the common user at the ith level in the product classification set, r m Representing the total number of users, r, of the core users in the user hierarchy n And representing the total number of users of the common users in the user grading set, wherein a and b are correlation coefficients of the preference similarity.
Further, the user behavior habit similarity is as described in the above step one, for example, the user behavior record performed by the life support company on the user includes the total value of life support products purchased by the user in the life support company and the total number of times of life support products purchased, the number of types of life support products collected by the user on the webpage of the life support company, the number of times of life support products browsed by the user on the webpage of the life support company, and the like, so that the user behavior habit similarity between different users can be obtained according to the behavior record calculation of the users at different levels in the user classification set.
Preferably, the user-product correspondence model is finally:
sim(x，[m，n])＝α*ps(x，[m，n])+β*hs(x，[m，n])
and when the unknown user x is received, the user-product corresponding model sequentially carries out similarity calculation on the unknown user x and the users [ m, n ] with different levels, sequentially traverses the user with the highest similarity with the unknown user x, and therefore, the corresponding products with the same level are recommended to the unknown user x according to the level of the user with the highest similarity, so that the purpose of product recommendation is achieved.
Alternatively, in other embodiments, the user-grading program may be divided into one or more modules, and one or more modules are stored in the memory 11 and executed by one or more processors (the processor 12 in this embodiment) to implement the present invention, where the modules refer to a series of instruction segments of a computer program capable of performing a specific function, for describing the execution of the user-grading program in the user-grading device.
For example, referring to fig. 3, a program module diagram of a user grading program in an embodiment of the user grading apparatus according to the present invention is shown, where the user grading program may be divided into a data receiving and processing module 10, a grading operation module 20, a model training module 30, and a user grading and product recommendation module 40, which are exemplary:
the data receiving and processing module 10 is configured to: and obtaining an original user tag set, an original user behavior record set and a product classification set, and performing exception removal processing on the original user tag set and the original user behavior record set to obtain a primary user tag set and a primary user behavior record set which have corresponding relations.
The hierarchical operation module 20 is configured to: and grading the primary user tag set according to the user attribute to obtain a common user tag set and a core user tag set.
The model training module 30 is configured to: and (3) preprocessing the primary user behavior record set to obtain a training set, inputting the training set, the common user tag set and the core user tag set into a pre-constructed grading model to train to obtain a loss value, judging the magnitude relation between the loss value and a preset loss value, if the loss value is larger than the preset loss value, continuing training internal parameters by the grading model, and if the loss value is smaller than or equal to the preset loss value, completing training by the grading model.
The user rating and product recommendation module 40 is configured to: and receiving a behavior record of a user, inputting the behavior record of the user into the grading model which is trained, grading the user to obtain a user grading set, establishing a user-product corresponding model according to the user grading set and the product grading set, and recommending a product corresponding to the user according to the user-product corresponding model.
The functions or operation steps implemented when the program modules of the data receiving and processing module 10, the classifying operation module 20, the model training module 30, the user classifying and product recommending module 40 and the like are executed are substantially the same as those of the above embodiments, and are not described herein again.
In addition, an embodiment of the present invention also proposes a computer-readable storage medium having stored thereon a user-grading program executable by one or more processors to implement the following operations:
and obtaining an original user tag set, an original user behavior record set and a product classification set, and performing exception removal processing on the original user tag set and the original user behavior record set to obtain a primary user tag set and a primary user behavior record set which have corresponding relations.
And grading the primary user tag set according to the user attribute to obtain a common user tag set and a core user tag set.
And (3) preprocessing the primary user behavior record set to obtain a training set, inputting the training set, the common user tag set and the core user tag set into a pre-constructed grading model to train to obtain a loss value, judging the magnitude relation between the loss value and a preset loss value, if the loss value is larger than the preset loss value, continuing training internal parameters by the grading model, and if the loss value is smaller than or equal to the preset loss value, completing training by the grading model.
And receiving a behavior record of a user, inputting the behavior record of the user into the grading model which is trained, grading the user to obtain a user grading set, establishing a user-product corresponding model according to the user grading set and the product grading set, and recommending a product corresponding to the user according to the user-product corresponding model.
It should be noted that, the foregoing reference numerals of the embodiments of the present invention are merely for describing the embodiments, and do not represent the advantages and disadvantages of the embodiments. And the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, apparatus, article or method that comprises the element.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) as described above, comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the method according to the embodiments of the present invention.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (7)
1. A method of user classification, the method comprising:
acquiring an original user tag set, an original user behavior record set and a product classification set, and performing exception removal processing on the original user tag set and the original user behavior record set to obtain a primary user tag set and a primary user behavior record set which have corresponding relations;
performing grading operation on the primary user tag set according to the user attribute to obtain a common user tag set and a core user tag set;
the primary user behavior record set is preprocessed to obtain a training set, the common user tag set and the core user tag set are input into a pre-built grading model to be trained to obtain a loss value, the relation between the loss value and a preset loss value is judged, if the loss value is larger than the preset loss value, the grading model continues training internal parameters, and if the loss value is smaller than or equal to the preset loss value, the grading model completes training;
Receiving a behavior record of a user, inputting the behavior record of the user into the grading model which is trained, and grading the user to obtain a user grading set;
establishing a user-product corresponding model according to the user grading set and the product grading set, and recommending a product corresponding to a user according to the user-product corresponding model;
wherein the establishing the user-product correspondence model comprises: calculating the preference similarity of different users m and n to the ith grade product in the product grading set according to the user grading set and the product grading set, wherein the formula of the preference similarity is as follows:
wherein ps (m, n) is the preference similarity of different users m, n to the i-th level product in the product classification set, wherein m, n represents different user level data in the user classification set, m is user level data of a core user in the user classification set, n is user level data of a common user in the user classification set, ps represents the product classification set, i is different classified product, r mi Representing the number of users, r, of core users at the ith level in the product hierarchy ni Representing the number of users of the common user at the ith level in the product classification set, r m Representing the total number of users, r, of the core users in the user hierarchy n Representing the total number of users of the common users in the user grading set, wherein a and b are correlation coefficients of the preference similarity; calculating according to the behavior records of the users in the user hierarchical set to obtain the user behavior habit similarity among the users in the user hierarchical set, and constructing the user-product corresponding model sim (x, [ m, n) based on the preference similarity and the user behavior habit similarity])＝α*ps(x,[m,n])+β*hs(x,[m,n]) Wherein sim (x, [ m, n)]) For the user-product corresponding model, alpha and beta are weights of the user-product corresponding model, respectively, x is an unknown user, hs (x, [ m, n)]) For unknown user x and users of different levels [ m, n ]]Similarity of user behavior habit between ps (x, [ m, n)]) For unknown user x and users of different levels [ m, n ]]The user-product correspondence model sequentially associates the unknown user x with users [ m, n ] of different levels when the unknown user x is received]And performing similarity calculation, traversing the user with the highest similarity with the unknown user x, recommending corresponding products with the same level to the unknown user x according to the level of the user with the highest similarity, wherein the user behavior habit comprises the total value of the products purchased by the user and the total times of the purchased products, the types of the products collected by the user on a company webpage selling the products, and the times of the products browsed by the user on the webpage.
2. The user grading method of claim 1 wherein the grading model comprises an input layer, a convolution layer, a pooling layer, a full connection layer, and an output layer; a kind of electronic device with high-pressure air-conditioning system
The training comprises:
receiving the primary user behavior record set by using the input layer, training by using the convolution layer, the pooling layer and the full-connection layer in combination with an activation function according to the primary user behavior record set to obtain a training value and outputting a predicted value through an output layer;
and carrying out loss calculation on the predicted value and the common user tag set as well as the core user tag set to obtain a loss value.
3. The user grading method according to claim 2, wherein the activation function comprises:
wherein O is j Representing the j-th neuron of the full connection layerOutput value of I j An input value representing a j-th neuron of the output layer, t representing the total amount of neurons of the output layer, e being an infinite non-circulating fraction;
the loss calculation includes:
wherein s is j For the predicted value, k is the number of primary user tag sets, y i For the common user tag set, y' i A tag set for the core user.
4. A user grading device, comprising a memory and a processor, wherein the memory has stored thereon a user grading program operable on the processor, which when executed by the processor, performs the steps of:
Acquiring an original user tag set, an original user behavior record set and a product classification set, and performing exception removal processing on the original user tag set and the original user behavior record set to obtain a primary user tag set and a primary user behavior record set which have corresponding relations;
performing grading operation on the primary user tag set according to the user attribute to obtain a common user tag set and a core user tag set;
the primary user behavior record set is preprocessed to obtain a training set, the common user tag set and the core user tag set are input into a pre-built grading model to be trained to obtain a loss value, the relation between the loss value and a preset loss value is judged, if the loss value is larger than the preset loss value, the grading model continues training internal parameters, and if the loss value is smaller than or equal to the preset loss value, the grading model completes training;
receiving a behavior record of a user, inputting the behavior record of the user into the grading model which is trained, and grading the user to obtain a user grading set;
establishing a user-product corresponding model according to the user grading set and the product grading set, and recommending a product corresponding to a user according to the user-product corresponding model;
Wherein the establishing the user-product correspondence model comprises: calculating the preference similarity of different users m and n to the ith grade product in the product grading set according to the user grading set and the product grading set, wherein the formula of the preference similarity is as follows:
wherein ps (m, n) is the preference similarity of different users m, n to the i-th level product in the product classification set, wherein m, n represents different user level data in the user classification set, m is user level data of a core user in the user classification set, n is user level data of a common user in the user classification set, ps represents the product classification set, i is different classified product, r mi Representing the number of users, r, of core users at the ith level in the product hierarchy ni Representing the number of users of the common user at the ith level in the product classification set, r m Representing the total number of users, r, of the core users in the user hierarchy n Representing the total number of users of the common users in the user grading set, wherein a and b are correlation coefficients of the preference similarity;
calculating according to the behavior records of the users in the user grading set to obtain the user behavior habit similarity between the users in the user grading set, constructing a user-product corresponding model sim (x, [ m, n ])=α×ps (x, [ m, n ]) based on the preference similarity and the user behavior habit similarity, wherein sim (x, [ m, n ]) is the user-product corresponding model, α, β is the weight of the user-product corresponding model, x is the unknown user, hs (x, [ m, n ]) is the user behavior habit similarity between the unknown user x and the users [ m, n ] of different levels, ps (x, [ m, n ]) is the preference similarity of the unknown user x and the users [ m, n ] of different levels for different products, when the unknown user x is received, the user-product corresponding model sequentially calculates the similarity of the unknown user x and the users [ m, n ] of different levels, α, β is the weight of the user-product corresponding model, x is the unknown user, hs (x, [ m, n ]) is the user behavior habit similarity between the unknown user x and the users [ m, n ] of different levels, ps is the preference similarity between the unknown user x and the users [ m, n ] of different levels, ps is the user's of the unknown user, the product is the user has the greatest similarity of the user's, and the user's has the greatest value of the user's, and the user's product is the product corresponding to the user class, the product is the product, and the product has the highest has the greatest value is the priority by the recommendation by the user class.
5. The user grading device of claim 4, wherein the grading model comprises an input layer, a convolution layer, a pooling layer, a full connection layer, and an output layer; a kind of electronic device with high-pressure air-conditioning system
The training comprises:
receiving the primary user behavior record set by using the input layer, training by using the convolution layer, the pooling layer and the full-connection layer in combination with an activation function according to the primary user behavior record set to obtain a training value and outputting a predicted value through an output layer;
and carrying out loss calculation on the predicted value and the common user tag set as well as the core user tag set to obtain a loss value.
6. The user grading device of claim 5, wherein the activation function comprises:
wherein O is j Representing the output value of the j-th neuron of the full connection layer, I j An input value representing a j-th neuron of the output layer, t representing the total amount of neurons of the output layer, e being an infinite non-circulating fraction;
the loss calculation includes:
wherein s is j For the predicted value, k is the number of primary user tag sets, y i For the common user tag set, y i ′ A tag set for the core user.
7. A computer readable storage medium having stored thereon a user rating program executable by one or more processors to implement the steps of the user rating method of any of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910976136.9A CN110852785B (en) | 2019-10-12 | 2019-10-12 | User grading method, device and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910976136.9A CN110852785B (en) | 2019-10-12 | 2019-10-12 | User grading method, device and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110852785A CN110852785A (en) | 2020-02-28 |
| CN110852785B true CN110852785B (en) | 2023-11-21 |
Family
ID=69596616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910976136.9A Active CN110852785B (en) | 2019-10-12 | 2019-10-12 | User grading method, device and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110852785B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112102011A (en) * | 2020-10-13 | 2020-12-18 | 平安科技（深圳）有限公司 | User grade prediction method, device, terminal and medium based on artificial intelligence |
| CN114860788B (en) * | 2022-04-22 | 2023-05-05 | 广东车卫士信息科技有限公司 | Technical popularization information service system and method |
| CN115511124B (en) * | 2022-09-27 | 2023-04-18 | 上海网商电子商务有限公司 | Customer grading method based on after-sale maintenance records |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009057103A2 (en) * | 2007-10-31 | 2009-05-07 | Musicgenome.Com Inc. | Content-based recommendations across domains |
| WO2010037286A1 (en) * | 2008-09-27 | 2010-04-08 | 华为技术有限公司 | Collaborative filtering-based recommendation method and system |
| CN107820619A (en) * | 2017-09-21 | 2018-03-20 | 达闼科技（北京）有限公司 | One kind classification interactive decision making method, interactive terminal and cloud server |
| CN110246012A (en) * | 2019-06-14 | 2019-09-17 | 哈尔滨哈银消费金融有限责任公司 | Consumer finance Products Show method, apparatus and equipment based on social data |
| CN110264306A (en) * | 2019-05-21 | 2019-09-20 | 平安银行股份有限公司 | Products Show method, apparatus, server and medium based on big data |
| CN110298574A (en) * | 2019-06-21 | 2019-10-01 | 国网辽宁省电力有限公司鞍山供电公司 | A kind of electricity consumption subscriber payment risk rating method based on convolutional neural networks |
-
2019
- 2019-10-12 CN CN201910976136.9A patent/CN110852785B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009057103A2 (en) * | 2007-10-31 | 2009-05-07 | Musicgenome.Com Inc. | Content-based recommendations across domains |
| WO2010037286A1 (en) * | 2008-09-27 | 2010-04-08 | 华为技术有限公司 | Collaborative filtering-based recommendation method and system |
| CN107820619A (en) * | 2017-09-21 | 2018-03-20 | 达闼科技（北京）有限公司 | One kind classification interactive decision making method, interactive terminal and cloud server |
| CN110264306A (en) * | 2019-05-21 | 2019-09-20 | 平安银行股份有限公司 | Products Show method, apparatus, server and medium based on big data |
| CN110246012A (en) * | 2019-06-14 | 2019-09-17 | 哈尔滨哈银消费金融有限责任公司 | Consumer finance Products Show method, apparatus and equipment based on social data |
| CN110298574A (en) * | 2019-06-21 | 2019-10-01 | 国网辽宁省电力有限公司鞍山供电公司 | A kind of electricity consumption subscriber payment risk rating method based on convolutional neural networks |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110852785A (en) | 2020-02-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109493199A (en) | Products Show method, apparatus, computer equipment and storage medium | |
| CN110163476A (en) | Project intelligent recommendation method, electronic device and storage medium | |
| CN107730389A (en) | Electronic installation, insurance products recommend method and computer-readable recording medium | |
| CN110852785B (en) | User grading method, device and computer readable storage medium | |
| US20150324939A1 (en) | Real-estate client management method and system | |
| CN106844407B (en) | Tag network generation method and system based on data set correlation | |
| CN111552870A (en) | Object recommendation method, electronic device and storage medium | |
| CN112818218B (en) | Information recommendation method, device, terminal equipment and computer readable storage medium | |
| CN112328909B (en) | Information recommendation method and device, computer equipment and medium | |
| CN111582932A (en) | Inter-scene information pushing method and device, computer equipment and storage medium | |
| CN111652278A (en) | User behavior detection method and device, electronic equipment and medium | |
| CN111612610A (en) | Risk early warning method and system, electronic equipment and storage medium | |
| CN113742492A (en) | Insurance scheme generation method and device, electronic equipment and storage medium | |
| CN111966886A (en) | Object recommendation method, object recommendation device, electronic equipment and storage medium | |
| CN113592605A (en) | Product recommendation method, device, equipment and storage medium based on similar products | |
| CN113706291A (en) | Fraud risk prediction method, device, equipment and storage medium | |
| CN106294410A (en) | A kind of determination method of personalized information push time and determine system | |
| CN111861679A (en) | Commodity recommendation method based on artificial intelligence | |
| CN113505273B (en) | Data sorting method, device, equipment and medium based on repeated data screening | |
| CN115222433A (en) | Information recommendation method and device and storage medium | |
| CN115204971B (en) | Product recommendation method, device, electronic equipment and computer readable storage medium | |
| CN113656690B (en) | Product recommendation method and device, electronic equipment and readable storage medium | |
| CN110389963A (en) | The recognition methods of channel effect, device, equipment and storage medium based on big data | |
| CN114693409A (en) | Product matching method, device, computer equipment, storage medium and program product | |
| CN113689233A (en) | Advertisement putting and selecting method and corresponding device, equipment and medium thereof |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |