WO2019061664A1

WO2019061664A1 - Electronic device, user's internet surfing data-based product recommendation method, and storage medium

Info

Publication number: WO2019061664A1
Application number: PCT/CN2017/108789
Authority: WO
Inventors: 刘睿恺; 吴振宇; 王建明; 肖京
Original assignee: 平安科技（深圳）有限公司
Priority date: 2017-09-30
Filing date: 2017-10-31
Publication date: 2019-04-04
Also published as: CN107818491A

Abstract

Disclosed in the present invention is a user's Internet surfing data-based product recommendation method. Said method comprises acquiring, from a predetermined database, Internet surfing data of each user within a preset time; using a predetermined service keyword table generation rule to analyze the acquired Internet surfing data of each user, so as to generate a service keyword table corresponding to each user; and analyzing, according to a predetermined user preferred product recommendation model, the service keyword table corresponding to the user, so as to determine a preferred product corresponding to the user, and sending, to a predetermined terminal, a product recommendation instruction for the user preference. In this way, the present invention can not only avoid the drawbacks in the prior art of a heavy workload caused by the necessarily analyzing mass redundant data having a low value density, but also improve the working efficiency and the accuracy of product recommendation.

Description

Electronic device, product recommendation method based on user internet data, and storage medium

This application claims the priority of the Chinese patent application filed on September 30, 2017, the Chinese Patent Office, the application number is 201710916189.2, and the invention name is "electronic device, product recommendation method based on user Internet data and storage medium", all contents thereof This is incorporated herein by reference.

Technical field

The present invention relates to the field of Internet product recommendation, and in particular, to an electronic device and a product recommendation method based on user internet data.

Background technique

In recent years, with the development of the Internet, better understanding and analysis of Internet users' online behavior has become crucial for the related applications of the Internet industry, for example, recommendation of personalized products, user security, and user behavior orientation. Etc. became crucial.

At present, there is no fast and efficient method for the Internet behavior of Internet users. This is mainly limited by the redundancy of the real behavior data of Internet users and the low value density. It is necessary to obtain data from massive redundancy and low value density. It is necessary to do a lot of tedious work in the process of analyzing the hobbies and interests of Internet users and recommending related products to relevant users according to hobbies and hobbies. Therefore, the existing methods have a large workload, low efficiency, and low accuracy.

Summary of the invention

In view of this, the present invention provides an electronic device and a product recommendation method based on user Internet data, which can automatically complete product recommendation of the user according to the user's online data, and improve recommendation efficiency and accuracy.

First, in order to achieve the above object, a first aspect of the present invention provides a product recommendation method based on user internet data, and the method includes the following steps:

A. If the product needs to be recommended to the user with the identification information, obtain the first online data of the user within the first preset time from the predetermined database, and the other users are in the first preset time. Second online data;

B. analyzing the obtained first Internet data and the second Internet data by using a predetermined business keyword table generation rule, generating a first service keyword table corresponding to the user, and performing the acquired second Internet data. The second business keyword table corresponding to each of the other users is generated, wherein the business keyword includes a verb describing a user's online behavior and a noun describing the product, and the business keyword table includes the same user at the first preset All business keywords during the time;

C. analyzing the first service keyword table and the second service keyword table according to a predetermined similarity analysis rule to analyze a similar user corresponding to the user;

D. analyzing the first service keyword table and the second service keyword table corresponding to the similar user according to a predetermined user preference product recommendation model, to determine a product that the user prefers, to The predetermined terminal sends a product recommendation instruction for the user.

In addition, in order to achieve the above object, a second aspect of the present invention provides an electronic device including a memory, a processor, and a user-based data-based product stored on the memory and operable on the processor The recommendation system, when the product recommendation system based on the user's Internet data is executed by the processor, implements the following steps:

B. analyzing the obtained first Internet data by using a predetermined business keyword table generation rule, generating a first service keyword table corresponding to the user, and analyzing the acquired second Internet data to generate other users. Corresponding second service keyword table, wherein the service keyword includes a verb describing a user's online behavior and a noun describing the product, and the business keyword table includes all services of the same user in the first preset time Key words;

D. The first service keyword table and the second service keyword table corresponding to the similar user are analyzed according to a predetermined user preference product recommendation model, and the product preferred by the user is determined, and sent to a predetermined terminal. Product recommendation instructions for this user.

Further, in order to achieve the above object, a third aspect of the present invention provides a computer readable storage medium, where the computer readable storage medium stores a product recommendation system based on user internet data, and the product recommendation system based on user internet data The at least one processor can be executed by the at least one processor to perform the following steps:

Compared with the prior art, the electronic device, the product recommendation method based on the user's Internet access data, and the computer readable storage medium obtain the Internet data of each user within a preset time from a predetermined database; The predetermined business keyword table generation rule analyzes the obtained online data of each user to generate a service keyword table corresponding to each user; The product preference recommendation model corresponding to the user is analyzed according to the predetermined user preference product recommendation model to determine the product corresponding to the preference of the user, and the product recommendation instruction for the user preference is sent to the predetermined terminal. In this way, the disadvantages of the large amount of work caused by the analysis of the data with large redundancy and low value density in the prior art can be avoided, and the work efficiency and the accuracy of the recommended products are also improved.

DRAWINGS

1 is a schematic diagram of an optional hardware architecture of an electronic device of the present invention;

2 is a schematic diagram of a program module of a product recommendation program based on user internet data in an implementation of an electronic device according to the present invention;

3 is a schematic flow chart of an implementation of a preferred embodiment of a product recommendation method based on user internet data according to the present invention.

The implementation, functional features, and advantages of the present invention will be further described in conjunction with the embodiments.

Detailed ways

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

It should be noted that the descriptions of "first", "second" and the like in the present invention are for the purpose of description only, and are not to be construed as indicating or implying their relative importance or implicitly indicating the number of indicated technical features. . Thus, features defining "first" and "second" may include at least one of the features, either explicitly or implicitly. In addition, the technical solutions between the various embodiments may be combined with each other, but must be based on the realization of those skilled in the art, and when the combination of the technical solutions is contradictory or impossible to implement, it should be considered that the combination of the technical solutions does not exist. It is also within the scope of protection required by the present invention.

Referring to FIG. 1 , it is a schematic diagram of an optional application environment of the electronic device of the present invention.

In this embodiment, the electronic device 1 may include, but is not limited to, the memory 11, the processor 12, and the network interface 13 being communicably connected to each other through a system bus. It should be noted that FIG. 1 only shows the electronic device 1 having the components 11-13, but it should be understood that not all illustrated components may be implemented, and more or fewer components may be implemented instead.

The memory 11 includes at least one type of readable storage medium including a flash memory, a hard disk, a multimedia card, a card type memory (for example, SD or DX memory, etc.), a random access memory (RAM), and a static random access. Memory (SRAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), programmable read only memory (PROM), magnetic memory, magnetic disk, optical disk, and the like. In some embodiments, the memory 11 may be an internal storage unit of the electronic device 1, such as a hard disk or a memory of the electronic device 1. In other embodiments, the memory 11 is also It may be an external storage device of the electronic device 1, such as a plug-in hard disk equipped on the electronic device 1, a smart memory card (SMC), a Secure Digital (SD) card, a flash card, or the like. . Of course, the memory 11 can also include both an internal storage unit of the electronic device 1 and an external storage device thereof. In this embodiment, the memory 11 is generally used to store an operating system installed on the electronic device 1 and various types of application software, such as a product recommendation program based on user Internet data. Further, the memory 11 can also be used to temporarily store various types of data that have been output or are to be output.

Processor 12 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor 12 is generally used to control the overall operation of the electronic device 1. For example, in this embodiment, the processor 12 is configured to run program code or processing data stored in the memory 11, such as a running product recommendation program based on user Internet data. .

The network interface 13 may include a wireless network interface or a wired network interface, and the network interface 13 is generally used to establish a communication connection between the electronic device 1 and other electronic devices.

Optionally, the device may further include a user interface, the user interface may include a display, an input unit such as a keyboard, and the optional user interface may further include a standard wired interface and a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch sensor, or the like. Wherein, the display may also be appropriately referred to as a display screen or a display unit for displaying information processed in the electronic device and a user interface for displaying visualization.

In the electronic device embodiment shown in FIG. 1 , the memory 11 stores a product recommendation program based on the user's Internet data. When the processor 12 executes the product recommendation program based on the user's Internet data stored in the memory 11 , the following steps are implemented:

Generally, the online data of the user within a preset time, for example, one week, one month, etc., may be obtained according to the user's identification information, such as a login name, a registered user name, a mobile phone number, and the like. And won The received Internet data is data for recording the user's real behavior. For example, at 2:40 pm on March 27, 2017, the user named A clicks the button of the certificate to send the traffic activity, and downloads the certificate APP. The user's behavior and preferred products can be analyzed according to the words with actual business meaning in the data recording the user's real behavior. For example, the verb can be clicked and downloaded according to the above-mentioned real behavior data to describe the user's online behavior, and the nouns describing the product. Proof of APP, traffic, and buttons to analyze the user's behavior and preferred products. Therefore, if the words with actual business meaning can be extracted from the obtained online data that records the user's real behavior, then it can be obtained according to the obtained The words of actual business meaning effectively analyze the products that the user prefers, and recommend related products to the user according to the products preferred by the user.

In this embodiment, if the product needs to be recommended to the user with the identification information, the first online data of the user within the first preset time (for example, one week) is first obtained from a predetermined database (the user) The real online behavior data), and other individual users (for example, other users of the predetermined preset number of users except the other users), the second Internet data in the first preset time (other individual users) Real online behavior data); then, using the predetermined business keyword table generation rule to analyze the obtained first Internet data, generating a first service keyword table corresponding to the user, and acquiring the second Internet data Performing an analysis to generate a second service keyword table corresponding to each of the other users, where the predetermined service keyword table generation rule includes:

Constructing a set of service keywords in the online data of the preset number of users in the second preset time, specifically, in an embodiment, constructing the business keywords in the Internet data of 300 users within two months The process of forming the set includes pre-acquiring the online data of 300 users in two months, and extracting keywords with actual business meanings from the obtained online data, wherein the keywords have actual business meanings. Including verbs that describe the user's online behavior (eg, click, download, purchase, etc.), and nouns describing the product (eg, food, credit card, wealth management products, etc.), the extracted verbs describing the user's online behavior, and description products The set of nouns is a collection of business keywords constructed to meet the above requirements.

And acquiring, according to the first online data, a plurality of first service keywords that match the business keywords in the set of the business keywords, and based on the acquired first The business keyword generates a corresponding first business keyword table; similarly, the second online data traverses the set of the business keyword to obtain the business keyword in the set composed of the business keyword from the second online data. Each of the second service keywords generates a corresponding second service keyword table based on the acquired second service keywords.

Further, in the embodiment, the predetermined similarity analysis rule may be a cosine angle similarity method, a Euclidean distance metric method, or a Pearson correlation coefficient method. For example, in one embodiment, the Euclidean distance metric is taken as an example, wherein the Euclidean distance metric uses the following formula to calculate the similarity:

It should be noted that x and y in the above formula are respectively normalized vectors, and the result of the above formula is the similarity between the vector x and the vector y. In this embodiment, the user and other users are calculated. The first service vector table corresponding to the user is normalized according to a preset normalization manner to obtain the first vector x in the above formula, and the second service corresponding to each other user is obtained. After the keyword table is normalized according to the preset normalization method, a second vector y different from the above formula is obtained, and then the first vector x and the other second vectors y are respectively substituted into the above similarity calculation. a formula that calculates a similarity between the user and each of the other users, and compares the calculated similarity with a preset similarity threshold, and if the calculated similarity is greater than a preset similarity threshold, Determined to be a similar user for this user.

It should be further noted that, in this embodiment, the predetermined user preference product recommendation model is a collaborative filtering recommendation model, and the training process of the user preference product recommendation model includes the following steps:

Obtaining a preset number of second service keyword table samples of similar users, and dividing the second service keyword table samples of each similar user into a corresponding first ratio training set and a second ratio test set;

The user preference product recommendation model is trained by using a second business keyword table of each similar user in the training set to obtain a trained user preference product recommendation model;

The user preference product recommendation model is tested by using the second service keyword table of each similar user in the test set. If the test passes, the training ends, or if the test fails, the second business keyword of the similar user in the training set is increased. The table samples and re-execute the steps described above for training the user preference product recommendation model.

The steps of testing the user preference product recommendation model using the second business keyword table of each similar user in the test set include:

Using the trained user product recommendation model to analyze the second business keyword table of each similar user in the test set to obtain the preference probability values of different similar users for different products;

If a similar user's preference probability value for the preset type product is greater than a preset preference probability threshold, the model accuracy test is performed for the similar user, and the user is confirmed to pay attention to the preset type product from the online data of the similar user. a frequency value, if the similar user pays attention to the frequency value of the preset type product exceeds a preset frequency threshold, determining that the model accuracy test result for the similar user is correct, or if the similar user pays attention to the preset type product If the frequency value is equal to or less than the preset frequency threshold, it is determined that the model accuracy test result for the similar user is an error;

If the correct model accuracy test results account for more than the preset percentage threshold for all model accuracy test results, then the test for the user preference product recommendation model is determined, or if the correct model accuracy test results account for all models If the percentage of the accuracy test result is less than or equal to the preset percentage threshold, it is determined that the test for the user preference product recommendation model fails.

In another embodiment, in order to make the calculation process faster, after the step B above, the first service keyword table and the second service keyword table are analyzed by using a predetermined clustering algorithm to analyze the second service. Among the users corresponding to the keyword table, the users corresponding to the first service keyword table belong to the same user group. The predetermined clustering algorithm includes a density-based clustering algorithm, and the specific process is: pre-setting a density hierarchy (for example, in the embodiment, the first service The keyword list includes different preset types of business keywords corresponding to different density levels, a highest density threshold (eg, including 6 preset types of business keywords), and a minimum density threshold (eg, including 3 pre- Set the type of business keyword), and according to the set density level, the highest density threshold, and the lowest density threshold, analyze the density levels included in each second business keyword table, and the density thresholds included in each density level, and analyze The density thresholds are arranged according to the density threshold from high to low, the highest density threshold is selected, and all preset types of business keywords are clustered for the first time based on the selected density threshold to generate cluster clusters. The residual density threshold repeatedly performs the above clustering process until the ith clustering is performed to generate a clustering family; wherein the clustering cluster generated by the i-th clustering process can only be expanded in the subsequent clustering process and cannot be Segmentation or merging into other clusters; and clusters that satisfy the current density threshold are preferentially extracted; the algorithm is for all preset classes in turn Keywords business at different cluster density threshold, the i-th cluster results directly as input i + 1-time clustering, until the remaining threshold value is less than the density of the lowest density threshold, then the algorithm ends. So far, the user corresponding to the first service keyword table corresponding to each user in the second service keyword table is clustered by the user group.

Then, analyzing, according to the predetermined similarity analysis rule, the first service keyword table and the fourth service keyword table corresponding to other users in the customer group to which the user belongs, to analyze the customer group to which the user belongs A similar user corresponding to the user.

Optionally, in other embodiments, the product recommendation program based on the user's Internet data may be divided into one or more program modules, and one or more program modules are stored in the memory 11 and processed by one or more The processor (which is processor 12 in this embodiment) is executed to complete the present invention. A program module as used herein refers to a series of computer program instructions that are capable of performing a particular function.

For example, referring to FIG. 2, it is a schematic diagram of a product recommendation program module based on user internet data in an embodiment of the electronic device of the present invention. In this embodiment, the product recommendation program based on the user's Internet access data may be divided into an acquisition module 201, a first analysis module 202, a second analysis module 203, and a third analysis module 204. The functions or operation steps implemented by the program modules 201-204 in this embodiment are similar to the above, and are not described in detail herein, for example, where:

The obtaining module 201 is configured to: if the product is recommended to the user with the identification information, obtain the first online data of the user in the first preset time from the predetermined database, and the other users are in the first pre-preparation Set the second online data in time;

The first analysis module 302 is configured to analyze the obtained first Internet data by using a predetermined service keyword table generation rule, generate a first service keyword table corresponding to the user, and generate the second Internet data. Performing an analysis to generate a second service keyword table corresponding to each of the other users, wherein the service keyword includes a verb describing the user's online behavior and a noun describing the product, and the business keyword table includes all services of the same user within the first preset time. Key words;

The second analysis module 303 is configured to analyze the first service keyword table and the second service keyword table according to the predetermined similarity analysis rule to analyze the similar user corresponding to the user;

The third analysis module 304 is configured to analyze the first service keyword table and the second service keyword table corresponding to the similar user according to the predetermined user preference product recommendation model, to determine the product that the user prefers, to a predetermined The terminal sends a product recommendation instruction for the user.

In addition, the present invention also provides a product recommendation method based on user online data. Referring to FIG. 3, it is a schematic flowchart of an implementation example of a preferred example of a product recommendation method based on user Internet data according to the present invention. The method can be performed by a device that can be implemented by software and/or hardware.

In this embodiment, the product recommendation method based on the user's online data includes:

Step S301: If the product needs to be recommended to the user with the identification information, the first online data of the user in the first preset time and the other users in the first preset time are obtained from the predetermined database. Second internet data;

Step S302, analyzing the acquired first Internet data by using a predetermined business keyword table generation rule, generating a first service keyword table corresponding to the user, and analyzing the acquired second Internet data to generate other each a second service keyword table corresponding to the user, wherein the service keyword includes a verb describing the user's online behavior and a noun describing the product, and the business keyword table includes all the business keywords of the same user within the first preset time;

Step S303, analyzing the first service keyword table and the second service keyword table according to the predetermined similarity analysis rule, to analyze the similar user corresponding to the user;

Step S304, analyzing the first service keyword table and the second service keyword table corresponding to the similar user according to the predetermined user preference product recommendation model, to determine the product preferred by the user, and sending the predetermined user to the predetermined terminal. Product recommendation instructions.

Generally, the online data of the user within a preset time, for example, one week, one month, etc., may be obtained according to the user's identification information, such as a login name, a registered user name, a mobile phone number, and the like. The obtained online data is data for recording the real behavior of the user. For example, at 2:40 pm on March 27, 2017, the user with the user name A clicks the button of the card to send the traffic activity, and downloads the certificate APP. The product that has the actual business meaning in the data that records the user's real behavior can be used to analyze the user's behavior and preferred products. For example, the verb can be clicked and downloaded according to the above-mentioned real behavior data to describe the user's online behavior, and the product description is described. Nouns, APPs, traffic, and buttons to analyze the user's behavior and preferences of the product, so if you can extract words with actual business meaning from the online data of the recorded user's real behavior, you can get the Words with actual business meaning effectively analyze the products that the user prefers, and recommend related products to the user according to the products preferred by the user.

Constructing a set of service keywords in the online data of the preset number of users in the second preset time, specifically, in an embodiment, constructing the business keywords in the Internet data of 300 users within two months The process of composing the collection includes pre-acquisition of 300 users online within two months. Data, and extracting keywords having actual business meanings from the obtained online data, wherein keywords having actual business meanings include verbs describing user online behavior (eg, clicking, downloading, purchasing, etc.), and Descriptive product nouns (eg, food, credit cards, wealth management products, etc.), the set of extracted verbs describing the user's online behavior, and the nouns describing the product are the set of business keywords that satisfy the above requirements. .

Traversing the set of the business keyword, obtaining the first words matching the business keywords in the set of the business keywords from the first online data, and generating the corresponding first service based on the acquired first words Keyword table; similarly, traversing a set of business keyword sets, obtaining second words matching the business keywords in the set of business keywords from the second internet data, and based on the acquired second words The word generates a corresponding second business keyword table.

Using the second business keyword table of each similar user in the test set to recommend the product preference product The steps for testing are:

In another specific embodiment, in order to make the calculation process faster, after the step B above, the first service keyword table and the second service keyword table are analyzed by using a predetermined clustering algorithm to analyze the Among the users corresponding to the two service keyword tables, the users corresponding to the first service keyword table belong to the same user group. The predetermined clustering algorithm includes a density-based clustering algorithm, and the specific process is: pre-setting a density level (for example, in the embodiment, the first service keyword table includes different preset types of business keywords. The number corresponds to setting different density levels), the highest density threshold (for example, including 6 preset types of business keywords), and the lowest density threshold (for example, including 3 preset types of business keywords), and according to the settings The density level, the highest density threshold, and the lowest density threshold analyze the density levels contained in each of the second business keyword tables, and the density thresholds included in each density level, and the respective density thresholds to be analyzed are based on the density threshold Low arrangement, selecting the highest density threshold and performing the first clustering on all preset types of business keywords based on the selected density threshold, generating a cluster cluster, and repeating the remaining density threshold to perform the clustering process until proceeding The i-th cluster generates a cluster family; wherein the cluster cluster generated by the i-th cluster process is in the subsequent clustering process Can only be extended and cannot be split or merged into other clusters; and clusters that satisfy the current density threshold are preferentially extracted; the algorithm sequentially clusters all preset types of business keywords at different density thresholds, i The result of the sub-cluster is directly used as the input of the i+1th cluster until the remaining density thresholds are less than the lowest density threshold, and the algorithm ends. So far, the user corresponding to the first service keyword table corresponding to each user in the second service keyword table is clustered by the user group.

In addition, the embodiment of the present invention further provides a computer readable storage medium, where the product recommendation program based on the user's Internet data is stored, and the product recommendation program based on the user's Internet data is executed by the processor as follows: operating:

If you need to recommend products to users with identification information, you can get them from a predetermined database. The first Internet access data of the user in the first preset time, and the second Internet access data of the other users in the first preset time;

The first online data collected by the user is analyzed by using a predetermined business keyword table generation rule, and the first service keyword table corresponding to the user is generated, and the acquired second online data is analyzed to generate corresponding addresses of other users. a second service keyword table, wherein the service keyword includes a verb describing a user's online behavior and a noun describing the product, and the business keyword table includes all business keywords of the same user in the first preset time ;

And analyzing the first service keyword table and the second service keyword table according to a predetermined similarity analysis rule to analyze a similar user corresponding to the user;

And analyzing, according to the predetermined user preference product recommendation model, the first service keyword table and the second service keyword table corresponding to the similar user, to determine a product that is preferred by the user, and sending the target to the predetermined terminal. User's product recommendation instructions.

Further, when the product recommendation program based on the user's Internet data is executed by the processor, the following operations are also implemented:

And analyzing the first service keyword table and the second service keyword table according to a predetermined clustering algorithm, respectively, to respectively respectively, the respective users corresponding to the second service keyword table and the first service keyword table The corresponding users perform clustering of user groups.

And analyzing, according to the predetermined similarity analysis rule, the first service keyword table and the fourth service keyword table corresponding to other users in the customer group to which the user belongs, to analyze the customer group to which the user belongs A similar user corresponding to the user.

The specific embodiment of the computer readable storage medium of the present invention is substantially the same as the above embodiments of the electronic device and the product recommendation method based on the user's Internet access data, and will not be described herein.

According to the foregoing embodiments, the electronic device of the present invention, the product recommendation method based on the user's Internet access data, and the computer readable storage medium utilize the advance data by acquiring the Internet data of each user within a preset time period from a predetermined database. The determined business keyword generation rule analyzes the obtained online data to generate a business keyword table corresponding to each user; analyzes the business keyword table according to the predetermined similarity analysis rule to analyze the corresponding user Similar user; analyzing the first service keyword table corresponding to the user and the second service keyword table corresponding to the similar user according to the predetermined user preference product recommendation model, to determine the product preferred by the user, to the predetermined terminal Send a product recommendation instruction for this user. It can automatically complete the product recommendation of the user's preference according to the user's online data, and improve the recommendation efficiency and accuracy.

The serial numbers of the embodiments of the present invention are merely for the description, and do not represent the advantages and disadvantages of the embodiments.

Through the description of the above embodiments, those skilled in the art can clearly understand that the foregoing embodiment method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be through hardware, but in many cases, the former is better. Implementation. Based on such understanding, the technical solution of the present invention, which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a storage medium (such as ROM/RAM, disk, CD) There are a number of instructions for causing a terminal device (which may be a cell phone, computer, server, air conditioner, or network device, etc.) to perform the methods described in various embodiments of the present invention.

The above are only the preferred embodiments of the present invention, and are not intended to limit the scope of the invention, and the equivalent structure or equivalent process transformations made by the description of the present invention and the drawings are directly or indirectly applied to other related technical fields. The same is included in the scope of patent protection of the present invention.

Claims

A product recommendation method based on user internet data, characterized in that the method comprises the following steps:

A. If the product needs to be recommended to the user with the identification information, obtain the first online data of the user within the first preset time from the predetermined database, and the other users are in the first preset time. Second online data;

B. analyzing the obtained first Internet data by using a predetermined business keyword table generation rule, generating a first service keyword table corresponding to the user, and analyzing the acquired second Internet data to generate other users. Corresponding second service keyword table, wherein the service keyword includes a verb describing a user's online behavior and a noun describing the product, and the business keyword table includes all services of the same user in the first preset time Key words;

C. analyzing the first service keyword table and the second service keyword table according to a predetermined similarity analysis rule to analyze a similar user corresponding to the user;

D. The first service keyword table and the second service keyword table corresponding to the similar user are analyzed according to a predetermined user preference product recommendation model, and the product preferred by the user is determined, and sent to a predetermined terminal. Product recommendation instructions for this user.
The product recommendation method based on the user's Internet access data according to claim 1, wherein the predetermined business keyword table generation rule comprises:

Constructing a set of business keywords in the online data of all users in the second preset time;

Matching the first Internet data with the service keyword according to a predetermined matching rule to obtain a first mapping relationship between the first Internet data and the service keyword;

Matching the second Internet data with the service keyword according to the predetermined matching rule to obtain a second mapping relationship between the second Internet data and the service keyword;

Generating, according to the first mapping relationship, a first service keyword table corresponding to the user;

And generating, according to the second mapping relationship, a second service keyword table corresponding to another user.
The product recommendation method based on user internet data according to claim 2, wherein the predetermined similarity analysis rule comprises a cosine angle similarity method, a Euclidean distance metric method, or a Pearson correlation coefficient method. .
The product recommendation method based on user internet data according to claim 2, wherein the user preference product recommendation model is a collaborative filtering recommendation model, and the training process of the user preference product recommendation model includes the following steps:

E. Obtain a preset number of third service keyword table samples of similar users, and divide the third service keyword table samples of each similar user into a corresponding first ratio training set and a second ratio test set;

F. training the user preference product recommendation model by using a third service keyword table of each similar user in the training set to obtain a trained user preference product recommendation model;

G. testing, by using a third service keyword table of each similar user in the test set, the user preference product recommendation model, if the test passes, the training ends, or if the test fails, Then, the third service keyword table sample of the similar user in the training set is added and the step E and the step F are re-executed.
The method for recommending a product based on user online data according to claim 4, wherein the step of testing the user preference product recommendation model by using a third service keyword table of each similar user in the test set comprises: :

Using the trained user product recommendation model to analyze the third service keyword table of each similar user in the test set to obtain a preference probability value of each similar user for different products;

If a similar user's preference probability value for the preset type product is greater than a preset preference probability threshold, the model accuracy test is performed for the similar user, and the user is confirmed to pay attention to the preset type product from the online data of the similar user. a frequency value, if the similar user pays attention to the frequency value of the preset type product exceeds a preset frequency threshold, determining that the model accuracy test result for the similar user is correct, or if the similar user pays attention to the preset type product If the frequency value is equal to or less than the preset frequency threshold, it is determined that the model accuracy test result for the similar user is an error;

If the correct model accuracy test result accounts for more than the preset percentage threshold for all model accuracy test results, then the test for the user preference product recommendation model is determined, or if the correct model accuracy test results account for If the percentage of all model accuracy test results is less than or equal to the preset percentage threshold, it is determined that the test of the user preference product recommendation model fails.
The product recommendation method based on the user's Internet access data according to claim 1, wherein after the step B, the method further comprises the following steps:

The first service keyword table and the second service keyword table are analyzed according to a predetermined clustering algorithm to cluster each customer into a customer group.
The product recommendation method based on user internet data according to claim 6, wherein the step C is replaced by:

And analyzing, according to the predetermined similarity analysis rule, the first service keyword table and the fourth service keyword table corresponding to other users in the customer group to which the user belongs, to analyze the customer group to which the user belongs A similar user corresponding to the user.
The product recommendation method based on user internet data according to claim 7, wherein the predetermined clustering algorithm comprises a density-based clustering algorithm.
An electronic device, comprising: a memory, a processor, and a product recommendation system based on user internet data stored on the memory and operable on the processor, the user internet data The product recommendation system is implemented by the processor to implement the following steps:

A. If the product needs to be recommended to the user with the identification information, obtain the first online data of the user within the first preset time from the predetermined database, and the other users are in the first preset time. Second online data;

B. analyzing the obtained first Internet data by using a predetermined business keyword table generation rule, generating a first service keyword table corresponding to the user, and analyzing the acquired second Internet data to generate other users. Corresponding second business keyword table, wherein the business keyword includes a verb describing a user's online behavior and a noun describing the product, and the business keyword table includes the same All business keywords of a user within the first preset time;

C. analyzing the first service keyword table and the second service keyword table according to a predetermined similarity analysis rule to analyze a similar user corresponding to the user;

D. The first service keyword table and the second service keyword table corresponding to the similar user are analyzed according to a predetermined user preference product recommendation model, and the product preferred by the user is determined, and sent to a predetermined terminal. Product recommendation instructions for this user.
The electronic device according to claim 9, wherein the predetermined business keyword table generation rule comprises:

Constructing a set of business keywords in the online data of all users in the second preset time;

Matching the first Internet data with the service keyword according to a predetermined matching rule to obtain a first mapping relationship between the first Internet data and the service keyword;

Matching the second Internet data with the service keyword according to the predetermined matching rule to obtain a second mapping relationship between the second Internet data and the service keyword;

Generating, according to the first mapping relationship, a first service keyword table corresponding to the user;

And generating, according to the second mapping relationship, a second service keyword table corresponding to another user.
The electronic device according to claim 10, wherein said predetermined similarity analysis rule comprises a cosine angle similarity method, a Euclidean distance metric, or a Pearson correlation coefficient method.
The electronic device according to claim 10, wherein the user preference product recommendation model is a collaborative filtering recommendation model, and the training process of the user preference product recommendation model comprises the following steps:

E. Obtain a preset number of third service keyword table samples of similar users, and divide the third service keyword table samples of each similar user into a corresponding first ratio training set and a second ratio test set;

F. training the user preference product recommendation model by using a third service keyword table of each similar user in the training set to obtain a trained user preference product recommendation model;

G. testing, by using a third service keyword table of each similar user in the test set, the user preference product recommendation model, if the test passes, the training ends, or if the test fails, the training set is increased. The third user keyword table sample of the similar user and the step E and the step F are re-executed.
The electronic device according to claim 12, wherein the step of testing the user preference product recommendation model by using a third service keyword table of each similar user in the test set comprises:

Using the trained user product recommendation model to analyze the third service keyword table of each similar user in the test set to obtain a preference probability value of each similar user for different products;

If a similar user's preference probability value for the preset type product is greater than a preset preference probability threshold, the model accuracy test is performed for the similar user, and the user is confirmed to pay attention to the preset type product from the online data of the similar user. Frequency value, if the similar user pays attention to the frequency of the preset type product If the rate value exceeds the preset frequency threshold, it is determined that the model accuracy test result for the similar user is correct, or if the similar user pays attention to the preset type product, the frequency value is equal to or less than the preset frequency threshold, The model accuracy test result of the similar user is an error;

If the correct model accuracy test result accounts for more than the preset percentage threshold for all model accuracy test results, then the test for the user preference product recommendation model is determined, or if the correct model accuracy test results account for If the percentage of all model accuracy test results is less than or equal to the preset percentage threshold, it is determined that the test of the user preference product recommendation model fails.
The electronic device according to claim 9, further comprising the following steps after said step B:

The first service keyword table and the second service keyword table are analyzed according to a predetermined clustering algorithm to cluster each customer into a customer group.
The electronic device according to claim 14, wherein said step C is replaced by:

And analyzing, according to the predetermined similarity analysis rule, the first service keyword table and the fourth service keyword table corresponding to other users in the customer group to which the user belongs, to analyze the customer group to which the user belongs A similar user corresponding to the user.
The electronic device of claim 15, wherein the predetermined clustering algorithm comprises a density based clustering algorithm.
A computer readable storage medium storing a product recommendation system based on user internet data, the product recommendation system based on user internet data being executable by at least one processor to enable the at least one The processor performs the following steps:

A. If the product needs to be recommended to the user with the identification information, obtain the first online data of the user within the first preset time from the predetermined database, and the other users are in the first preset time. Second online data;

B. analyzing the obtained first Internet data by using a predetermined business keyword table generation rule, generating a first service keyword table corresponding to the user, and analyzing the acquired second Internet data to generate other users. Corresponding second service keyword table, wherein the service keyword includes a verb describing a user's online behavior and a noun describing the product, and the business keyword table includes all services of the same user in the first preset time Key words;

C. analyzing the first service keyword table and the second service keyword table according to a predetermined similarity analysis rule to analyze a similar user corresponding to the user;

D. The first service keyword table and the second service keyword table corresponding to the similar user are analyzed according to a predetermined user preference product recommendation model, and the product preferred by the user is determined, and sent to a predetermined terminal. Product recommendation instructions for this user.
The computer readable storage medium of claim 17, wherein the predetermined business keyword table generation rule comprises:

Constructing a set of business keywords in the online data of all users in the second preset time;

Matching the first Internet data with the service keyword according to a predetermined matching rule to obtain a first mapping relationship between the first Internet data and the service keyword;

Matching the second Internet data with the service keyword according to the predetermined matching rule to obtain a second mapping relationship between the second Internet data and the service keyword;

Generating, according to the first mapping relationship, a first service keyword table corresponding to the user;

And generating, according to the second mapping relationship, a second service keyword table corresponding to another user.
The computer readable storage medium of claim 18, wherein the predetermined similarity analysis rule comprises a cosine angle similarity method, a Euclidean distance metric, or a Pearson correlation coefficient method.
The computer readable storage medium according to claim 18, wherein the user preference product recommendation model is a collaborative filtering recommendation model, and the training process of the user preference product recommendation model comprises the following steps:

E. Obtain a preset number of third service keyword table samples of similar users, and divide the third service keyword table samples of each similar user into a corresponding first ratio training set and a second ratio test set;

F. training the user preference product recommendation model by using a third service keyword table of each similar user in the training set to obtain a trained user preference product recommendation model;

G. testing, by using a third service keyword table of each similar user in the test set, the user preference product recommendation model, if the test passes, the training ends, or if the test fails, the training set is increased. The third user keyword table sample of the similar user and the step E and the step F are re-executed.