CN114579858A

CN114579858A - Content recommendation method and device, electronic equipment and storage medium

Info

Publication number: CN114579858A
Application number: CN202210209468.6A
Authority: CN
Inventors: 陶醉; 徐宁
Original assignee: Ping An Technology Shenzhen Co Ltd
Current assignee: Ping An Technology Shenzhen Co Ltd
Priority date: 2022-03-03
Filing date: 2022-03-03
Publication date: 2022-06-03

Abstract

The embodiment of the invention provides a content recommendation method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence. The content recommendation method comprises the following steps: the method comprises the steps of obtaining user behavior information, carrying out feature processing on the user behavior information to obtain user behavior factor data, carrying out feature analysis on the user behavior factor data to obtain interest content, generating a user interest correlation matrix according to the interest content, calculating according to the user interest correlation matrix to obtain user similarity, selecting candidate content according to the user similarity, and carrying out content recommendation. According to the embodiment of the invention, the user behavior information including the wireless data information and the track data information is acquired, and the multiple types of user data are combined for feature extraction, so that the interest content is obtained by acquiring the implicit related information among different data, and the interest content is utilized to carry out corresponding content recommendation, so that the accuracy of content recommendation and the correlation between the recommended content and the user requirements are improved.

Description

Content recommendation method and device, electronic equipment and storage medium

Technical Field

The present invention relates to the field of artificial intelligence technologies, and in particular, to a content recommendation method and apparatus, an electronic device, and a storage medium.

Background

With the development of internet technology and mobile terminals, information shows explosive expansion, and people are moving to the information overload era from the information deficiency era, and the way of acquiring information is more various and more convenient. As an information consumer, how to find information of interest from a large amount of information, as an information producer, how to make information put in by the consumer stand out and pay attention is a problem that needs to be solved today. The content recommendation system is a method for alleviating the above problems, and the task thereof is to contact users and information, help users find valuable interest points for themselves, and accurately present information in front of users interested in the information, thereby realizing win-win of information consumers and information producers.

In the related art, most content recommendation systems capture historical data of a certain aspect of a user, for example, browsing history or search keywords, to perform related content recommendation, or acquire a geographic position of the user in combination with positioning information of the user to perform recommendation of functions such as peripheral search, coupon search, card punching, and the like, but in the related art, multiple types of user data are not combined to acquire implicit related information to perform corresponding content recommendation, so that accuracy of content recommendation is not high, and relevance of recommended content and user requirements is low.

Disclosure of Invention

The embodiment of the invention mainly aims to provide a content recommendation method and device, electronic equipment and a storage medium, which can improve the accuracy of content recommendation and the correlation between push content and user requirements.

In order to achieve the above object, an aspect of the embodiments of the present invention provides a content recommendation method, including:

acquiring user behavior information, wherein the user behavior information comprises: wireless data information and trajectory data information;

performing feature processing on the user behavior information to obtain user behavior factor data, wherein the user behavior factor data comprises: wireless factor data and trajectory factor data;

performing characteristic analysis on the user behavior factor data to obtain interest content;

generating a user interest correlation matrix according to the interest content;

calculating according to the user interest correlation matrix to obtain user similarity;

and selecting candidate contents according to the user similarity to recommend the contents.

In some embodiments, the performing feature processing on the user behavior information to obtain user behavior factor data includes:

performing characteristic processing on the wireless data information to obtain the wireless factor data, wherein the wireless factor data comprises at least one of the following data: the system comprises local area network name factor data, browsing content factor data, browsing time factor data, browsing frequency factor data and browsing time factor data;

and performing characteristic processing on the track data information to obtain the track factor data.

In some embodiments, the performing feature analysis on the user behavior factor data to obtain interest content includes:

performing characteristic analysis according to the user behavior factor data to obtain location correlation content, browsing content correlation content and browsing duration correlation content;

and generating the interest content according to the location correlation content, the browsing content correlation content and the browsing duration correlation content.

In some embodiments, before generating the user interest correlation matrix according to the interest content, the content recommendation method further includes:

performing feature selection on the interest content to obtain interest screening content;

and performing numerical operation on the interest screening content to obtain numerical interest content so as to generate the user interest correlation matrix according to the numerical interest content.

In some embodiments, the generating a user interest relevance matrix from the interest content includes:

generating a relevance value of the user for the interest content according to the use frequency;

and generating the user interest correlation matrix according to the correlation value.

In some embodiments, the calculating the user similarity according to the user interest correlation matrix includes:

inputting the user interest correlation matrix into a recommendation system, and calculating correlation coefficients among different users;

and calculating the user similarity among different users according to the correlation coefficient.

In some embodiments, the selecting candidate content according to the user similarity for content recommendation includes:

sorting the users according to the user similarity;

selecting a preset number of users as a candidate user set;

acquiring candidate interest content;

obtaining a grading and sorting result of the candidate interest content according to the candidate user set;

and selecting the candidate content for content recommendation according to the grading sorting result and the preset recommendation quantity.

To achieve the above object, a further aspect of the present invention provides a content recommendation apparatus including:

the user behavior information acquisition module is used for acquiring user behavior information;

the characteristic processing module is used for carrying out characteristic processing on the user behavior information to obtain user behavior factor data;

the characteristic analysis module is used for carrying out characteristic analysis on the user behavior factor data to obtain interest content;

the interest correlation calculation module is used for generating a user interest correlation matrix according to the interest content;

the user similarity calculation module is used for calculating the user similarity according to the user interest correlation matrix;

and the content recommendation module is used for selecting candidate content according to the user similarity and recommending the content.

To achieve the above object, another aspect of the present invention provides an electronic device, including:

at least one memory;

at least one processor;

at least one program;

the programs are stored in a memory and a processor executes the at least one program to implement the content recommendation method of the present invention as described above.

To achieve the above object, a further aspect of the present invention proposes a storage medium which is a computer-readable storage medium storing computer-executable instructions for causing a computer to execute:

the content recommendation method as described above.

According to the content recommendation method and device, the electronic device and the storage medium provided by the embodiment of the invention, the user behavior information is acquired, and the user behavior information comprises the following steps: the wireless data information and the track data information are used for carrying out characteristic processing on the user behavior information to obtain user behavior factor data, and the user behavior factor data comprise: the wireless factor data and the track factor data perform characteristic analysis on the user behavior factor data to obtain interest content, a user interest correlation matrix is generated according to the interest content, user similarity is obtained through calculation according to the user interest correlation matrix, candidate content is selected according to the user similarity, and content recommendation is performed. According to the embodiment of the invention, the user behavior information including the wireless data information and the track data information is acquired, and the multiple types of user data are combined for feature extraction, so that the interest content is obtained by acquiring the implicit related information among different data, and the interest content is utilized to carry out corresponding content recommendation, so that the accuracy of content recommendation and the correlation between the recommended content and the user requirements are improved.

Drawings

Fig. 1 is a flowchart of a content recommendation method according to an embodiment of the present invention.

Fig. 2 is a partial flowchart of a content recommendation method according to another embodiment of the present invention.

Fig. 3 is a partial flowchart of a content recommendation method according to another embodiment of the present invention.

Fig. 4 is a partial flowchart of a content recommendation method according to another embodiment of the present invention.

Fig. 5 is a partial flowchart of a content recommendation method according to another embodiment of the present invention.

Fig. 6 is a partial flowchart of a content recommendation method according to another embodiment of the present invention.

Fig. 7 is a block diagram of a content recommendation apparatus according to an embodiment of the present invention.

Fig. 8 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further 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 invention and are not intended to limit the invention.

It should be noted that although functional blocks are partitioned in a schematic diagram of an apparatus and a logical order is shown in a flowchart, in some cases, the steps shown or described may be performed in a different order than the partitioning of blocks in the apparatus or the order in the flowchart.

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 invention belongs. The terminology used herein is for the purpose of describing embodiments of the invention only and is not intended to be limiting of the invention.

First, several terms related to the present invention are analyzed:

location Based Services (LBS): the current position of the positioning equipment is obtained by utilizing various types of positioning technologies, and information resources and basic services are provided for the positioning equipment through the mobile internet. First, the user can determine the spatial position of the user by using a positioning technology, and then the user can acquire resources and information related to the position through the mobile internet. The LBS service integrates various information technologies such as mobile communication, internet, space positioning, position information, big data and the like, and utilizes a mobile internet service platform to update and interact data, and provides various services related to the position for the user according to the position information and the query information of the user and through the network.

Collaborative Filtering algorithm (Collaborative Filtering): the collaborative filtering algorithm is a commonly used recommendation algorithm, finds preference bias of a user based on mining of historical behavior data of the user, and predicts products which the user may prefer to recommend. Collaborative filtering generally finds a small part similar to a target among a large number of users, in the collaborative filtering, the users become a neighbor set, and then an ordering directory is organized as a recommendation according to other things liked by the users in the neighbor set. The method mainly comprises the following steps: recommending according to people with common preferences, recommending similar articles according to favorite articles and comprehensively recommending. Commonly used collaborative filtering algorithms are divided into two categories, user-based collaborative filtering algorithms and item-based collaborative filtering algorithms.

Pearson correlation coefficient (Pearson correlation coefficient): also known as pearson product-moment correlation coefficient, is a method proposed by the british statistician pearson in the 20 th century to calculate a straight-line correlation, which measures the correlation (linear correlation) between two variables X and Y, with a value between-1 and 1.

Nearest Neighbor algorithm (K-Nearest-Neighbor, KNN): KNN classifies by measuring the distance between different characteristic values, and the main idea is as follows: a sample belongs to a class if most of the K most similar samples in feature space (i.e. the nearest neighbors in feature space) belong to this class, where K is typically an integer no greater than 20. In the KNN algorithm, all selected neighbors are objects which are classified correctly, and the method only determines the class of the sample to be classified according to the class of the nearest sample or samples in the classification decision.

Based on this, embodiments of the present invention provide a content recommendation method and apparatus, an electronic device, and a storage medium, which may acquire user behavior information including wireless data information and trajectory data information, and combine multiple types of user data to perform feature extraction, so as to obtain interest content by acquiring implicit related information between different data, and perform corresponding content recommendation using the interest content, thereby improving accuracy of content recommendation and correlation between recommended content and user requirements.

Embodiments of the present invention provide a content recommendation method and apparatus, an electronic device, and a storage medium, which are described in detail with reference to the following embodiments, and first describe a content recommendation method in an embodiment of the present invention.

The embodiment of the invention 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. In other words, artificial intelligence is a comprehensive technique of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a manner similar to human intelligence. Artificial intelligence is the research of the design principle and the realization method of various intelligent machines, so that the machines have the functions of perception, reasoning and decision making. The artificial intelligence technology is a comprehensive subject, and relates to the field of extensive technology, namely the technology of a hardware level and the technology of a software level.

The artificial intelligence infrastructure generally includes 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 invention provides a content recommendation method, and relates to the technical field of artificial intelligence, in particular to the technical field of data mining. The content recommendation method provided by the embodiment of the invention 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, smart watch, or the like; the server can be an independent server, and can also be a cloud server providing basic cloud computing services such as cloud service, a cloud database, cloud computing, a cloud function, cloud storage, Network service, cloud communication, middleware service, domain name service, security service, Content Delivery Network (CDN), big data and artificial intelligence platform and the like; the software may be an application or the like implementing a content recommendation method, but is not limited to the above form.

The invention 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 invention 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 invention 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.

Fig. 1 is an alternative flowchart of a content recommendation method according to an embodiment of the present invention, and the method in fig. 1 may include, but is not limited to, steps S110 to S160.

Step S110, user behavior information is acquired.

In an embodiment, with the rapid popularization of the smart mobile device and the development of the wireless communication technology, the WiFi network achieves an extremely wide range of coverage, has technical features and application values such as low energy consumption, accurate positioning, trajectory analysis and the like, and can reflect the relevant information of the user, so the user behavior information obtained in the embodiment includes wireless data information, that is, the WiFi use relevant information of the user.

In an embodiment, since the relevant information of the user can also be reflected in the moving behavior track of the user, the acquired user behavior information further includes track data information, for example, the track data information may be LBS track data (i.e. information that the user appears in a certain type of place at a certain time and the like can be acquired through the LBS track data).

Some recommendation systems in the related art, such as a news-type text recommendation system (for example, the top of the day), a shopping-type item recommendation system (for example, the commodity), a life service-type service recommendation system (for example, the popular comment), and the like, perform related recommendation by using the location information of the user, such as functions of peripheral search, coupon search, card punching, and the like of the service recommendation system. However, due to the limitation of the location information, multiple types of user data are not combined, and implicit related information is obtained to recommend the corresponding content, so that the accuracy of content recommendation is not high, and the relevance between the recommended content and the user requirements is low. According to the embodiment, the data analysis is performed on the user by acquiring the wireless data information and the track data information and combining the two types of user behavior information, so that important information such as life preference, travel rules and risk exposure of the user is obtained, abundant user images can be established, similar user matching is realized, and corresponding insurance products and services are recommended according to matching results.

In an embodiment, the step S110 of acquiring the user behavior information includes, but is not limited to, the steps S111 to S112:

and step S111, acquiring the original information of the user behavior.

In an embodiment, the obtaining of the user behavior raw information may be obtaining raw wireless data information and raw track data information of the user terminal through user authorization.

And step S112, performing data cleaning and data association on the original user behavior information to obtain user behavior information.

In one embodiment, data cleaning is performed on the user behavior original information, the purpose of the data cleaning is to improve the usability of data, and the loaded user behavior original information is subjected to data cleaning to remove blank data, abnormal data or error data contained in the user behavior original information.

In an embodiment, data cleaning mainly identifies possible error values or abnormal values through a statistical analysis method, for example, through deviation analysis, identifies values that do not comply with a distribution or regression equation, and may also check user behavior raw information through a simple business rule. It should be understood that the method of data cleansing is only an example and is not meant to limit the method of data cleansing in the embodiments of the present invention.

In one embodiment, the data association is mainly to associate user personal information, user insurance information and user behavior original data after data cleaning, and establish a corresponding mapping relation to obtain user behavior information. For example, the corresponding user personal information, user insurance information, wireless data information and track data information can be obtained through the number of the user.

And step S120, performing characteristic processing on the user behavior information to obtain user behavior factor data.

In an embodiment, after obtaining the user behavior information, performing feature processing on the user behavior information, and extracting normalized feature information from the obtained user behavior information to obtain user behavior factor data, where the user behavior factor data corresponds to the user behavior information, in this embodiment, the user behavior factor data includes: radio factor data and trajectory factor data.

Specifically, referring to fig. 2, the specific steps of processing the characteristics of the user behavior information in this embodiment include, but are not limited to, step S121 to step S122:

and step S121, performing characteristic processing on the wireless data information to obtain wireless factor data.

In an embodiment, the radio factor data includes at least one of:

1) local area network name factor data: and extracting the place name from the connection name of the IP local area network in the wireless data information, and obtaining the place information contained in the connection name to form local area network name factor data, such as a catering place, a hotel place, a financial institution place, a hospital place, a market place, a pet store place, a clothing store place, a gymnasium place or a 4S store place.

2) Browsing content factor data: and extracting specific browsed contents from the browsing history record in the wireless data information to form browsed content factor data, such as food related contents, accommodation related contents, financial information related contents, health related contents, fashion information related contents or entertainment information related contents.

3) Browsing duration factor data: and extracting browsing duration corresponding to browsing different contents from the browsing history record in the wireless data information to form browsing duration factor data, wherein the browsing duration factor data comprises the corresponding browsing duration of each type of contents in the browsing content factor data. For example, the browsing duration may be represented by a specific number of minutes, or may be represented by a duration interval, which is not specifically limited herein.

4) And browsing frequency factor data: and setting a preset period, and extracting browsing times corresponding to different contents browsed by the user in the preset period from a browsing history record in the wireless data information to form browsing frequency factor data, wherein the browsing frequency factor data can represent the preference of the user to different types of browsing contents to a certain extent.

5) Browsing time factor data: and acquiring the browsing time of the current type of content browsed by the user from the wireless data information to form browsing time factor data.

It is to be understood that the above-mentioned wireless factor data is only an example, and does not represent the present embodiment to limit the same.

And step S122, performing characteristic processing on the track data information to obtain track factor data.

In one embodiment, the track factor data is mainly used for extracting the place information related to the interest from the track data of the user, for example, the track factor data includes: a dining venue, a hotel venue, a financial institution venue, a hospital venue, a mall venue, a pet store venue, a clothing store venue, a gym venue, a 4S store venue, or the like.

And step S130, performing characteristic analysis on the user behavior factor data to obtain interest content.

In an embodiment, after the wireless factor data and the track factor data are obtained, the two types of user behavior information are combined to perform data analysis, and some information which may be ignored in the traditional client operation mode is mined, so that important information such as life preference, travel rules, risk exposure and the like of a user is obtained and is bound with the existing personal basic information and application information, rich user pictures can be established, similar user matching is realized, corresponding insurance products and services are recommended according to matching results, the vacancy of existing insurance recommendation and innovation services is filled, risks of the client are analyzed, and better insurance services are performed.

Specifically, referring to fig. 3, step S130 includes, but is not limited to, step S131 to step S132:

step S131, performing characteristic analysis according to the user behavior factor data to obtain location correlation content, browsing content correlation content and browsing duration correlation content.

In one embodiment, a user in need of insurance services typically has the following characteristics: 1) insurance demand dependencies can be analyzed from their frequented venues; 2) insurance demand relevance can be analyzed from the browsing content thereof; 3) insurance demand relevance can be analyzed from its browsing duration. Therefore, in this embodiment, feature analysis is performed according to the user behavior factor data to obtain location-related content, browsing content-related content, and browsing duration-related content.

Step S132, generating the interest content according to the location correlation content, the browsing content correlation content and the browsing duration correlation content.

In one embodiment, the interest content is generated according to the location-related content, the browsing content-related content and the browsing duration-related content, for example:

1) insurance demand dependencies can be analyzed from their frequented venues: generating corresponding interest content according to the location-related content, for example, the location-related content is: and if the user frequently visits the gymnasium, generating corresponding interest contents as follows: health risks and sport protectors; for example, the location-related content is: and generating corresponding interest contents as follows by the house property intermediary place: property risk; for example, the location-related content is: according to hotel place information obtained by frequently connecting hotel WiFi information, if the user is judged to belong to a person who may frequently go on a business trip or travel, corresponding interest content is generated as follows: accident hazards and luggage.

2) Insurance requirement dependencies can be analyzed from their browsed content: generating corresponding interest content according to the browsing content related content, and discovering client interest and preference by analyzing the browsing content, for example, the browsing content related content is: the pet is related, and if the specific pet is a cat, the generated corresponding interest content is as follows: cat climbs frame and molar stick.

3) Insurance demand relevance can be analyzed from its browsing duration: and generating corresponding interest content according to the related content of the browsing duration, wherein the longer the browsing duration is, the higher the interest degree of the user in the corresponding content is possibly. For example, if relevant content of paying attention to financing of the user is found according to browsing duration, the corresponding interest content is generated as follows: property risk.

It is to be understood that the above-mentioned generation of the interest content according to the location related content, the browsing content related content and the browsing duration related content is only an example and is not meant to limit the present embodiment.

Step S140, generating a user interest correlation matrix according to the interest content.

In an embodiment, the obtained content of interest includes: interest goods and/or interest services, such as property, accident or health risks, belong to the interest services, while cat climbers, molar rods or luggage all belong to the interest goods.

In an embodiment, referring to fig. 4, obtaining the content of interest further includes steps S141 to S142:

step S141, performing feature selection on the interest content to obtain interest filtering content.

In one embodiment, the strategy for feature selection includes: one or more of a high base class data selection strategy, a saturation data selection strategy and a correlation data selection strategy, wherein the specific data selection strategy is as follows:

1) high base class data selection strategy: and when the strategy for selecting the characteristics is a high base class data selection strategy, deleting the high base class data in the interest content to obtain the interest screening content.

In the fields of data mining and machine learning, the collection and processing of contents are an important link influencing the quality of a data model. In this embodiment, the high-class data refers to some IP address class and time series class data, and since such data has different values and no specific meaning on each interested content, the contribution degree to the recommendation result is not large when subsequently recommending the content. For example, the interest content includes IP address information, even if the IP addresses of the user devices are different, the data does not contribute to content recommendation, and therefore the data belongs to high base class data, and the high base class data in the interest content needs to be deleted to obtain the interest filtering content.

2) Saturation data selection strategy: when the strategy for selecting the features is a saturation data selection strategy, deleting low-saturation data in the interest content to obtain interest screening content;

in one embodiment, when content recommendation is performed on data with low saturation, misleading of a recommendation result is large, and therefore, the misleading needs to be removed. In the operation process of the recommendation system, the problem of unbalanced information saturation of data is often encountered, in this embodiment, the information saturation is defined as the content of the content feature variable contained in the machine learning explaining the target variable, and the more information content, the more likely the content is correctly predicted. Generally, the low data saturation means that data omission is caused or data itself is lost during non-standardized data acquisition due to human factors. Therefore, in this embodiment, for example, data with serious data loss, such as scattered browsing contents or low browsing duration, belongs to data with low saturation, and low saturation data in the interest content needs to be deleted to obtain the interest screening content.

3) Relevance data selection strategy: and when the strategy for selecting the features is a relevance data selection strategy, deleting high-relevance data in the interest content to obtain the interest screening content.

In an embodiment, the high-relevance data indicates that the relevance of the recommendation result corresponding to the content of interest is high, for example, by means of single-factor analysis or relevance analysis, it is analyzed whether the relevance of each piece of data in the content of interest and the corresponding recommendation result is high, if there is high-relevance data, it is indicated that there is a possibility of data leakage, and using this type of content may cause overfitting of the recommendation system, the system bias is high, the prediction effect only for the training content is good, and the prediction accuracy on the test set or the actual data is low, so it is necessary to delete the high-relevance data in the content of interest to obtain the interest screening content.

It can be understood that the three strategies for performing feature selection may be used alternatively or in combination, so as to implement data selection on the interest content according to the strategy for performing feature selection, and obtain the interest filtering content, and the embodiment does not limit the strategy for performing feature selection here.

Step S142, performing a digitization operation on the interest filtering content to obtain a digitized interest content, so as to generate a user interest correlation matrix according to the digitized interest content.

In an embodiment, after the data selection is performed on the interest content to obtain the interest filtering content, the method further includes performing a digitization operation on the interest filtering content, and performing labeling and training on a vertical interest feature of the user, so that the interest filtering content can be used as an input of a recommendation system.

In one embodiment, the digitizing operation comprises: one or more of noise reduction digitization processing and feature cross-selection digitization processing. The specific numerical operation process is as follows:

1) noise reduction digitization processing: and dividing the numerical value in the preset range into determined blocks in a partitioning mode to reduce the interference of noise. For example, some data in the interest screening content is discrete, and if each data is input into the recommendation system for content recommendation, the recommendation system learns inaccurately on limited content, and there is a possibility of under-fitting. For example, the user is labeled by combining personal information and interest content of the user, in order to carry out normalization processing on the label, the value of the label data is in a [0,10] range, then the label data is divided into determined blocks according to the value and input into a recommendation system, and the interference of noise is reduced. In order to avoid interference caused by noise, the embodiment divides the numerical value in the preset range into the determined blocks in a partitioning or binning mode to reduce the interference of the noise, so that the recommendation accuracy of the content recommendation system can be improved.

2) Performing characteristic cross selection digitization treatment: since the data involved in part of the interest screening content is more, but not every feature has an influence on the recommendation result of the recommendation system, and the recognition accuracy of the recommendation system may be reduced if these redundant features are not removed, the embodiment searches out the feature subset through correlation or other methods for determining the feature importance.

It should be understood that the above-mentioned specific procedures of the numerical operation are only illustrative, and the numerical operation in the embodiment of the present invention is not limited to be implemented by the above-mentioned method.

In an embodiment, the digitized interest content is obtained, and a corresponding user interest correlation matrix is obtained by calculating the interest content according to a requirement of a collaborative filtering algorithm, with reference to fig. 5, the method specifically includes, but is not limited to, steps S510 to S520:

step S510: and generating a relevance value of the user for the interest content according to the use frequency.

In one embodiment, to measure the user's interest level in a certain interested item or a certain interested service (i.e. the correlation value of the present embodiment), the correlation value is quantified by calculating the frequency of use:

for example, the usage frequency is judged by counting the number of experiences of the experience ticket in a manner of issuing the experience ticket (e.g., audio/video experience ticket), or the investigation result is obtained in a manner of network questionnaire investigation, and the usage frequency is quantified (e.g., in a manner of scoring) to analyze the relevance value of the user to the content of interest.

Step S520: and generating a user interest correlation matrix according to the correlation value.

In one embodiment, the recommendation system applies a collaborative filtering algorithm based on users, and the embodiment uses an m × n matrix to represent the preference of users for the interesting content, i.e. a user interest correlation matrix, wherein the relevance value of the interesting content of the users is represented by a score, the higher the score is, the more the users like the interesting content, and 0 represents that the users have no correlation with the interesting content. A user is represented by a row in an interest correlation matrix U, a list represents an interest content, and a value U in the matrix_ijA score value of the user i for the interest content j, i.e., a relevance value of the user i for the interest content j (which can be calculated by using the frequency as described above) is represented.

And step S150, calculating according to the user interest correlation matrix to obtain the user similarity.

In an embodiment, after obtaining a user interest correlation matrix representing interest contents of a plurality of users, calculating similarities between different users, and selecting recommended contents for a target user according to the similarities.

In one embodiment, the pearson correlation coefficient is used to represent a measure of similarity of different users to the content of interest. The pearson correlation coefficient is used to measure the correlation (linear correlation) between the two variables X and Y, and the higher the calculated pearson correlation coefficient between the two users is, the higher the accuracy of predicting the content of interest of one user by using the content of interest of the other user is, because the higher the pearson correlation coefficient of the two users is, the more common the two users are, the more the change of one user is known from the change of the other user.

In this embodiment, the pearson correlation coefficient is obtained by calculating the quotient of the covariance and the standard deviation between the two variables (i.e., the users). If the correlation coefficient between two users (represented by variable X and variable Y) is 1 or-1, the larger the absolute value of the pearson correlation coefficient, the stronger the correlation between the two users, and the closer the pearson correlation coefficient is to 1 and-1, meaning that the two variables can be described by straight-line equations, all data points fall on a straight line, and one variable increases as the other increases. A coefficient value of-1 means that all data points fall on a straight line and one variable decreases as the other increases. The closer the pearson correlation coefficient is to 0, meaning that there is no linear relationship between the two variables, the weaker the correlation between the two users.

In one embodiment, the correlation between the pearson correlation coefficient and the liana is as follows:

when the pearson correlation coefficient is 0, there is no relationship between the variable X and the variable Y.

When the Pearson correlation coefficient is between 0.00 and 1.00, the value of the variable X is increased, and the variable Y is also increased, so that a positive correlation is formed.

When the Pearson correlation coefficient is between 0.00 and 1.00, the variable X is decreased, and the variable Y is also decreased, so that the positive correlation is formed.

When the Pearson correlation coefficient is between-1.00 and 0.00, the variable X is increased in value and the variable Y is decreased, so that a negative correlation relationship is formed.

When the Pearson correlation coefficient is between-1.00 and 0.00, the variable X is decreased, and the variable Y is increased, so that the correlation relationship is negative.

And step S160, selecting candidate contents according to the similarity of the users, and recommending the contents.

In an embodiment, the above-mentioned obtaining of user similarities among different users, in order to recommend a content to a certain user, K other users having similar behaviors or preferences to past behavior history of a current user need to be found, the embodiment uses a KNN algorithm to find a nearest neighbor set of the user, where the users in the nearest neighbor set and the user have similar preferences (i.e., interesting content), the KNN algorithm predicts the preferences of the user according to the preferences of the neighbors, for example, user a has an interest point a, user B has three interests a, B, and C, user C likes a and C, then user a is considered similar to users B and C, and the user who likes a also likes C, so that a class C product is recommended to user a, for example, a dangerous type recommendation can be made to the user, and a class C insurance is recommended to user a.

In one embodiment, referring to fig. 6, step S160 includes, but is not limited to, steps S161 to S165:

step S161, user sorting is performed according to the user similarity.

In an embodiment, the users in the nearest neighbor set of the target user obtained by the KNN algorithm are ranked according to user similarity with the target user, for example, the top ranked user is the user most similar to the target user.

Step S162, selecting a preset number of users as a candidate user set.

In an embodiment, the preset number may select 1 to 3 neighbors as a candidate user set according to actual requirements, and the interest of the target user is predicted by using scores of different interest contents in the candidate user set.

Step S163, obtaining candidate interest content;

in an embodiment, the target user does not purchase or see the item or service as the candidate content of interest, for example, the candidate content of interest may also be obtained from historical purchases or historical collections of users in the candidate user set.

And S164, acquiring a grading and sorting result of the candidate interest content according to the candidate user set.

In an embodiment, for candidate interest contents, scoring values of the candidate interest contents by the users in the candidate user set are obtained, the scores are subjected to weighted summation, scoring prediction results (serving as scores of the candidate interest contents) of the candidate interest contents by the target users are obtained through calculation, and the scoring prediction results of each candidate interest content are ranked to obtain scoring ranking results of the candidate interest contents.

And S165, selecting candidate contents for content recommendation according to the grading sorting result and the preset recommendation quantity.

In an embodiment, a preset recommendation number (e.g., 3) may be set according to an actual recommendation number requirement, and candidate content is selected for content recommendation according to a score ranking result of the candidate interest content (e.g., the candidate interest content ranked in the top three is recommended).

In the embodiment, the wireless data information and the track data information are acquired, the characteristics are extracted from the historical behavior information of the user, the browsing behaviors such as common travel places, browsing contents, browsing duration and the like of the user are obtained, then the characteristic information is used for carrying out cross analysis, the interest contents of the user are mined out, an interest set is established, abundant user figures are established, the risk exposure is effectively identified, a user interest correlation matrix for different users is formed, content recommendation is carried out on the target user according to the similarity between the users, the existing insurance recommendation gap is filled up, and the user is effectively helped to carry out risk avoidance and insurance guarantee.

According to the technical scheme provided by the embodiment of the invention, the user behavior information is acquired, and the user behavior information comprises the following steps: the wireless data information and the track data information are used for carrying out characteristic processing on the user behavior information to obtain user behavior factor data, and the user behavior factor data comprise: the wireless factor data and the track factor data perform characteristic analysis on the user behavior factor data to obtain interest content, a user interest correlation matrix is generated according to the interest content, user similarity is obtained through calculation according to the user interest correlation matrix, candidate content is selected according to the user similarity, and content recommendation is performed. According to the embodiment of the invention, the user behavior information including the wireless data information and the track data information is acquired, and the multiple types of user data are combined for feature extraction, so that the interest content is obtained by acquiring the implicit related information among different data, and the interest content is utilized to carry out corresponding content recommendation, so that the accuracy of content recommendation and the correlation between the recommended content and the user requirements are improved.

An embodiment of the present invention further provides a content recommendation apparatus, which can implement the content recommendation method described above, and with reference to fig. 7, the apparatus includes:

a user behavior information obtaining module 710, configured to obtain user behavior information;

the feature processing module 720 is configured to perform feature processing on the user behavior information to obtain user behavior factor data;

the characteristic analysis module 730 is used for carrying out characteristic analysis on the user behavior factor data to obtain interest content;

an interest relevance calculation module 740, configured to generate a user interest relevance matrix according to the interest content;

the user similarity calculation module 750 is configured to calculate user similarity according to the user interest correlation matrix;

and the content recommendation module 760 is configured to select candidate content according to the user similarity and recommend the content.

In an embodiment, the characteristic processing module 720 is further configured to perform characteristic processing on the wireless data information to obtain wireless factor data, where the wireless factor data includes at least one of: the method comprises the steps of obtaining local area network name factor data, browsing content factor data, browsing duration factor data, browsing frequency factor data and browsing time factor data, and performing characteristic processing on track data information to obtain track factor data.

In an embodiment, the feature analysis module 730 is further configured to perform feature analysis according to the user behavior factor data to obtain location-related content, browsing content-related content, and browsing duration-related content, and then generate interest content according to the location-related content, browsing content-related content, and browsing duration-related content.

In an embodiment, the interest correlation calculation module 740 is further configured to generate a correlation value of the user for the interest content according to the usage frequency, and then generate a user interest correlation matrix according to the correlation value.

In an embodiment, the user similarity calculating module 750 is further configured to input the user interest correlation matrix into the recommendation system, calculate a correlation coefficient between different users, and calculate user similarities between different users according to the correlation coefficient.

In an embodiment, the content recommendation module 760 is further configured to perform user sorting according to user similarity, select a preset number of users as a candidate user set, acquire candidate interest content, acquire a score sorting result of the candidate interest content according to the candidate user set, and finally select the candidate content according to the preset recommendation number and the score sorting result for content recommendation.

The specific implementation of the content recommendation apparatus in this embodiment is substantially the same as the specific implementation of the content recommendation method, and is not described herein again.

An embodiment of the present invention further provides an electronic device, including:

at least one memory;

at least one processor;

at least one program;

the programs are stored in a memory and a processor executes the at least one program to implement the present invention to implement the content recommendation method described above. The electronic device can be any intelligent terminal including a mobile phone, a tablet computer, a Personal Digital Assistant (PDA for short), 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 invention;

the memory 802 may be implemented in a ROM (read only memory), a static memory device, a dynamic memory device, or a RAM (random access memory). 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 content recommendation method for executing the embodiments of the present disclosure;

an input/output interface 803 for realizing information input and output;

the communication interface 804 is used for realizing communication interaction between the device and other devices, and can realize communication in a wired manner (such as USB, network cable, and the like) or in a wireless manner (such as mobile network, WIFI, bluetooth, and the like); and

a bus 805 that transfers information between the various components of the device (e.g., the processor 801, memory 802, input/output interfaces 803, and communication 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.

An embodiment of the present invention further provides a storage medium, which is a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are configured to enable a computer to execute the content recommendation method.

According to the content recommendation method, the content recommendation device, the electronic equipment and the storage medium, the user behavior information including the wireless data information and the track data information is obtained, the multiple types of user data are combined to perform feature extraction, so that the relevant information implicit between different data is obtained to obtain the interest content, the interest content is used for performing corresponding content recommendation, and the accuracy of content recommendation and the relevance of the recommended content and the user requirements are improved.

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.

The embodiment described in the embodiment of the present invention is for more clearly illustrating the technical solution of the embodiment of the present invention, and does not constitute a limitation to the technical solution provided in the embodiment of the present invention, and it can be known by those skilled in the art that the technical solution provided in the embodiment of the present invention is also applicable to similar technical problems with the evolution of technology and the occurrence of new application scenarios.

It will be appreciated by those skilled in the art that the solutions shown in fig. 1-6 are not meant to limit embodiments of the present invention, 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 the present 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 invention and in the drawings described above, 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 invention 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 is to be understood that, in the present invention, "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 embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, 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 invention 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 invention may be embodied in the form of a software product, which is stored in a storage medium and includes multiple instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. 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 invention have been described above with reference to the accompanying drawings, and are not intended to limit the scope of the embodiments of the invention. 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 invention are intended to be within the scope of the claims of the embodiments of the present invention.

Claims

1. A method for recommending content, comprising:

performing characteristic processing on the user behavior information to obtain user behavior factor data;

2. The content recommendation method according to claim 1, wherein the user behavior factor data includes: wireless factor data and trajectory factor data; the processing the characteristics of the user behavior information to obtain user behavior factor data includes:

performing characteristic processing on the wireless data information to obtain the wireless factor data, wherein the wireless factor data comprises at least one of the following data: the local area network name factor data, browsing content factor data, browsing duration factor data, browsing frequency factor data and browsing time factor data;

3. The content recommendation method according to claim 1, wherein said performing feature analysis on said user behavior factor data to obtain interest content comprises:

4. The content recommendation method according to claim 1, wherein before generating the user interest correlation matrix from the interest content, the content recommendation method further comprises:

and carrying out numerical operation on the interest screening content to obtain numerical interest content so as to generate the user interest correlation matrix according to the numerical interest content.

5. The content recommendation method according to claim 1, wherein said generating a user interest correlation matrix according to the interest content comprises:

6. The content recommendation method according to any one of claims 1 to 5, wherein said calculating user similarity according to said user interest correlation matrix comprises:

7. The content recommendation method according to claim 6, wherein said selecting candidate content according to the user similarity for content recommendation comprises:

sorting the users according to the user similarity;

selecting a preset number of users as a candidate user set;

acquiring candidate interest content;

8. A content recommendation apparatus characterized by comprising:

9. An electronic device, comprising:

at least one memory;

at least one processor;

at least one program;

the programs are stored in a memory, and a processor executes the at least one program to implement:

the method of any one of claims 1 to 7.

10. A storage medium that is a computer-readable storage medium having stored thereon computer-executable instructions for causing a computer to perform:

the method of any one of claims 1 to 7.