CN110264306B - Big data-based product recommendation method, device, server and medium - Google Patents
Big data-based product recommendation method, device, server and medium Download PDFInfo
- Publication number
- CN110264306B CN110264306B CN201910424733.0A CN201910424733A CN110264306B CN 110264306 B CN110264306 B CN 110264306B CN 201910424733 A CN201910424733 A CN 201910424733A CN 110264306 B CN110264306 B CN 110264306B
- Authority
- CN
- China
- Prior art keywords
- user
- target
- product
- matrix
- candidate
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention is applicable to the technical field of artificial intelligence, and provides a product recommendation method, a device, a server and a medium based on big data.
Description
Technical Field
The invention belongs to the technical field of artificial intelligence, and particularly relates to a product recommendation method, device, server and medium based on big data.
Background
With the development of artificial intelligence, product recommendation methods based on big data are increasingly applied to various industries, and these products include not only physical products that can be sold in physical stores, but also virtual internet products such as movies, music, and financial resources. Obviously, reasonable recommendations for products may improve user experience while saving purchase costs for the user in some cases.
However, when the big data for recommending the product is too complicated, the big data is difficult to reasonably and comprehensively integrate and comb in the prior art, and only one of the user data or the product data is often analyzed, so that overall analysis of the data of the two parties and future prediction are lacking. This results in the prior art that often results in inaccurate recommendations, including: the user is not interested in the recommended product, and the user is difficult to pay for the recommended product, etc.
In summary, the current product recommendation method has the problem that the recommended product has poor adaptation with the user.
Disclosure of Invention
In view of the above, the embodiments of the present invention provide a product recommendation method, device, server and medium based on big data, so as to solve the problem in the prior art that the recommended product has poor adaptation with the user.
The first aspect of the embodiment of the invention provides a product recommendation method based on big data, which comprises the following steps: acquiring a target user matrix for describing target users, and calling a plurality of reference user sets corresponding to user grades respectively; the reference user set corresponding to the user level comprises a plurality of reference user matrixes belonging to the user level and storage time corresponding to each reference user matrix; calculating the similarity between the target user matrix and the reference user matrix corresponding to each user grade according to the reference user matrix corresponding to each user grade and the storage time corresponding to each reference user matrix, and taking the user grade corresponding to the reference user matrix with the highest similarity to the target user matrix as the selected grade; invoking the data value transfer record of the target user, and calculating a target scheduling data value corresponding to the target user according to the data value transfer record of the target user and the selected grade; classifying the target user matrix through a preset classification tree model to obtain leaf nodes of the classification tree model corresponding to the target user matrix, and determining a target product set corresponding to the target user according to the corresponding relation between the preset leaf nodes and the product set; the negative data value and the positive data value of each product in the target product set are called, products with the negative data value not larger than the target scheduling data value in the target product set are used as candidate products, and positive parameters corresponding to each candidate product are calculated according to the positive data value of each candidate product; and recommending the candidate products corresponding to the forward parameters with preset quantity to a target user as selected products in sequence from the largest forward parameter according to the sequence from the large forward parameter to the small forward parameter.
A second aspect of an embodiment of the present invention provides a product recommendation device based on big data, including:
the acquisition module is used for acquiring a target user matrix for describing target users and calling a plurality of reference user sets corresponding to user grades respectively; the reference user set corresponding to the user level comprises a plurality of reference user matrixes belonging to the user level and storage time corresponding to each reference user matrix; the grading module is used for calculating the similarity between the target user matrix and the reference user matrix corresponding to each user grade according to the reference user matrix corresponding to each user grade and the storage time corresponding to each reference user matrix, and taking the user grade corresponding to the reference user matrix with the highest similarity to the target user matrix as the selected grade; the calling module is used for calling the data value transfer record of the target user and calculating a target scheduling data value corresponding to the target user according to the data value transfer record of the target user and the selected grade; the classification module is used for classifying the target user matrix through a preset classification tree model to obtain leaf nodes of the classification tree model corresponding to the target user matrix, and determining a target product set corresponding to the target user according to the corresponding relation between the preset leaf nodes and the product set; the calculation module is used for calling the negative data value and the positive data value of each product in the target product set, taking the product with the negative data value not greater than the target scheduling data value in the target product set as a candidate product, and calculating the positive parameter corresponding to each candidate product according to the positive data value of each candidate product; and the recommending module is used for recommending candidate products corresponding to the forward parameters with preset quantity to a target user as selected products in sequence from the largest forward parameter according to the sequence from the largest forward parameter to the small forward parameter.
A third aspect of an embodiment of the present invention provides a server, including: a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the method provided by the first aspect of the embodiments of the invention when the computer program is executed by the processor.
A fourth aspect of the embodiments of the present invention provides a computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the method provided by the first aspect of the embodiments of the present invention.
In the embodiment of the invention, the grade of the target user is intelligently determined according to the data corresponding to each user grade by acquiring the target user matrix for describing the target user and calling the reference user sets corresponding to a plurality of user grades and determining the selected grade corresponding to the target user according to the reference user set corresponding to the user grade; the target scheduling data value of the target user is calculated based on the data value transfer record and the selected level of the target user, so that a callable threshold value is set according to the related historical data of the target user, and the safety of subsequent data scheduling is ensured; classifying the target user matrix through a preset classification tree model, determining a target product set corresponding to a target user, calling negative data values and positive data values of all products in the target product set, taking products with the negative data values not larger than target scheduling data values in the target product set as candidate products, and calculating positive parameters corresponding to all candidate products according to the positive data values of all candidate products so as to determine a plurality of candidate products suitable for the target user after analyzing the product data of all candidate products; and recommending the candidate products corresponding to the forward parameters with preset quantity as selected products to a target user in sequence from the largest forward parameter according to the sequence from the largest forward parameter to the smallest forward parameter, so that user data and product data are considered simultaneously, and the accuracy of product recommendation is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of an implementation of a big data based product recommendation method provided by an embodiment of the present invention;
FIG. 2 is a flowchart of a specific implementation of a big data based product recommendation method S102 according to an embodiment of the present invention;
FIG. 3 is a flowchart of a specific implementation of a big data based product recommendation method S105 according to an embodiment of the present invention;
FIG. 4 is a block diagram of a product recommendation device based on big data according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a server according to an embodiment of the present invention.
Detailed Description
In the following description, for purposes of explanation and not limitation, specific details are set forth such as the particular system architecture, techniques, etc., in order to provide a thorough understanding of the embodiments of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present invention with unnecessary detail.
In order to illustrate the technical scheme of the invention, the following description is made by specific examples.
Fig. 1 shows an implementation flow of a big data based product recommendation method according to an embodiment of the present invention, where the method flow includes steps S101 to S106. The specific implementation principle of each step is as follows.
In S101, a target user matrix for describing a target user is acquired, and a reference user set corresponding to a plurality of user classes is called; the reference user set corresponding to the user level comprises a plurality of reference user matrixes belonging to the user level and storage time corresponding to each reference user matrix.
In the embodiment of the invention, after receiving the task of recommending a product for a target user, the server acquires a target user matrix of the target user, which is generated in advance. Since the target user matrix is generated according to the identity data (such as age, sex, region, occupation, and bank card type) and the historical transaction data (such as purchase product type, month purchase price, and purchase time) of the target user, the target user matrix is equivalent to a user portrait, and can be used for describing the target user and representing relevant parameters of the target user.
Notably, in the embodiment of the present invention, there are a plurality of user levels, each corresponding to a reference user set, where each reference user matrix stored, similar to the target user matrix, is capable of functioning to describe a user. Optionally, the storage time corresponding to the reference user matrix is: when a user is determined to be of a user class, the reference user matrix of the user is stored in the server for a certain reference user set time.
In S102, according to the reference user matrix corresponding to each user level and the storage time corresponding to each reference user matrix, calculating the similarity between the target user matrix and the reference user matrix corresponding to each user level, and taking the user level corresponding to the reference user matrix with the highest similarity to the target user matrix as the selected level.
Obviously, by the reference user sets corresponding to the user grades, the target user matrix can be classified by judging which reference user matrix in the reference user set is closer to the target user matrix. Considering summary of historical trends and prediction of future trends, the embodiment of the present invention needs to consider time factors in the subsequent grading process, so that each reference user matrix corresponds to one storage time.
As an embodiment of the present invention, as shown in fig. 2, the step S102 includes:
s1021, calculating the time difference between the storage time corresponding to each reference user matrix and the current time, and taking the time difference as the time difference corresponding to each reference user matrix.
It will be appreciated that in order to show that the influence of the reference user matrices stored at different times on the final calculation result is different in the embodiment of the present invention, it is necessary to first calculate the time difference between each storage time and the current time.
S1022, calculating the grade matrix corresponding to each user grade through the grade matrix formula.
Optionally, the rank matrix formula is:
wherein the P is f For the class matrix corresponding to the user class f, the v fi Representing a reference user matrix i corresponding to a user class f, said t fi And the time difference of the reference user matrix i corresponding to the user level f is represented, Y is a preset constant, and n is the number of the reference user matrices corresponding to the user level f.
It will be appreciated that different reference user matrices may be weighted differently when calculating the rank matrix by the rank matrix formula described above. Obviously, in the above rank matrix formula, the larger the time difference corresponding to the reference user matrix, the smaller the weight thereof in calculating the rank matrix.
In the embodiment of the invention, the grade matrix which can be used for representing the user grade is determined through the integrated calculation of each reference user matrix corresponding to the user grade.
S1023, calculating cosine similarity of the target user matrix and the grade matrix corresponding to each user grade, and taking the cosine similarity as similarity of the target user matrix and the reference user matrix corresponding to each user grade.
In the embodiment of the invention, the similarity between the target user matrix and the reference user matrix corresponding to each user grade is calculated by adopting a cosine similarity calculation formula in consideration of the possible sparseness of the grade matrix.
In S103, the data value transfer record of the target user is called, and the target scheduling data value corresponding to the target user is calculated according to the data value transfer record of the target user and the selected level.
In the embodiment of the invention, in order to ensure that the target user can purchase the recommended product in the subsequent process, the condition of unsatisfied recommendation results caused by excessive purchase burden is avoided, so that the target scheduling data value corresponding to the target user needs to be calculated first. It is apparent that the target schedule data value in the embodiment of the present invention is used to represent the highest amount that the target user is eligible for scheduling.
Optionally, the data value transfer record contains an inflow data value per unit time, a current total inflow data value, and a current total outflow data value. It will be appreciated that the data value transfer record corresponds to a stream of target users in accounts provided by the server. In this case, the specific implementation manner of S103 is:
firstly, determining a grade coefficient corresponding to the selected grade according to a corresponding relation between a preset user grade and the grade coefficient.
Finally, the formula is passed: tran=unit×classpre+ (ClassPre-1) × (InTo-OutTo) calculates a target scheduling data value corresponding to the target user, where Tran represents the target scheduling data value, unit represents a Unit time inflow data value of the target user, classPre represents a class coefficient corresponding to the selected class, inTo represents a current total inflow data value of the target user, and OutTo represents a current total outflow data value of the target user.
In the embodiment of the present invention, the inflow data value per unit time is an inflow data value per unit time before the current time, the parameter is used to describe the recent funds inflow situation of the target user, the current total inflow data value is used to describe the long-term funds inflow situation of the target user, and the current total outflow data value is used to describe the long-term funds outflow situation of the target user.
It will be appreciated that the final calculated target schedule data value is proportional to the ranking factor, the inflow data value per unit time and the current total inflow data value, and the target schedule data value is inversely proportional to the current total outflow data value. Accordingly, the target schedule data value calculated by the above formula may represent the degree of acceptance of the target user for the product in one aspect, since various factors are considered.
In S104, classifying the target user matrix through a preset classification tree model, obtaining leaf nodes of the classification tree model corresponding to the target user matrix, and determining a target product set corresponding to the target user according to a corresponding relation between the preset leaf nodes and the product set.
It will be appreciated that a predetermined classification tree model comprises a plurality of nodes, each node other than leaf nodes certainly corresponding to a classification rule, each classification rule classifying for one or more data categories. As described above, since the target user matrix is generated according to the identity data of the target user and the historical transaction data, the server can naturally parse the data value corresponding to a certain data class from the target user matrix. Therefore, through analysis of the target user matrix, data values corresponding to all data types can be obtained, then the target user matrix is classified through one node in each layer of the preset classification tree model, and finally the target user matrix is classified into one leaf node in the lowest layer of the preset classification tree model. Therefore, further through the corresponding relation between the leaf nodes and the product sets, the product set corresponding to the target user matrix can be determined and used as the target product set.
In the embodiment of the invention, each product set including the target product set comprises a plurality of products.
In S105, the negative data value and the positive data value of each product in the target product set are called, products with negative data values not greater than the target scheduling data value in the target product set are used as candidate products, and the positive parameters corresponding to each candidate product are calculated according to the positive data values of each candidate product.
In the embodiment of the present invention, the steps before S105 basically use the data of the target users as the reference factors of classification and grading, but in order to improve the fitness between the last recommended product and the target users, in the embodiment of the present invention, in S105, the data of the products in the target product set needs to be analyzed, and the products need to be screened after the analysis.
Notably, each product in the embodiment of the present invention corresponds to 2 types of data values, which are a negative data value and a plurality of positive data values, wherein the negative data value represents the cost to be paid for purchasing the product by the user, and the positive data value represents the value historically brought by different users by the product, so that one product corresponds to a plurality of positive data values. In the embodiment of the invention, the negative data value and the positive data value are set and evaluated in advance, and are stored in the server after corresponding to the product in advance. Therefore, after the target product set is determined, the positive data value and the negative data value of each product in the target product set can be directly called.
It will be appreciated that, based on the description of the target schedule data value in S103 above, once the negative data value of a product is greater than the target schedule data value of the target user, if the product is recommended to the target user, the probability of purchase by the target user will be small and the satisfaction of the recommendation will be poor. Therefore, products with negative data values larger than the target scheduling data value of the target user are firstly excluded, and only products with negative data values not larger than the target scheduling data value in the target product set are used as candidate products. And in the subsequent process, further screening the candidate products according to the forward data values of the candidate products.
As an embodiment of the present invention, as shown in fig. 3, the step S105 includes:
s1051, taking the maximum value in the forward data values of all candidate products as the forward extremum corresponding to each candidate product independently, and taking the maximum value in the forward data values of all candidate products as the forward maximum value corresponding to all candidate products together.
It will be appreciated that, in accordance with the above description, a candidate product corresponds to a plurality of forward data values, so that a maximum value may be selected therefrom as the forward extremum for the candidate product. On the other hand, there is a total maximum value in the forward data values corresponding to all candidate products, that is, one maximum value in the forward extremum values of all candidate products.
S1052, calculating the average value of the forward data values of each candidate product as the forward data average value corresponding to each candidate product individually, and calculating the total average value of the forward data values of all candidate products as the total forward data average value corresponding to all candidate products together.
S1053, calculating the variance of the forward data value of each candidate product as the forward data variance corresponding to each candidate product individually, and calculating the average value of the forward data variances corresponding to all candidate products individually as the average value of the forward variances corresponding to all candidate products together.
S1054, calculating the forward parameters corresponding to the candidate products through a preset parameter calculation formula.
Optionally, the preset parameter calculation formula is:
wherein said p x For the forward parameter corresponding to the candidate product x, the r x For the average value of the forward data corresponding to the candidate product x, R is the total average value of the forward data corresponding to all the candidate products, and s x For the positive extremum of the candidate product x, S is the positive maximum value commonly corresponding to all candidate products, and t is x And the T is the forward variance average value which corresponds to all candidate products in common.
It will be appreciated that the forward data variance of a candidate product may reflect the stability of the product in value to the target user, with the larger variance proving less stable, so the forward parameter is inversely proportional to the forward data variance. In addition, the embodiment of the invention also considers the forward data average value and the forward extremum of one product at the same time, so that the average value and the maximum value of one product have forward influence on the finally calculated forward parameter. It should be noted that, considering that there may be a large deviation in the parameters of one product, the parameters of one product need to be measured among the parameters of all the products, so in the embodiment of the present invention, the ratio of the parameters of one product to the average conditions of all the products corresponding to the parameters of one product is calculated, and the forward parameter is calculated according to the ratio.
In S106, according to the order of the forward parameters from the large to the small, the candidate products corresponding to the preset number of forward parameters are sequentially recommended to the target user as the selected products from the largest forward parameter.
Optionally, arranging the candidate products according to the sequence of the forward parameters from large to small to generate a product queue, and taking the candidate products with the preset quantity before the product queue as selected products; generating a recommendation report according to the arrangement sequence of each selected product in the product queue, and sending the recommendation report to terminal equipment corresponding to the target user so as to recommend the selected product to the target user.
It can be understood that, by acquiring a target user matrix for describing a target user, and calling a plurality of reference user sets corresponding to user grades, a selected grade corresponding to the target user is determined according to the reference user set corresponding to the user grade, so as to intelligently determine the grade of the target user according to the data corresponding to each user grade; the target scheduling data value of the target user is calculated based on the data value transfer record and the selected level of the target user, so that a callable threshold value is set according to the related historical data of the target user, and the safety of subsequent data scheduling is ensured; classifying the target user matrix through a preset classification tree model, determining a target product set corresponding to a target user, calling negative data values and positive data values of all products in the target product set, taking products with the negative data values not larger than target scheduling data values in the target product set as candidate products, and calculating positive parameters corresponding to all candidate products according to the positive data values of all candidate products so as to determine a plurality of candidate products suitable for the target user after analyzing the product data of all candidate products; and recommending the candidate products corresponding to the forward parameters with preset quantity as selected products to a target user in sequence from the largest forward parameter according to the sequence from the largest forward parameter to the smallest forward parameter, so that user data and product data are considered simultaneously, and the accuracy of product recommendation is improved.
Corresponding to the big data based product recommendation method described in the above embodiments, fig. 4 shows a block diagram of a big data based product recommendation device according to an embodiment of the present invention, and for convenience of explanation, only the parts related to the embodiment of the present invention are shown.
Referring to fig. 4, the apparatus includes:
an obtaining module 401, configured to obtain a target user matrix for describing a target user, and invoke a reference user set corresponding to a plurality of user classes; the reference user set corresponding to the user level comprises a plurality of reference user matrixes belonging to the user level and storage time corresponding to each reference user matrix;
the grading module 402 is configured to calculate, according to the reference user matrices corresponding to the user ranks and the storage time corresponding to the reference user matrices, the similarity between the target user matrix and the reference user matrix corresponding to the user ranks, and take, as the selected rank, the user rank corresponding to the reference user matrix with the highest similarity to the target user matrix;
a retrieving module 403, configured to retrieve the data value transfer record of the target user, and calculate a target scheduling data value corresponding to the target user according to the data value transfer record of the target user and the selected level;
The classification module 404 is configured to classify the target user matrix through a preset classification tree model, obtain leaf nodes of the classification tree model corresponding to the target user matrix, and determine a target product set corresponding to the target user according to a corresponding relationship between the preset leaf nodes and the product set;
a calculation module 405, configured to call a negative data value and a positive data value of each product in the target product set, take a product whose negative data value is not greater than the target scheduling data value in the target product set as a candidate product, and calculate a positive parameter corresponding to each candidate product according to the positive data value of each candidate product;
and a recommending module 406, configured to sequentially recommend, according to the order from the large forward parameter to the small forward parameter, a preset number of candidate products corresponding to the forward parameter as selected products to a target user, starting from the largest forward parameter.
Optionally, the grading module is specifically configured to:
calculating the time difference between the storage time corresponding to each reference user matrix and the current time, and taking the time difference as the time difference corresponding to each reference user matrix;
by the formula:
calculating a grade matrix corresponding to each user grade, wherein the P is as follows f For the class matrix corresponding to the user class f, the v fi Representing a reference user matrix i corresponding to a user class f, said t fi The time difference of a reference user matrix i corresponding to a user level f is represented, Y is a preset constant, and n is the number of the reference user matrices corresponding to the user level f;
and calculating cosine similarity of the target user matrix and a grade matrix corresponding to each user grade, and taking the cosine similarity as similarity of the target user matrix and a reference user matrix corresponding to each user grade.
Optionally, the data value transfer record includes an inflow data value per unit time, a current total inflow data value, and a current total outflow data value, and the scheduling module is specifically configured to:
determining a grade coefficient corresponding to the selected grade according to the corresponding relation between the preset user grade and the grade coefficient;
by the formula: tran=unit×classpre+ (ClassPre-1) × (InTo-OutTo) calculates a target scheduling data value corresponding to the target user, where Tran represents the target scheduling data value, unit represents a Unit time inflow data value of the target user, classPre represents a class coefficient corresponding to the selected class, inTo represents a current total inflow data value of the target user, and OutTo represents a current total outflow data value of the target user.
Optionally, the calculating the forward parameter corresponding to each candidate product according to the forward data value of each candidate product includes:
taking the maximum value in the forward data values of all candidate products as the forward extremum corresponding to each candidate product independently, and taking the maximum value in the forward data values of all candidate products as the forward maximum value corresponding to all candidate products together;
calculating the average value of the forward data values of all candidate products as the forward data average value corresponding to all candidate products independently, and calculating the total average value of the forward data values of all candidate products as the total forward data average value corresponding to all candidate products together;
calculating the variance of the forward data value of each candidate product as the forward data variance corresponding to each candidate product independently, and calculating the average value of the forward data variances corresponding to all candidate products independently as the forward variance average value corresponding to all candidate products together;
by the formula:
calculating the corresponding forward parameters of each candidate product, wherein p is x For the forward parameter corresponding to the candidate product x, the r x For the average value of the forward data corresponding to the candidate product x, R is the total average value of the forward data corresponding to all the candidate products, and s x For the positive extremum of the candidate product x, S is the positive maximum value commonly corresponding to all candidate products, and t is x And the T is the forward variance average value which corresponds to all candidate products in common.
Optionally, the recommending, as the selected product, the candidate product corresponding to the largest forward parameter in the preset number to the target user includes:
arranging the candidate products according to the sequence from large to small of the corresponding forward parameters of the candidate products to generate a product queue, and taking the candidate products with the preset quantity before the product queue as selected products;
generating a recommendation report according to the arrangement sequence of each selected product in the product queue, and sending the recommendation report to terminal equipment corresponding to the target user so as to recommend the selected product to the target user.
It will be appreciated that embodiments of the invention
The method comprises the steps of acquiring a target user matrix for describing target users, calling a plurality of reference user sets corresponding to user grades, determining selected grades corresponding to the target users according to the reference user sets corresponding to the user grades, and intelligently determining the grades of the target users according to data corresponding to the user grades; the target scheduling data value of the target user is calculated based on the data value transfer record and the selected level of the target user, so that a callable threshold value is set according to the related historical data of the target user, and the safety of subsequent data scheduling is ensured; classifying the target user matrix through a preset classification tree model, determining a target product set corresponding to a target user, calling negative data values and positive data values of all products in the target product set, taking products with the negative data values not larger than target scheduling data values in the target product set as candidate products, and calculating positive parameters corresponding to all candidate products according to the positive data values of all candidate products so as to determine a plurality of candidate products suitable for the target user after analyzing the product data of all candidate products; and recommending the candidate products corresponding to the forward parameters with preset quantity as selected products to a target user in sequence from the largest forward parameter according to the sequence from the largest forward parameter to the smallest forward parameter, so that user data and product data are considered simultaneously, and the accuracy of product recommendation is improved.
Fig. 5 is a schematic diagram of a server according to an embodiment of the present invention. As shown in fig. 5, the server 5 of this embodiment includes: a processor 50, a memory 51 and a computer program 52 stored in the memory 51 and executable on the processor 50, such as a big data based product recommendation program. The processor 50, when executing the computer program 52, implements the steps of the various big data based product recommendation method embodiments described above, such as steps 101 to 106 shown in fig. 1. Alternatively, the processor 50, when executing the computer program 52, performs the functions of the modules/units of the apparatus embodiments described above, e.g. the functions of the units 401 to 406 shown in fig. 4.
By way of example, the computer program 52 may be partitioned into one or more modules/units that are stored in the memory 51 and executed by the processor 50 to complete the present invention. The one or more modules/units may be a series of computer program instruction segments capable of performing the specified functions, which instruction segments are used to describe the execution of the computer program 52 in the server 5.
The server 5 may be a computing device such as a desktop computer, a notebook computer, a palm computer, a cloud server, etc. The server may include, but is not limited to, a processor 50, a memory 51. It will be appreciated by those skilled in the art that fig. 5 is merely an example of the server 5 and is not meant to be limiting as the server 5 may include more or fewer components than shown, or may combine certain components, or different components, e.g., the server may further include input and output devices, network access devices, buses, etc.
The processor 50 may be a central processing unit (Central Processing Unit, CPU), other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The storage 51 may be an internal storage unit of the server 5, for example, a hard disk or a memory of the server 5. The memory 51 may be an external storage device of the server 5, 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 server 5. Further, the memory 51 may also include both an internal storage unit and an external storage device of the server 5. The memory 51 is used for storing the computer program and other programs and data required by the server. The memory 51 may also be used to temporarily store data that has been output or is to be output.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules, that is, the internal structure of the apparatus is divided into different functional units or modules, so as to perform all or part of the above-described functions. The functional units and modules in the embodiment may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit, where the integrated units may be implemented in a form of hardware or a form of a software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working process of the units and modules in the above system may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
The integrated modules/units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the present invention may implement all or part of the flow of the method of the above embodiment, or may be implemented by a computer program to instruct related hardware, where the computer program may be stored in a computer readable storage medium.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention, and are intended to be included in the scope of the present invention.
Claims (5)
1. A big data based product recommendation method, comprising:
acquiring a target user matrix for describing target users, and calling a plurality of reference user sets corresponding to user grades respectively; the reference user set corresponding to the user level comprises a plurality of reference user matrixes belonging to the user level and storage time corresponding to each reference user matrix;
calculating the similarity between the target user matrix and the reference user matrix corresponding to each user grade according to the reference user matrix corresponding to each user grade and the storage time corresponding to each reference user matrix, and taking the user grade corresponding to the reference user matrix with the highest similarity to the target user matrix as the selected grade;
invoking the data value transfer record of the target user, and calculating a target scheduling data value corresponding to the target user according to the data value transfer record of the target user and the selected grade;
classifying the target user matrix through a preset classification tree model to obtain leaf nodes of the classification tree model corresponding to the target user matrix, and determining a product set corresponding to the target user as a target product set according to the corresponding relation between the preset leaf nodes and the product set;
The negative data value and the positive data value of each product in the target product set are called, products with the negative data value not larger than the target scheduling data value in the target product set are used as candidate products, and positive parameters corresponding to each candidate product are calculated according to the positive data value of each candidate product;
according to the sequence from the big forward parameter to the small forward parameter, recommending candidate products corresponding to the preset quantity of forward parameters to a target user as selected products in sequence from the largest forward parameter;
according to the reference user matrix corresponding to each user level and the storage time corresponding to each reference user matrix, calculating the similarity between the target user matrix and the reference user matrix corresponding to each user level, including:
calculating the time difference between the storage time corresponding to each reference user matrix and the current time, and taking the time difference as the time difference corresponding to each reference user matrix;
by the formula:
calculating a grade matrix corresponding to each user grade, wherein P f For the class matrix corresponding to the user class f, v fi Representing a reference user matrix i, t corresponding to a user class f fi The time difference of the reference user matrix i corresponding to the user level f is represented, Y is a preset constant, and n is the number of the reference user matrices corresponding to the user level f;
Calculating cosine similarity of the target user matrix and a grade matrix corresponding to each user grade, and taking the cosine similarity as similarity of the target user matrix and a reference user matrix corresponding to each user grade;
the data value transfer record comprises a unit time inflow data value, a current total inflow data value and a current total outflow data value;
and calculating a target scheduling data value corresponding to the target user according to the data value transfer record of the target user and the selected level, wherein the method comprises the following steps:
determining a grade coefficient corresponding to the selected grade according to the corresponding relation between the preset user grade and the grade coefficient;
by the formula: tran=Unit×ClassPre+ (ClassPre-1) × (InTo-OutTo) calculating a target scheduling data value corresponding to the target user, wherein Tran represents the target scheduling data value, unit represents a Unit time inflow data value of the target user, classPre represents a class coefficient corresponding to the selected class, inTo represents a current total inflow data value of the target user, and OutTo represents a current total outflow data value of the target user;
and calculating the forward parameters corresponding to each candidate product according to the forward data value of each candidate product, wherein the forward parameters comprise:
Taking the maximum value in the forward data values of all candidate products as the forward extremum corresponding to each candidate product independently, and taking the maximum value in the forward data values of all candidate products as the forward maximum value corresponding to all candidate products together;
calculating the average value of the forward data values of all candidate products as the forward data average value corresponding to all candidate products independently, and calculating the total average value of the forward data values of all candidate products as the total forward data average value corresponding to all candidate products together;
calculating the variance of the forward data value of each candidate product as the forward data variance corresponding to each candidate product independently, and calculating the average value of the forward data variances corresponding to all candidate products independently as the forward variance average value corresponding to all candidate products together;
by the formula:
computing individual candidate product pairsA corresponding forward parameter, wherein p x For the forward parameter, r, corresponding to the candidate product x x For the average value of the forward data corresponding to the candidate product x, R is the total average value of the forward data corresponding to all the candidate products together, s x S is the positive maximum value, t, of the total candidate products which are commonly corresponding to the positive extremum of the candidate products x x And (3) for the forward data variance of the candidate product x, T is the forward variance average value which corresponds to all the candidate products in common.
2. The big data based product recommendation method of claim 1, wherein the recommending candidate products corresponding to a predetermined number of the forward parameters as selected products to a target user in order from the largest forward parameter according to the order of the forward parameters from the largest forward parameter comprises:
arranging the candidate products according to the sequence from large to small of the corresponding forward parameters of the candidate products to generate a product queue, and taking the candidate products with the preset quantity before the product queue as selected products;
generating a recommendation report according to the arrangement sequence of each selected product in the product queue, and sending the recommendation report to terminal equipment corresponding to the target user so as to recommend the selected product to the target user.
3. A big data based product recommendation device, the device comprising:
the acquisition module is used for acquiring a target user matrix for describing target users and calling a reference user set corresponding to a plurality of user grades; the reference user set corresponding to the user level comprises a plurality of reference user matrixes belonging to the user level and storage time corresponding to each reference user matrix;
The grading module is used for calculating the similarity between the target user matrix and the reference user matrix corresponding to each user grade according to the reference user matrix corresponding to each user grade and the storage time corresponding to each reference user matrix, and taking the user grade corresponding to the reference user matrix with the highest similarity to the target user matrix as the selected grade;
the calling module is used for calling the data value transfer record of the target user and calculating a target scheduling data value corresponding to the target user according to the data value transfer record of the target user and the selected grade;
the classification module is used for classifying the target user matrix through a preset classification tree model to obtain leaf nodes of the classification tree model corresponding to the target user matrix, and determining a target product set corresponding to the target user according to the corresponding relation between the preset leaf nodes and the product set;
the calculation module is used for calling the negative data value and the positive data value of each product in the target product set, taking the product with the negative data value not greater than the target scheduling data value in the target product set as a candidate product, and calculating the positive parameter corresponding to each candidate product according to the positive data value of each candidate product;
The recommending module is used for recommending candidate products corresponding to the forward parameters with preset quantity to a target user as selected products in sequence from the largest forward parameter according to the sequence from the largest forward parameter to the small forward parameter;
the grading module is specifically configured to:
calculating the time difference between the storage time corresponding to each reference user matrix and the current time, and taking the time difference as the time difference corresponding to each reference user matrix;
by the formula:
calculating a grade matrix corresponding to each user grade, wherein P f For the class matrix corresponding to the user class f, v fi Representing a reference user matrix i, t corresponding to a user class f fi The time difference of the reference user matrix i corresponding to the user grade f is represented, Y is a preset constant, and n is the reference corresponding to the user grade fThe number of the user matrix;
calculating cosine similarity of the target user matrix and a grade matrix corresponding to each user grade, and taking the cosine similarity as similarity of the target user matrix and a reference user matrix corresponding to each user grade;
the data value transfer record comprises a unit time inflow data value, a current total inflow data value and a current total outflow data value, and the calling module is specifically configured to:
Determining a grade coefficient corresponding to the selected grade according to the corresponding relation between the preset user grade and the grade coefficient;
by the formula: tran=Unit×ClassPre+ (ClassPre-1) × (InTo-OutTo) calculating a target scheduling data value corresponding to the target user, wherein Tran represents the target scheduling data value, unit represents a Unit time inflow data value of the target user, classPre represents a class coefficient corresponding to the selected class, inTo represents a current total inflow data value of the target user, and OutTo represents a current total outflow data value of the target user;
the computing module is specifically configured to:
taking the maximum value in the forward data values of all candidate products as the forward extremum corresponding to each candidate product independently, and taking the maximum value in the forward data values of all candidate products as the forward maximum value corresponding to all candidate products together;
calculating the average value of the forward data values of all candidate products as the forward data average value corresponding to all candidate products independently, and calculating the total average value of the forward data values of all candidate products as the total forward data average value corresponding to all candidate products together;
calculating the variance of the forward data value of each candidate product as the forward data variance corresponding to each candidate product independently, and calculating the average value of the forward data variances corresponding to all candidate products independently as the forward variance average value corresponding to all candidate products together;
By the formula:
calculating the corresponding forward parameters of each candidate product, wherein p x For the forward parameter, r, corresponding to the candidate product x x For the average value of the forward data corresponding to the candidate product x, R is the total average value of the forward data corresponding to all the candidate products together, s x S is the positive maximum value, t, of the total candidate products which are commonly corresponding to the positive extremum of the candidate products x x And (3) for the forward data variance of the candidate product x, T is the forward variance average value which corresponds to all the candidate products in common.
4. A server comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any of claims 1 to 2 when the computer program is executed.
5. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the method according to any one of claims 1 to 2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910424733.0A CN110264306B (en) | 2019-05-21 | 2019-05-21 | Big data-based product recommendation method, device, server and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910424733.0A CN110264306B (en) | 2019-05-21 | 2019-05-21 | Big data-based product recommendation method, device, server and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110264306A CN110264306A (en) | 2019-09-20 |
| CN110264306B true CN110264306B (en) | 2023-06-30 |
Family
ID=67914967
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910424733.0A Active CN110264306B (en) | 2019-05-21 | 2019-05-21 | Big data-based product recommendation method, device, server and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110264306B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110852785B (en) * | 2019-10-12 | 2023-11-21 | 中国平安人寿保险股份有限公司 | User grading method, device and computer readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017041541A1 (en) * | 2015-09-08 | 2017-03-16 | 北京邮电大学 | Method for pushing recommendation information, and server and storage medium |
| CN108550078A (en) * | 2018-05-09 | 2018-09-18 | 平安普惠企业管理有限公司 | The recommendation method and terminal device of loan product |
| CN108681970A (en) * | 2018-05-28 | 2018-10-19 | 深圳市零度智控科技有限公司 | Finance product method for pushing, system and computer storage media based on big data |
| CN108734587A (en) * | 2018-05-22 | 2018-11-02 | 深圳壹账通智能科技有限公司 | The recommendation method and terminal device of financial product |
| CN108804564A (en) * | 2018-05-22 | 2018-11-13 | 深圳壹账通智能科技有限公司 | The combined recommendation method and terminal device of financial product |
-
2019
- 2019-05-21 CN CN201910424733.0A patent/CN110264306B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017041541A1 (en) * | 2015-09-08 | 2017-03-16 | 北京邮电大学 | Method for pushing recommendation information, and server and storage medium |
| CN108550078A (en) * | 2018-05-09 | 2018-09-18 | 平安普惠企业管理有限公司 | The recommendation method and terminal device of loan product |
| CN108734587A (en) * | 2018-05-22 | 2018-11-02 | 深圳壹账通智能科技有限公司 | The recommendation method and terminal device of financial product |
| CN108804564A (en) * | 2018-05-22 | 2018-11-13 | 深圳壹账通智能科技有限公司 | The combined recommendation method and terminal device of financial product |
| CN108681970A (en) * | 2018-05-28 | 2018-10-19 | 深圳市零度智控科技有限公司 | Finance product method for pushing, system and computer storage media based on big data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110264306A (en) | 2019-09-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109858740B (en) | Enterprise risk assessment method and device, computer equipment and storage medium | |
| AU2021232839B2 (en) | Updating Attribute Data Structures to Indicate Trends in Attribute Data Provided to Automated Modelling Systems | |
| WO2019214248A1 (en) | Risk assessment method and apparatus, terminal device, and storage medium | |
| WO2019205325A1 (en) | Method for determining risk level of user, terminal device, and computer-readable storage medium | |
| CN109242612B (en) | Product recommendation method and device | |
| WO2019061976A1 (en) | Fund product recommendation method and apparatus, terminal device, and storage medium | |
| CN108683530A (en) | Data analysing method, device and the storage medium of multi-dimensional data | |
| CN109299356B (en) | Activity recommendation method and device based on big data, electronic equipment and storage medium | |
| EP3961384A1 (en) | Automatic derivation of software engineering artifact attributes from product or service development concepts | |
| JP2018536937A (en) | Push information coarse selection sorting method, device, and computer storage medium | |
| CN109325818B (en) | Product recommendation method, computer readable storage medium and terminal equipment | |
| CN111340244A (en) | Prediction method, training method, device, server and medium | |
| WO2023000491A1 (en) | Application recommendation method, apparatus and device, and computer-readable storage medium | |
| CN112116245A (en) | Credit risk assessment method, credit risk assessment device, computer equipment and storage medium | |
| CN110750658A (en) | Recommendation method of media resource, server and computer readable storage medium | |
| CN112184046A (en) | Advertisement service user value evaluation method, device, equipment and storage medium | |
| CN112396211A (en) | Data prediction method, device, equipment and computer storage medium | |
| CN110866698A (en) | Device for assessing service score of service provider | |
| CN114782201A (en) | Stock recommendation method and device, computer equipment and storage medium | |
| CN110264306B (en) | Big data-based product recommendation method, device, server and medium | |
| CN116737373A (en) | Load balancing method, device, computer equipment and storage medium | |
| CN109460778B (en) | Activity evaluation method, activity evaluation device, electronic equipment and storage medium | |
| CN110717653A (en) | Risk identification method and device and electronic equipment | |
| CN109472454B (en) | Activity evaluation method, activity evaluation device, electronic equipment and storage medium | |
| CN112950225A (en) | Customer category determination method, device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |