CN115375484A - Matrix decomposition-based insurance product extraction method and device, equipment and medium - Google Patents
Matrix decomposition-based insurance product extraction method and device, equipment and medium Download PDFInfo
- Publication number
- CN115375484A CN115375484A CN202210933439.4A CN202210933439A CN115375484A CN 115375484 A CN115375484 A CN 115375484A CN 202210933439 A CN202210933439 A CN 202210933439A CN 115375484 A CN115375484 A CN 115375484A
- Authority
- CN
- China
- Prior art keywords
- insurance product
- matrix
- insurance
- user
- scoring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000011159 matrix material Substances 0.000 title claims abstract description 198
- 238000000605 extraction Methods 0.000 title claims abstract description 84
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 35
- 230000006399 behavior Effects 0.000 claims description 53
- 230000006870 function Effects 0.000 claims description 27
- 230000015654 memory Effects 0.000 claims description 22
- 238000012163 sequencing technique Methods 0.000 claims description 21
- 238000004891 communication Methods 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000013473 artificial intelligence Methods 0.000 abstract description 8
- 239000000047 product Substances 0.000 description 353
- 238000005516 engineering process Methods 0.000 description 11
- 239000000284 extract Substances 0.000 description 9
- 238000012545 processing Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000000670 limiting effect Effects 0.000 description 2
- 230000002829 reductive effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 239000012466 permeate Substances 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/06—Asset management; Financial planning or analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Development Economics (AREA)
- Technology Law (AREA)
- General Engineering & Computer Science (AREA)
- Operations Research (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Algebra (AREA)
- Entrepreneurship & Innovation (AREA)
- Computing Systems (AREA)
- Game Theory and Decision Science (AREA)
- Human Resources & Organizations (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The embodiment of the application provides an insurance product extraction method and device based on matrix decomposition, electronic equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring historical click behavior data of a plurality of users on each insurance product; generating a user-insurance product scoring matrix according to historical click behavior data; calculating a scoring result for each of the insurance products in a user-insurance product scoring matrix; extracting insurance products according to the sorting result of the grading result; or determining the scoring result of each insurance product in the user-insurance product scoring matrix as an insurance product vector; determining a target insurance product; calculating cosine similarity between each insurance product and the target insurance product; and extracting the insurance products according to the sorting result of the cosine similarity. Compared with the traditional extraction method, the insurance products extracted by the embodiment of the application are richer and more various, and the insurance products which are interested by the user can be extracted.
Description
Technical Field
The application relates to the technical field of artificial intelligence, in particular to an insurance product extraction method and device based on matrix decomposition, electronic equipment and a storage medium.
Background
In recent years, the rapid development of machine learning unconsciously permeates all aspects of life of people such as clothes, eating, housing and the like, for example, various delicious foods are taken out on various take-out platforms for selection by users, various clothes, commodities and the like are taken out on shopping websites for browsing and selecting by users, the life rules of people are mastered by machine learning through mining big data, and meanwhile, the life habits of people are also changed slowly. At the same time, the insurance industry is beginning to undergo significant transformation in this regard. On one hand, the consumer behavior starts to change fundamentally, and more young people start to interact with various industries through the online mobile internet; on the other hand, the insurance industry is also developing on the channel side, the internet channel is more and more active, and the digital platform is replacing the traditional channel to carry out insurance recommendation service.
Currently, an online insurance product recommendation system, which extracts the logic of insurance products, is based on the personal basic attribute information provided by user registration, and is based on the insurance products purchased by the user history, so that the decision time of the user can be reduced to a certain extent. However, the insurance product records purchased based on the user history are extracted, so that the insurance products selected by the user can be extracted insufficiently, and the potential real requirements of the user cannot be met.
Disclosure of Invention
The embodiment of the application mainly aims to provide an insurance product extraction method and device based on matrix decomposition, electronic equipment and a storage medium, wherein all insurance products possibly liked by users are considered from the global perspective, all insurance products interested by different users can be covered, and the extracted insurance products are richer and more diverse than the traditional extraction method.
In order to achieve the above object, a first aspect of an embodiment of the present application provides a method for extracting an insurance product based on matrix decomposition, including:
acquiring historical click behavior data of a plurality of users on each insurance product;
generating a user-insurance product scoring matrix according to the historical click behavior data;
extracting the insurance product according to the user-insurance product scoring matrix by at least one of:
calculating a scoring result for each of the insurance products in the user-insurance product scoring matrix;
ranking the grading results corresponding to the insurance products, and extracting the insurance products according to the ranking results of the grading results;
or,
determining the scoring result of each of the insurance products in the user-insurance product scoring matrix as an insurance product vector;
determining a target insurance product according to the historical click behavior data;
calculating cosine similarity between each insurance product and the target insurance product;
and sequencing the cosine similarity corresponding to each insurance product, and extracting the insurance products according to the sequencing result of the cosine similarity.
In some embodiments, after obtaining the historical click behavior data of the user on the insurance product, the method further comprises:
and intercepting data of which the clicking times exceed a first threshold value in the historical clicking behavior data within a preset time period so as to retain the historical clicking behavior data of X insurance products clicked within the preset time period, wherein X is a positive integer.
In some embodiments, said step of calculating the scoring result for each of said insurance products in said user-insurance product scoring matrix comprises:
decomposing the user-insurance product scoring matrix into a product of a first matrix and a second matrix, and constructing a third matrix;
(iii) taking the square of the error of the user-insurance product scoring matrix and the third matrix as a loss function;
solving the first matrix and the second matrix by using an alternating least square method according to the loss function;
and determining the scoring result of each insurance product according to the values of the first matrix and the second matrix obtained by solving.
In some embodiments, the loss function is expressed as:
wherein,
in the formula,
square of error, R, representing user-insurance product scoring matrix and said third matrix m,n A user-insurance product scoring matrix is represented,
denotes a third matrix, P m,k Representing a first matrix, Q k,m Representing a second matrix, m representing the number of users, n representing the number of insurance products, and K representing the spatial dimension.
In some embodiments, said step of solving said first and second matrices using alternating least squares based on said loss function comprises:
initializing values of the first matrix;
solving the second matrix by minimizing the loss function according to the values of the first matrix;
solving the first matrix by minimizing the loss function according to the solved value of the second matrix;
and returning to execute the step of solving the second matrix by minimizing the loss function according to the value of the first matrix until the first matrix and the second matrix converge to have errors meeting a threshold condition.
In some embodiments, the step of ranking the scoring results corresponding to each insurance product and extracting the insurance products according to the ranking results of the scoring results includes:
sorting the scoring results corresponding to the insurance products according to the numerical value from large to small to obtain a first sorting list;
selecting the first N insurance products in the first ranking list for extraction;
or,
sorting the scoring results corresponding to the insurance products according to the numerical value from small to large to obtain a second sorting list;
and selecting the next N insurance products in the second sorted list for extraction, wherein N is a positive integer.
In some embodiments, the step of sorting the cosine similarity corresponding to each insurance product and extracting the insurance products according to the sorting result of the cosine similarity includes:
sorting the cosine similarity corresponding to each insurance product according to the numerical value from large to small to obtain a third sorting list;
selecting the top M insurance products in the third sorted list for extraction;
or,
sorting the cosine similarity corresponding to each insurance product according to the numerical value from small to large to obtain a fourth sorting list;
and selecting the back M insurance products in the fourth sorted list for extraction, wherein M is a positive integer.
To achieve the above object, a second aspect of the embodiments of the present application proposes an insurance product extraction apparatus based on matrix factorization, the apparatus including:
the acquisition module is used for acquiring historical click behavior data of a plurality of users on each insurance product;
the generating module is used for generating a user-insurance product scoring matrix according to the historical click behavior data;
an extraction module for extracting the insurance product according to the user-insurance product scoring matrix by at least one of:
a first calculation unit for calculating a scoring result for each of the insurance products in the user-insurance product scoring matrix;
the first sorting and extracting unit is used for sorting the scoring results corresponding to the insurance products and extracting the insurance products according to the sorting results of the scoring results;
or,
a first determining unit, configured to determine a scoring result of each of the insurance products in the user-insurance product scoring matrix as an insurance product vector;
the second determining unit is used for determining a target insurance product according to the historical click behavior data;
the second calculation unit is used for calculating cosine similarity between each insurance product and the target insurance product;
and the second sequencing and extracting unit is used for sequencing the cosine similarity corresponding to each insurance product and extracting the insurance products according to the sequencing result of the cosine similarity.
In order to achieve the above object, a third aspect of the embodiments of the present application provides an electronic device, which includes a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for implementing connection communication between the processor and the memory, wherein the program, when executed by the processor, implements the method of the first aspect.
To achieve the above object, a fourth aspect of the embodiments of the present application proposes a storage medium, which is a computer-readable storage medium for computer-readable storage, and stores one or more programs, which are executable by one or more processors to implement the method of the first aspect.
According to the insurance product extraction method and device based on matrix decomposition, the electronic equipment and the storage medium, historical click behavior data of a plurality of users on each insurance product are obtained, then a user-insurance product scoring matrix is generated according to the historical click behavior data, and then the insurance products are extracted in at least one mode of 2 modes according to the user-insurance product scoring matrix. The first way is to calculate the scoring results for each insurance product in the user-insurance product scoring matrix, then rank the scoring results, and extract the insurance products according to the ranked results of the scoring results. In the extraction mode, the user-insurance product scoring matrix covers the scores of all users on all insurance products, the insurance products which are probably liked by all users are considered from the global perspective, all the insurance products which are interested by different users can be covered, and the extracted insurance products are richer and more diverse than the traditional extraction method. The second mode is that the scoring result of each insurance product in the user-insurance product scoring matrix is determined as an insurance product vector, then a target insurance product is determined, and the cosine similarity between each insurance product and the target insurance product is calculated; and sequencing the cosine similarity, and extracting the insurance products according to the sequencing result of the cosine similarity. In the extraction mode, once the target insurance product clicked by the user is detected, other insurance products most similar to the target insurance product can be extracted for the user to compare, refer and select, so that the method is more targeted, and meanwhile, the user can be assisted to make a selection.
Drawings
FIG. 1 is a schematic diagram of an implementation environment of an insurance product extraction method based on matrix factorization provided in an embodiment of the present application;
FIG. 2 is a flow chart of a matrix factorization based insurance product extraction method provided by an embodiment of the present application;
FIG. 3 is a flowchart of step S204 in FIG. 2;
fig. 4 is a flowchart of step S303 in fig. 3;
fig. 5 is a flowchart of step S205 in fig. 2;
fig. 6 is a flowchart of step S209 in fig. 2;
FIG. 7 is a schematic structural diagram of an insurance product extraction device based on matrix decomposition according to an embodiment of the present application;
fig. 8 is a schematic hardware structure diagram of an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
It is noted that while functional block divisions are provided in device diagrams and logical sequences are shown in flowcharts, in some cases, steps shown or described may be performed in sequences other than block divisions within devices or flowcharts. The terms first, second and the like in the description and in the claims, and the drawings described above, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.
Currently, the development of the internet profoundly changes the life style of people, and the software platform can selectively extract insurance products according to personal basic attribute information of users or record information of the insurance products purchased by users in history, so that the time for searching the insurance products by the users can be shortened to a certain extent, and more suitable insurance products can be extracted for the users. However, at the same time, the product extraction is inaccurate or repeated, which may cause browsing fatigue of the user to some extent and may cause a sense of discomfort to the user.
In the related art, the online insurance product extraction method generally extracts insurance products according to insurance product records purchased by a user in history, for example, by obtaining the insurance products purchased by the user in history, determining the type of the insurance product that the user is interested in, and then extracting the insurance products with similar types for the user according to the type of the insurance product. However, the insurance products extracted by the extraction mode are single, namely, the insurance products selected by the user can be extracted insufficiently; and for users who have not purchased insurance products, suitable insurance products cannot be extracted for the users to select.
In view of this, an embodiment of the present application provides an insurance product extraction method based on matrix decomposition, which includes obtaining historical click behavior data of multiple users on each insurance product, generating a user-insurance product scoring matrix according to the historical click behavior data, and extracting the insurance product through at least one of 2 ways according to the user-insurance product scoring matrix. The first way is to calculate the scoring results for each insurance product in the user-insurance product scoring matrix, then sort the scoring results, and extract the insurance products according to the sorted results of the scoring results. In the extraction mode, the user-insurance product scoring matrix covers the scores of all the insurance products of all the users, the insurance products which are probably liked by all the users are considered from the global perspective, all the insurance products which are interested by different users can be covered, and the extracted insurance products are richer and more diverse than the traditional extraction method. The second mode is that the scoring result of each insurance product in the user-insurance product scoring matrix is determined as an insurance product vector, then a target insurance product is determined, and the cosine similarity between each insurance product and the target insurance product is calculated; and sorting the cosine similarity, and extracting the insurance products according to the sorting result of the cosine similarity. In the extraction mode, once the target insurance product clicked by the user is detected, other insurance products most similar to the target insurance product can be extracted for the user to compare, refer and select, so that the method is more targeted, and meanwhile, the user can be assisted to make a selection.
The matrix decomposition-based insurance product extraction method and apparatus, the electronic device, and the storage medium provided in the embodiments of the present application are specifically described in the following embodiments, and first, the matrix decomposition-based insurance product extraction method in the embodiments of the present application is described.
The embodiment of the application can acquire and process related data based on an artificial intelligence technology. Among them, artificial Intelligence (AI) is a theory, method, technique and application system that simulates, extends and expands human Intelligence using a digital computer or a machine controlled by a digital computer, senses the environment, acquires knowledge and uses the knowledge to obtain the best result.
The artificial intelligence base technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a robot technology, a biological recognition technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.
The embodiment of the application provides an insurance product extraction method based on matrix decomposition, and relates to the technical field of artificial intelligence. The method can be applied to a terminal, a server side and software running in the terminal or the server side. In some embodiments, the terminal may be a smartphone, tablet, laptop, desktop computer, or the like; the server side can be configured as an independent physical server, can also be configured as a server cluster or a distributed system formed by a plurality of physical servers, and can also be configured as a cloud server for providing basic cloud computing services such as cloud service, a cloud database, cloud computing, cloud functions, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN (content distribution network) and big data and artificial intelligence platforms; the software may be an application or the like implementing a matrix factorization based insurance product extraction method, but is not limited to the above form.
The application is operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like. The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
In each embodiment of the present application, when data related to the user identity or characteristic, such as user information, user behavior data, user history data, and user location information, is processed, permission or consent of the user is obtained, and the data collection, use, and processing comply with relevant laws and regulations and standards of relevant countries and regions. In addition, when the embodiment of the present application needs to acquire sensitive personal information of a user, individual permission or individual consent of the user is obtained through a pop-up window or a jump to a confirmation page, and after the individual permission or individual consent of the user is definitely obtained, necessary user-related data for enabling the embodiment of the present application to operate normally is acquired.
The following describes in detail specific embodiments of the present application with reference to the drawings.
Referring to fig. 1, fig. 1 is a schematic diagram illustrating an implementation environment of the matrix factorization-based insurance product extraction method provided in the embodiment of the present application. In this implementation environment, the main software and hardware entities involved include a first user terminal 110, a server 120 and several second user terminals 130. The first user terminal 110 and the second user terminal 130 may run thereon related software or a network platform having insurance product browsing and purchasing functions, for example, the software or the network platform may be an online insurance product purchasing application, a web page, or an applet running depending on a host program. In some embodiments, the first user terminal 110 and the second user terminal 130 may be any one of a smart watch, a smart phone, a tablet computer, a Personal Digital Assistant (PDA), a laptop computer, or a desktop computer.
The server 120 may be a background server corresponding to the aforementioned software or network platform, and is mainly used for executing the matrix decomposition-based insurance product extraction method in the embodiment of the present application. In some embodiments, the server 120 may be configured as an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, and a cloud server providing services such as cloud services, cloud databases, cloud computing, cloud storage, and network services. In some embodiments, the communication connection between the first user terminal 110 and the server 120, and between the server 120 and the second user terminal 130 may be established through a wireless network or a wired network. The wireless or wireline networks may be implemented using standard communication technologies and/or protocols, and may be configured as the internet or any other Network including, but not limited to, a Local Area Network (LAN), a Metropolitan Area Network (MAN), a Wide Area Network (WAN), any combination of mobile, wireline or wireless networks, private or virtual private networks, for example.
In the implementation environment shown in fig. 1, the server 120 may obtain historical click behavior data of a plurality of users for each insurance product through the first user terminal 110 or the second user terminal 130. Then, the server 120 generates a user-insurance product scoring matrix according to the acquired historical click behavior data, calculates a scoring result of each insurance product in the user-insurance product scoring matrix, sorts the scoring results, and extracts the insurance products according to the sorting results of the scoring results. Or determining the scoring result as an insurance product vector, determining a target insurance product, calculating cosine similarity between each insurance product and the target insurance product, sequencing the cosine similarity, and extracting the insurance products according to the sequencing result of the cosine similarity. Of course, it is understood that the implementation environment in fig. 1 is only an optional application scenario of the matrix decomposition-based insurance product extraction method provided in the embodiment of the present application, and the actual application is not fixed to the software and hardware environment shown in fig. 1.
Referring to fig. 2, fig. 2 is an alternative flowchart of the matrix decomposition-based insurance product extraction method provided in the embodiment of the present application, and the method in fig. 2 may include, but is not limited to, steps S201 to S209.
Step S201, obtaining historical click behavior data of a plurality of users on each insurance product;
step S202, generating a user-insurance product scoring matrix according to historical click behavior data;
step S203, extracting the insurance products by at least one of the following modes according to the user-insurance product scoring matrix:
step S204, calculating the scoring result of each insurance product in the user-insurance product scoring matrix;
step S205, sorting the grading results corresponding to each insurance product, and extracting the insurance products according to the sorting results of the grading results;
step S206, determining the scoring result of each insurance product in the user-insurance product scoring matrix as an insurance product vector;
step S207, determining a target insurance product according to historical click behavior data;
step S208, calculating cosine similarity between each insurance product and a target insurance product;
step S209, the cosine similarity corresponding to each insurance product is sorted, and the insurance products are extracted according to the sorting result of the cosine similarity.
In steps S201 to S209 illustrated in the embodiment of the present application, historical click behavior data of a plurality of users on each insurance product is obtained, then a user-insurance product scoring matrix is generated according to the historical click behavior data, and then the insurance product is further extracted according to the generated user-insurance product scoring matrix. The embodiment of the application provides 2 extraction modes, and corresponding extraction can be carried out according to actual conditions. The first way is to calculate the scoring results for each insurance product in the user-insurance product scoring matrix and then extract the insurance products according to the ranking of the scoring results. The extraction mode is suitable for users who do not purchase the insurance products, namely, the users who do not purchase the insurance products can extract the insurance products which are relatively interested by the public, and the extraction accuracy can be ensured to a certain extent. The second mode is that the scoring result is used as an insurance product vector, then a target insurance product is determined, and the cosine similarity between each insurance product and the target insurance product is calculated; and sequencing the cosine similarity, and extracting the insurance products according to the sequencing result of the cosine similarity. The extraction method is suitable for extracting similar insurance products for the user to select when the user clicks a certain insurance product. For example, it is detected that the insurance product a is clicked, and at this time, the insurance product a is a target insurance product, the ranking result of the cosine similarity between each insurance product and the insurance product a is inquired, and other insurance products similar to the insurance product a are extracted according to the ranking result. The extraction mode has pertinence, and the insurance products required by the user can be extracted more accurately according to the interest degree of the user in the insurance products.
Regarding step S201, historical click behavior data of a plurality of users in recent 3 months may be obtained from the user behavior buried point table, and considering that the purchase frequency of insurance products is relatively low compared with that of other e-commerce products, most users may purchase only one product, so the historical click behavior data of the plurality of users may be properly extended to recent 6 months. Of course, it may be understood that obtaining the historical click behavior data of a plurality of users from the user behavior buried point table is only an exemplary obtaining manner, and other manners capable of obtaining the historical click behavior data may be used as long as the historical click behavior data can be obtained, and the obtaining manner is not specifically limited in the embodiment of the present application. Likewise, the time period of the historical click behavior data is not particularly limited, and may be 3 months, 4 months, 6 months or even longer, and the specific time period may be determined according to actual conditions.
After step S201 of some embodiments, the method further comprises step S201-1:
step S201-1, performing truncation processing on data, of which the number of clicks exceeds a first threshold value in a preset time period, in the historical click behavior data so as to retain the historical click behavior data of X insurance products clicked in the preset time period, wherein X is a positive integer.
Specifically, considering that there is data with too dense partial clicks in the obtained historical click behavior data, for example, data with the number of clicks exceeding 100 in the last 1 month, only historical click behavior data of, for example, 100 insurance products clicked in the last 1 month is retained, and data in other time periods is directly truncated and discarded. The processing mode can accelerate the processing speed of the data without influencing the accuracy of the acquired data.
With respect to step S202, after obtaining historical click behavior data, a corresponding user-insurance product scoring matrix may further be generated. For example, n users and m insurance products are obtained through historical click behavior data, and click and purchase behaviors of all the users on the insurance products can be mapped to an n x m matrix, namely, a user-insurance product scoring matrix R is generated m,n . For example, a score of 1 may be set for insurance products that the user has clicked on, and a score of 5 may be set for insurance products that have been purchased.
In the embodiment of the present application, after the user-insurance product scoring matrix is generated, the insurance products can be extracted by at least one of the following methods:
(1) And extracting according to the sorting result of the grading result, wherein the steps S204-S205 are included.
Referring to fig. 3, in some embodiments, step S204 may include, but is not limited to including, steps S301 to S304:
step S301, decomposing a user-insurance product scoring matrix into a product of a first matrix and a second matrix, and constructing a third matrix;
step S302, taking the square of the error of the user-insurance product scoring matrix and the third matrix as a loss function;
step S303, solving a first matrix and a second matrix by using an alternating least square method according to the loss function;
and step S304, determining the scoring result of each insurance product according to the values of the first matrix and the second matrix obtained by solving.
In this embodiment, considering that insurance products are various, the number of insurance products that can be touched by the user is very small, that is, the insurance products that the user has interacted only occupy a small part of the products in the mall, so the generated user-insurance product scoring matrix R m,n Is a high-dimensional sparse matrix. According to matrix theory in mathematics, the matrix R can be expressed m,n Approximately decomposed into the product of two matrixes, and the decomposition process can reduce the original matrix R as much as possible m,n I.e. by
Wherein the first matrix P m,k Can represent a user matrix, a second matrix Q k,m Can represent insurance product matrix, wherein K represents hidden vector space dimension, mapping clicked relation of user and insurance product to potential space of K dimension, in first matrix P m,k In the second matrix Q, K dimensions are each potentially represented as a behavior feature of the user, such as the user's preference when selecting a product, and the like k,m In (e), the K dimensions each potentially represent characteristics of the insurance product, such as product planogram characteristics, premium size range, and the like. It can be understood that user attribute information and implicit feedback, temporal context, etc. can be added to the user behavior feature vector and the insurance product feature vector subsequently to enhance the matrix prediction effect. The embodiment solves the problem of each element value in the matrix and transformsTo solve the regression problem in machine learning, a user-insurance product scoring matrix R is used m,n And a constructed third matrix
As a function of the loss. Then solving the first matrix P by an alternating least square method m,k And a second matrix Q k,m . Finally, obtaining a first matrix P according to the solution m,k And a second matrix Q k,m To determine the scoring result for each insurance product.
In this embodiment, the expression of the loss function is:
wherein,
in the formula,
square of error, R, representing user-insurance product scoring matrix and said third matrix m,n A user-insurance product scoring matrix is represented,
denotes a third matrix, P m,k Representing a first matrix, Q k,m Representing a second matrix, m representing the number of users, n representing the number of insurance products, and K representing the spatial dimension.
Referring to fig. 4, in some embodiments, step S303 may include, but is not limited to including, step S401 to step S404:
step S401, initializing the value of a first matrix;
step S402, solving a second matrix by minimizing the loss function according to the value of the first matrix;
step S403, solving a first matrix through a minimum loss function according to the value of the second matrix obtained through solving;
and S404, returning to execute the step of solving the second matrix through a minimum loss function according to the value of the first matrix until the first matrix and the second matrix converge to the error meeting the threshold condition.
This embodiment solves the first matrix P by the alternating least squares method according to the loss function m,k And a second matrix Q k,m . Specifically, the first matrix P is initialized first m,k Then solving the second matrix Q by minimizing a loss function k,m Then fix the solved second matrix Q k,m By minimizing a loss function to solve a first matrix P m,k Iteratively repeating the above steps until the first matrix P is reached m,k And a second matrix Q k,m Convergence is made until the error satisfies the threshold condition.
In the embodiment of the application, the potential factors liked by the user to the insurance products are considered in the matrix decomposition calculation process, the characteristics of the insurance products are considered, and the relation between deeper users and the insurance products can be mined by combining the characteristics of hidden semantics and machine learning. Meanwhile, the high-dimensional matrix is mapped into two low-dimensional matrices, so that the calculation complexity is reduced, and the method is easier to realize through programming.
Referring to FIG. 5, in some embodiments, after calculating the scoring result for each insurance product in the user-insurance product scoring matrix through step S204, step S205 is further performed, and step S205 may include, but is not limited to, steps S501 to S504:
s501, sorting the scoring results corresponding to the insurance products according to the numerical value from large to small to obtain a first sorting list;
step S502, selecting the first N insurance products in the first ranking list for extraction;
or,
step S503, sorting the scoring results corresponding to the insurance products according to the numerical value from small to large to obtain a second sorting list;
and step S504, selecting the next N insurance products in the second sorted list for extraction, wherein N is a positive integer.
In this embodiment, since the user-insurance product scoring matrix covers the scores of all users on all insurance products, insurance products of interest to the public can be extracted from the users. Meanwhile, in the extraction process, if the user purchases insurance products, the insurance products purchased by the user can be filtered firstly and then extracted, so that the problem of poor user experience caused by repeated extraction can be avoided, and the extraction effect can be improved.
(2) And extracting according to the sorting result of the cosine similarity, comprising steps S206-S209.
In the embodiment of the application, the scoring result of each insurance product in the user-insurance product scoring matrix can be determined as an insurance product vector, then the target insurance product is determined according to historical click behavior data, and the cosine similarity between each insurance product and the target insurance product is calculated, so that the similarity between each insurance product and the target insurance product can be obtained. It is to be understood that the target insurance product may be any one of all insurance products. For example, the user-insurance product scoring matrix has 100 different insurance products in total, which are respectively numbered 1-100, if the insurance product numbered 1 is taken as the target insurance product, the cosine similarity between the other 99 insurance products and the insurance product numbered 1 can be calculated, similarly, if the insurance product numbered 2 is taken as the target insurance product, the cosine similarity between the other 99 insurance products and the insurance product numbered 2 can be calculated, and by analogy, the cosine similarity between each insurance product and the other insurance products can be calculated, and each insurance product has a corresponding cosine similarity ranking list. For example, the insurance product numbered 1 has its corresponding sorted list X, and the insurance product numbered 2 has its corresponding sorted list Y, where the sorted list X reflects the degree of similarity between the other 99 insurance products and the insurance product numbered 1, and the sorted list Y reflects the degree of similarity between the other 99 insurance products and the insurance product numbered 2.
Referring to fig. 6, in some embodiments, after the cosine similarity between each insurance product and the target insurance product is calculated in step 208, step S209 is further performed, and step S209 may include, but is not limited to, steps S601 to S604:
s601, sorting the cosine similarity corresponding to each insurance product according to the numerical value from large to small to obtain a third sorting list;
step S602, selecting the top M insurance products in the third sorted list for extraction;
or,
step S603, sorting the cosine similarity corresponding to each insurance product according to the numerical value from small to large to obtain a fourth sorting list;
and step S604, selecting the next M insurance products in the fourth sorted list for extraction, wherein M is a positive integer.
In this embodiment, after the cosine similarity corresponding to each insurance product is obtained, the cosine similarity values may be sorted in the order from large to small or from small to large, so as to obtain a cosine similarity sorted list corresponding to each insurance product. The extraction method is suitable for the situation that a user clicks a certain insurance product, for example, the insurance product A is clicked, at the moment, the cosine similarity ranking list corresponding to the insurance product A is inquired, and the insurance products similar to the insurance product A in the list can be extracted. By the extraction method, insurance products required by the user can be extracted from the user in a targeted manner, and the extraction effect is better compared with that of the traditional extraction method.
It can be understood that, according to the embodiment of the present application, one corresponding extraction manner may be selected from the 2 extraction manners to extract the insurance product, the 2 extraction manners may be used to extract the insurance product at the same time, and the 2 extraction manners may be combined to extract the insurance product. For example, when it is detected that the user logs in a purchase webpage of the insurance product for the first time through the software platform to browse, a mode of extracting according to the ranking result of the scoring result can be selected, and the insurance product with the top 10 of the popular interest degree ranking can be extracted from the user. Or when detecting that the user clicks a certain insurance product or detecting that the user purchases a certain insurance product, selecting a mode of extracting according to the sorting result of the cosine similarity, and extracting the insurance product which the user may be interested in from the user. Or after the user logs in the webpage through the corresponding software platform, the mode of extracting according to the ranking result of the scoring result can be selected firstly, and the first few insurance products which are interested by the public are extracted from the user. And when detecting that the user clicks one insurance product, further selecting a mode of extracting according to the sorting result of the cosine similarity, and extracting the first few insurance products similar to the clicked insurance product from the user. Or when detecting that a user logs in a purchase webpage of an insurance product through a software platform to browse, firstly selecting a mode of extracting according to the sorting result of the grading result to generate an insurance product extraction list which is interested by the public and marked as a list 1; then, detecting that a certain insurance product is clicked, selecting a mode of extracting according to the sorting result of cosine similarity, generating an insurance product extraction list similar to the clicked insurance product, and recording the insurance product extraction list as a list 2; and combining the list 1 with the list 2 to extract insurance products which are similar to the clicked insurance products and are interested by the public to the user.
Referring to fig. 7, an embodiment of the present application further provides an insurance product extraction apparatus based on matrix decomposition, which can implement the insurance product extraction method based on matrix decomposition, and the apparatus includes:
the acquisition module is used for acquiring historical click behavior data of a plurality of users on each insurance product;
the generating module is used for generating a user-insurance product scoring matrix according to historical click behavior data;
an extraction module for extracting the insurance product according to the user-insurance product scoring matrix by at least one of:
a first calculation unit for calculating a scoring result for each of the insurance products in a user-insurance product scoring matrix;
the first sorting and extracting unit is used for sorting the grading results corresponding to the insurance products and extracting the insurance products according to the sorting results of the grading results;
or,
a first determination unit, configured to determine a scoring result of each insurance product in the user-insurance product scoring matrix as an insurance product vector;
the second determining unit is used for determining a target insurance product according to historical click behavior data;
the second calculating unit is used for calculating cosine similarity between each insurance product and the target insurance product;
and the second sequencing and extracting unit is used for sequencing the cosine similarity corresponding to each insurance product and extracting the insurance products according to the sequencing result of the cosine similarity.
The specific implementation of the insurance product device based on matrix decomposition is substantially the same as the specific embodiment of the insurance product extraction method based on matrix decomposition, and is not described herein again.
An embodiment of the present application further provides an electronic device, where the electronic device includes: the insurance product extraction method based on matrix decomposition comprises a memory, a processor, a program stored on the memory and capable of running on the processor, and a data bus for realizing connection communication between the processor and the memory, wherein when the program is executed by the processor, the insurance product extraction method based on matrix decomposition is realized. The electronic equipment can be any intelligent terminal including a tablet computer, a vehicle-mounted computer and the like.
Referring to fig. 8, fig. 8 illustrates a hardware structure of an electronic device according to another embodiment, where the electronic device includes:
the processor 801 may be implemented by a general-purpose CPU (central processing unit), a microprocessor, an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits, and is configured to execute a relevant program to implement the technical solution provided in the embodiment of the present application;
the memory 802 may be implemented in a form of a Read Only Memory (ROM), a static storage device, a dynamic storage device, or a Random Access Memory (RAM). The memory 802 may store an operating system and other application programs, and when the technical solution provided by the embodiments of the present disclosure is implemented by software or firmware, the relevant program codes are stored in the memory 802, and the processor 801 calls the matrix decomposition-based insurance product extraction method for implementing the embodiments of the present disclosure;
an input/output interface 803 for realizing input and output of information;
the communication interface 804 is configured to implement communication interaction between the device and another device, and may implement communication in a wired manner (e.g., USB, network cable, etc.) or in a wireless manner (e.g., mobile network, WIFI, bluetooth, etc.);
a bus 805 that transfers information between the various components of the device (e.g., the processor 801, memory 802, input/output interface 803, and communications interface 804);
wherein the processor 801, the memory 802, the input/output interface 803 and the communication interface 804 are communicatively connected to each other within the device via a bus 805.
The embodiment of the present application further provides a storage medium, which is a computer-readable storage medium for computer-readable storage, and the storage medium stores one or more programs, and the one or more programs are executable by one or more processors to implement the method for extracting insurance products based on matrix decomposition.
The memory, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs as well as non-transitory computer executable programs. Further, the memory may include high speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory optionally includes memory located remotely from the processor, and these remote memories may be connected to the processor through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
According to the extraction method, the extraction device, the electronic equipment and the storage medium of the insurance product based on matrix decomposition, the user-insurance product scoring matrix is generated through historical click behavior data, and the insurance product is extracted through at least one mode of 2 modes according to the user-insurance product scoring matrix. The first way is to calculate the scoring results for each insurance product in the user-insurance product scoring matrix, then sort the scoring results, and extract the insurance products according to the sorted results of the scoring results. In the extraction mode, the user-insurance product scoring matrix covers the scores of all the insurance products of all the users, the insurance products which are probably liked by all the users are considered from the global perspective, all the insurance products which are interested by different users can be covered, and the extracted insurance products are richer and more diverse than the traditional extraction method. The second mode is that the scoring result of each insurance product in the user-insurance product scoring matrix is determined as an insurance product vector, then a target insurance product is determined, and the cosine similarity between each insurance product and the target insurance product is calculated; and sequencing the cosine similarity, and extracting the insurance products according to the sequencing result of the cosine similarity. In the extraction mode, once the target insurance product clicked by the user is detected, other insurance products most similar to the target insurance product can be extracted for the user to compare, refer and select, and the method is more pertinent and can assist the user in making a selection.
The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation to the technical solutions provided in the embodiments of the present application, and it is obvious to those skilled in the art that the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems with the evolution of technology and the emergence of new application scenarios.
It will be appreciated by those skilled in the art that the embodiments shown in fig. 2-6 are not limiting of the embodiments of the present application and may include more or fewer steps than those shown, or some of the steps may be combined, or different steps may be included.
The above-described embodiments of the apparatus are merely illustrative, wherein the units illustrated as separate components may or may not be physically separate, i.e. may be located in one place, or may also be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
One of ordinary skill in the art will appreciate that all or some of the steps of the methods, systems, functional modules/units in the devices disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof.
The terms "first," "second," "third," "fourth," and the like in the description of the application and the above-described figures, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It should be understood that in the present application, "at least one" means one or more, "a plurality" means two or more. "and/or" for describing an association relationship of associated objects, indicating that there may be three relationships, e.g., "a and/or B" may indicate: only A, only B and both A and B are present, wherein A and B may be singular or plural. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. "at least one of the following" or similar expressions refer to any combination of these items, including any combination of single item(s) or plural items. For example, at least one (one) of a, b, or c, may represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, c may be single or plural.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the above-described division of units is only one type of division of logical functions, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed 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 can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be substantially implemented or contributed to by the prior art, or all or part of the technical solution may be embodied in a software product, which is stored in a storage medium and includes multiple instructions for causing a computer device (which may be a personal computer, a server, or a network device) to perform all or part of the steps of the method of the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing programs, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The preferred embodiments of the present application have been described above with reference to the accompanying drawings, and the scope of the claims of the embodiments of the present application is not limited thereto. Any modifications, equivalents and improvements that may occur to those skilled in the art without departing from the scope and spirit of the embodiments of the present application are intended to be within the scope of the claims of the embodiments of the present application.
Claims (10)
1. An insurance product extraction method based on matrix decomposition is characterized by comprising the following steps:
acquiring historical click behavior data of a plurality of users on each insurance product;
generating a user-insurance product scoring matrix according to the historical click behavior data;
extracting the insurance product according to the user-insurance product scoring matrix by at least one of:
calculating a scoring result for each of the insurance products in the user-insurance product scoring matrix;
ranking the grading results corresponding to the insurance products, and extracting the insurance products according to the ranking results of the grading results;
or,
determining the scoring result of each of the insurance products in the user-insurance product scoring matrix as an insurance product vector;
determining a target insurance product according to the historical click behavior data;
calculating cosine similarity between each insurance product and the target insurance product;
and sequencing the cosine similarity corresponding to each insurance product, and extracting the insurance products according to the sequencing result of the cosine similarity.
2. The matrix factorization based insurance product extraction method of claim 1, wherein after obtaining historical click behavior data of users on insurance products, the method further comprises:
and intercepting data of which the clicking times exceed a first threshold value in the historical clicking behavior data within a preset time period so as to retain the historical clicking behavior data of X insurance products clicked within the preset time period, wherein X is a positive integer.
3. The matrix factorization based insurance product extraction method of claim 1, wherein said step of calculating a scoring result for each said insurance product in said user-insurance product scoring matrix comprises:
decomposing the user-insurance product scoring matrix into a product of a first matrix and a second matrix, and constructing a third matrix;
(iv) squaring the error of the user-insurance product scoring matrix and the third matrix as a loss function;
solving the first matrix and the second matrix by using an alternating least square method according to the loss function;
and determining the scoring result of each insurance product according to the values of the first matrix and the second matrix obtained by solving.
4. The matrix factorization based insurance product extraction method of claim 3, wherein the loss function is expressed by:
wherein,
in the formula,
a square of error, R, representing a user-insurance product scoring matrix and said third matrix m,n A user-insurance product scoring matrix is represented,
denotes a third matrix, P m,k Representing a first matrix, Q k,m Representing a second matrix, m representing the number of users, n representing the number of insurance products, and K representing the spatial dimension.
5. The matrix factorization based insurance product extraction method of claim 3, wherein the step of solving the first and second matrices using an alternating least squares method according to the loss function comprises:
initializing values of the first matrix;
solving the second matrix by minimizing the loss function according to the values of the first matrix;
solving the first matrix by minimizing the loss function according to the value of the second matrix obtained by the solving;
and returning to execute the step of solving the second matrix by minimizing the loss function according to the value of the first matrix until the first matrix and the second matrix converge to have errors meeting a threshold condition.
6. The matrix factorization based insurance product extraction method according to claim 1, wherein the step of ranking the scoring results corresponding to each insurance product and extracting the insurance product according to the ranking results of the scoring results comprises:
sorting the scoring results corresponding to the insurance products according to the numerical value from large to small to obtain a first sorting list;
selecting the first N insurance products in the first ranking list for extraction;
or,
sorting the scoring results corresponding to the insurance products according to the numerical order from small to large to obtain a second sorting list;
and selecting the last N insurance products in the second sorted list for extraction, wherein N is a positive integer.
7. The matrix decomposition-based insurance product extraction method according to claim 1, wherein the step of sorting the cosine similarity corresponding to each insurance product and extracting the insurance product according to the sorting result of the cosine similarity comprises:
sorting the cosine similarity corresponding to each insurance product according to the numerical value from large to small to obtain a third sorting list;
selecting the top M insurance products in the third sorted list for extraction;
or,
sorting the cosine similarity corresponding to each insurance product according to the numerical value from small to large to obtain a fourth sorting list;
and selecting the back M insurance products in the fourth sorted list for extraction, wherein M is a positive integer.
8. An insurance product extraction apparatus based on matrix factorization, comprising:
the acquisition module is used for acquiring historical click behavior data of a plurality of users on each insurance product;
the generating module is used for generating a user-insurance product scoring matrix according to the historical click behavior data;
an extraction module for extracting the insurance product according to the user-insurance product scoring matrix by at least one of:
a first calculation unit for calculating a scoring result for each of the insurance products in the user-insurance product scoring matrix;
the first sorting and extracting unit is used for sorting the scoring results corresponding to the insurance products and extracting the insurance products according to the sorting results of the scoring results;
or,
a first determining unit, configured to determine a scoring result of each of the insurance products in the user-insurance product scoring matrix as an insurance product vector;
the second determining unit is used for determining a target insurance product according to the historical click behavior data;
the second calculation unit is used for calculating cosine similarity between each insurance product and the target insurance product;
and the second sequencing and extracting unit is used for sequencing the cosine similarity corresponding to each insurance product and extracting the insurance products according to the sequencing result of the cosine similarity.
9. An electronic device, characterized in that the electronic device comprises a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for enabling a connection communication between the processor and the memory, which program, when executed by the processor, realizes the steps of the method according to any one of claims 1 to 7.
10. A storage medium, being a computer readable storage medium, for computer readable storage, characterized in that the storage medium stores one or more programs executable by one or more processors to implement the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210933439.4A CN115375484A (en) | 2022-08-04 | 2022-08-04 | Matrix decomposition-based insurance product extraction method and device, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210933439.4A CN115375484A (en) | 2022-08-04 | 2022-08-04 | Matrix decomposition-based insurance product extraction method and device, equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115375484A true CN115375484A (en) | 2022-11-22 |
Family
ID=84064135
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210933439.4A Pending CN115375484A (en) | 2022-08-04 | 2022-08-04 | Matrix decomposition-based insurance product extraction method and device, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115375484A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117522352A (en) * | 2024-01-08 | 2024-02-06 | 安徽国元保险经纪股份有限公司 | Block chain-based safe production responsibility insurance informatization management system and method |
-
2022
- 2022-08-04 CN CN202210933439.4A patent/CN115375484A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117522352A (en) * | 2024-01-08 | 2024-02-06 | 安徽国元保险经纪股份有限公司 | Block chain-based safe production responsibility insurance informatization management system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9607010B1 (en) | Techniques for shape-based search of content | |
| CN108805598B (en) | Similarity information determination method, server and computer-readable storage medium | |
| CN108108821A (en) | Model training method and device | |
| CN109492180A (en) | Resource recommendation method, device, computer equipment and computer readable storage medium | |
| CN110008973B (en) | Model training method, method and device for determining target user based on model | |
| CN110008397B (en) | Recommendation model training method and device | |
| CN111198939B (en) | Statement similarity analysis method and device and computer equipment | |
| CN110580278A (en) | personalized search method, system, equipment and storage medium according to user portrait | |
| CN111967914A (en) | User portrait based recommendation method and device, computer equipment and storage medium | |
| CN113657087B (en) | Information matching method and device | |
| Wang et al. | CROWN: a context-aware recommender for web news | |
| CN114240552A (en) | Product recommendation method, device, equipment and medium based on deep clustering algorithm | |
| CN114266443A (en) | Data evaluation method and device, electronic equipment and storage medium | |
| CN111159578B (en) | Method and system for recommending objects | |
| US11269896B2 (en) | System and method for automatic difficulty level estimation | |
| CN115375484A (en) | Matrix decomposition-based insurance product extraction method and device, equipment and medium | |
| CN114579858A (en) | Content recommendation method and device, electronic equipment and storage medium | |
| CN111882224A (en) | Method and device for classifying consumption scenes | |
| CN115879455A (en) | Word emotion polarity prediction method and device, electronic equipment and storage medium | |
| CN115965439A (en) | Data recall method, data recall device, electronic equipment and storage medium | |
| CN116089616A (en) | Theme text acquisition method, device, equipment and storage medium | |
| CN106445934A (en) | Data processing method and apparatus | |
| CN114896294A (en) | Method, device and equipment for generating product test data table and storage medium | |
| US20220172008A1 (en) | An apparatus and method for performing image-based dish recognition | |
| US10223728B2 (en) | Systems and methods of providing recommendations by generating transition probability data with directed consumption |
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 |