CN109829108B

CN109829108B - Information recommendation method and device, electronic equipment and readable storage medium

Info

Publication number: CN109829108B
Application number: CN201910081528.9A
Authority: CN
Inventors: 陈文石; 王强; 卢文羊; 李春阳
Original assignee: Beijing Sankuai Online Technology Co Ltd
Current assignee: Beijing Sankuai Online Technology Co Ltd
Priority date: 2019-01-28
Filing date: 2019-01-28
Publication date: 2020-12-04
Anticipated expiration: 2039-01-28
Also published as: WO2020155877A1; CN109829108A

Abstract

The invention discloses an information recommendation method and device, electronic equipment and a readable storage medium. The method comprises the following steps: determining an association degree value of a target user and each scene theme based on first user data of the target user; determining a matching degree value of the information to be recommended and the scene theme according to second user data corresponding to the information to be recommended; and determining information to be recommended matched with the target user based on the association degree value and the matching degree value, and sending the information to be recommended to the target user. Therefore, the technical problems that the existing information recommendation method is insufficient in accuracy and diversity and low in attraction to users are solved. The method and the device have the advantages of improving the accuracy and diversity of the recommendation information and improving the user attraction.

Description

Information recommendation method and device, electronic equipment and readable storage medium

Technical Field

The invention relates to the technical field of internet, in particular to an information recommendation method and device, electronic equipment and a readable storage medium.

Background

With the rise of mobile internet services, people can conveniently access the network through mobile terminals To further acquire or customize required services, so that an O2O (Online To Offline/Online To Offline) mode is in force. The essence of the mode is that the user and the service can be found more conveniently, and the user can select the required offline service on line at any time; and information recommendation can be carried out on the user through mining of the user portrait and the merchant information, so that the user experience is improved, and the merchant can find the customer.

However, the core idea of the existing information recommendation scheme is to mine similar users or similar information to be recommended, and recommend the similar users or recommend the similar information to be recommended to the users. Therefore, the existing information recommendation method mainly aims at optimizing a recommendation algorithm, namely how to more accurately match users and merchants, and does not pay attention to the actual situation of the users, so that the accuracy and diversity of recommendation results determined based on the existing information recommendation method are insufficient, and the attractiveness to the users is low.

Disclosure of Invention

The invention provides an information recommendation method, an information recommendation device, an electronic device and a readable storage medium, which are used for partially or completely solving the problems related to the information recommendation process in the prior art.

According to a first aspect of the present invention, there is provided an information recommendation method, including:

determining an association degree value of a target user and each scene theme based on first user data of the target user;

determining a matching degree value of the information to be recommended and the scene theme according to second user data corresponding to the information to be recommended;

and determining information to be recommended matched with the target user based on the association degree value and the matching degree value, and sending the information to be recommended to the target user.

According to a second aspect of the present invention, there is provided an information recommendation apparatus comprising:

the system comprises a correlation degree determining module, a correlation degree determining module and a scene theme matching module, wherein the correlation degree determining module is used for determining a correlation degree value between a target user and each scene theme based on first user data of the target user;

the matching degree determining module is used for determining a matching degree value of the information to be recommended and the scene theme according to second user data corresponding to the information to be recommended;

and the recommendation information matching module is used for determining information to be recommended matched with the target user based on the association degree value and the matching degree value and sending the information to be recommended to the target user.

According to a third aspect of the present invention, there is provided an electronic apparatus comprising:

processor, memory and computer program stored on the memory and executable on the processor, characterized in that the processor implements the aforementioned information recommendation method when executing the program.

According to a fourth aspect of the present invention, there is provided a readable storage medium, wherein instructions, when executed by a processor of an electronic device, enable the electronic device to perform the aforementioned information recommendation method.

According to the information recommendation method, the association degree value of the target user and each scene theme can be determined based on the first user data of the target user; determining a matching degree value of the information to be recommended and the scene theme according to second user data corresponding to the information to be recommended; and determining information to be recommended matched with the target user based on the association degree value and the matching degree value, and sending the information to be recommended to the target user. Therefore, the technical problems that the existing information recommendation method is insufficient in accuracy and diversity and low in attraction to users are solved. The method and the device have the advantages of improving the accuracy and diversity of the recommendation information and improving the user attraction.

The foregoing description is only an overview of the technical solutions of the present invention, and the embodiments of the present invention are described below in order to make the technical means of the present invention more clearly understood and to make the above and other objects, features, and advantages of the present invention more clearly understandable.

Drawings

Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:

FIG. 1 is a flow chart illustrating the steps of a method for information recommendation according to one embodiment of the present invention;

FIG. 2 is a flow chart illustrating the steps of a method for information recommendation according to one embodiment of the present invention;

fig. 3 is a schematic structural diagram of an information recommendation apparatus according to an embodiment of the present invention; and

fig. 4 is a schematic structural diagram of an information recommendation apparatus according to an embodiment of the present invention.

Detailed Description

Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

Example one

The information recommendation method provided by the embodiment of the invention is described in detail.

Referring to fig. 1, a flowchart illustrating steps of an information recommendation method according to an embodiment of the present invention is shown.

Step 110, determining the association degree value of the target user and each scene topic based on the first user data of the target user.

In the embodiment of the invention, when recommending information to a user, the matching degree of the recommended information and the scene where the corresponding user is located is improved, so that the matching accuracy between the recommended information and the corresponding user is further improved. For a target user needing information recommendation, first user data of the target user can be obtained, and association degree values of the corresponding target user for all scene topics are determined based on the first user data.

The first user data may include any available data related to the target user. For example, User portrait data, UGC (User Generated Content) data, PGC (professional-produced Content) data, OGC (professional-produced Content) data, User location data, current POI (Point of Interest) data, and the like. The POI data may in turn include, but is not limited to, POI tags, UGC content of the POI, POI-associated articles, and the like. In the embodiment of the present invention, specific content included in the first user data may be preset according to a requirement, and in the embodiment of the present invention, the first user data may be obtained by any available method, which is not limited to this embodiment of the present invention.

The scenario topic can be predefined by any available method according to the requirement, for example, the scenario topic can be set by an expert-defined method, or the scenario topic can be mined by data mining through a large amount of reference data, and the like. And the representation of the scenario topic may include that one scenario topic is represented by at least one word, and so on.

For example, a certain scenario theme may be set as a scenario theme of romantic appointments, or a certain scenario theme may be set as a scenario theme of dinner parties, and so on.

In addition, in the embodiment of the present invention, any available method may be adopted to determine the association degree of the corresponding target user to each scenario topic based on the first user data, which is not limited in the embodiment of the present invention. For example, the degree of matching of the first user data with each scenario topic may be used as the association degree value of the corresponding target user and the corresponding scenario topic, and so on.

And step 120, determining a matching degree value of the information to be recommended and the scene theme according to second user data corresponding to each piece of information to be recommended.

In practical application, second user data of different users aiming at certain information to be recommended can represent corresponding information to be recommended to a certain extent. For example, if the second user data corresponding to a certain piece of information to be recommended includes a place selected by a plurality of users to be recommended as an appointment place, it may be inferred that the matching degree between the corresponding piece of information to be recommended and the scene theme of the appointment is high.

Therefore, in the embodiment of the present invention, in order to recommend information to be recommended, which has a high degree of matching with the situation where the target user is located, the matching degree value between the corresponding information to be recommended and each situation topic may be determined according to the second user data corresponding to each information to be recommended. Specifically, the matching degree value of the information to be recommended and the scenario topic may be determined in any available manner based on the second user data corresponding to the information to be recommended, which is not limited in this embodiment of the present invention.

Furthermore, in the embodiment of the present invention, the information to be recommended may be any information that can be recommended to the user, and for example, the information to be recommended may include, but is not limited to, recommendation information for at least one item, recommendation information for at least one place, recommendation information for at least one web page, and the like. The specific information to be recommended may be set according to the requirement, and the embodiment of the present invention is not limited.

In addition, the second user data corresponding to each piece of recommendation information may include second user data of a related user who performs operations such as purchasing, browsing, sharing, and the like on the corresponding piece of recommendation information. And the second user data may also include, but is not limited to, the aforementioned user profile data, UGC data, PGC data, OGC data, user positioning data, POI data, and the like. The specific data type included in the second user data may be preset according to a requirement, and the embodiment of the present invention is not limited thereto.

Step 130, determining information to be recommended matched with the target user based on the association degree value and the matching degree value, and sending the information to be recommended to the target user.

As described above, the association degree value may represent the degree of correlation between the target user and each contextual theme, and the matching degree may represent the degree of correlation between each piece of information to be recommended and each contextual theme. The method and the device aim to recommend corresponding information to the target user based on the situation of the target user. Therefore, the information to be recommended matched with the target user can be determined based on the association degree value and the matching degree value, and then the corresponding information to be recommended can be sent to the corresponding target user.

The method comprises the steps that a contextual theme corresponding to a target user can be determined based on a correlation degree value, information to be recommended matched with the corresponding contextual theme can be determined based on a matching degree value, and then the information to be recommended matched with the target user is obtained; or in the embodiment of the invention, the association degree of the corresponding information to be recommended and the target user can be obtained directly based on the association degree value and the matching degree value, and the information to be recommended matched with the target user is selected from the information to be recommended based on the association degree, and the like.

Furthermore, in the embodiment of the present invention, the determined information to be recommended may be sent to the target user by any available method, which is not limited in the embodiment of the present invention.

Example two

Referring to fig. 2, a flowchart illustrating steps of an information recommendation method according to an embodiment of the present invention is shown.

And step 210, performing data mining based on third user data of the referenceable user, and extracting a specific scene theme.

In practical application, if the type of the scenario theme is customized by the user, the scenario themes defined by different requirements of different users are different, and different scenario themes may be set for different users in the same scenario, so that the scenario themes are easily confused, and the accuracy of recommendation information is easily influenced. Therefore, in the embodiment of the present invention, in order to avoid the above situation, different scene themes may be uniformly defined in advance. Specifically, data mining can be performed based on third user data of the referenceable user, and a specific scene topic is extracted. The third user data that may be referred to by the user may specifically include any user data that may be obtained for contextual topic mining. Such as user data for users on a group purchase, take-away, etc., platform, etc.

And the third user data may also include, but is not limited to, the aforementioned user profile data, UGC data, PGC data, OGC data, user positioning data, POI data, and the like. The specific data type included in the third user data may be preset according to a requirement, and the embodiment of the present invention is not limited thereto.

Moreover, in the embodiment of the present invention, based on the third user data of the referenceable user, data mining may be performed by any available method, so as to extract a specific scenario topic, which is not limited in the embodiment of the present invention.

Optionally, in an embodiment of the present invention, the step 210 further includes:

and a substep 211, performing vectorization processing on the third user data of the referenceable user to obtain a multi-dimensional word vector corresponding to the third user data.

In the embodiment of the present invention, in order to facilitate mining of the scene topic from the third user data of the referenceable user, vectorization processing may be performed on the third user data of the referenceable user, so as to obtain a multi-dimensional word vector corresponding to the third user data. Specifically, the vectorization processing may be performed on the third user data of the referenceable user by any available vector processing method, which is not limited in this embodiment of the present invention.

For example, the third user data of the referenceable user may be vectorized by a doc2vec model, or may be vectorized by a word2vec model, and so on. Furthermore, the word2vec model may include a skip-gram model, a continuous bag-of-words (CBOW) model, and so on.

Optionally, in an embodiment of the present invention, the substep 211 further may include:

a substep 2111 of performing word segmentation processing on the third user data of the referenceable user;

in practical applications, in order to perform vectorization processing on the third user data, word segmentation processing may be performed on the third user data of the referent user, and specifically, any available word segmentation processing method may be used to perform word segmentation processing on the third user data of the referent user, which is not limited to the embodiment of the present invention.

A substep 2112 of removing invalid words in the third user data after the word segmentation processing, and extracting feature words in the third user data, wherein the invalid words comprise at least one of stop words and high-frequency words;

in practical application, partial words in the user data do not have any effect on determination of scene topics, and the words can be defined as invalid words, so that the invalid words can be not considered when a multidimensional vector is generated, and then the invalid words in the third user data after word segmentation can be further removed, and further characteristic words are extracted. The invalid Words may include, but are not limited to Stop Words (Stop Words), high frequency Words, and the like.

The specific content of the stop word can be preset according to the requirement, and the embodiment of the invention is not limited. For example, stop words may be placed with reference to existing stop word lists, such as a Hadamard stop word list, a Baidu stop word list, and so forth. Moreover, in the embodiment of the invention, words which are not helpful or meaningless to the service can be specially arranged according to the service requirement aiming at specific services. Even the stop word may include a stop "sentence", such as "this user has no comments on the e-commerce". "may also be set as a stop word.

For each participle obtained after the participle processing, after removing the invalid word therein, the remaining participles can be directly used as the feature word; or, each participle obtained after removing the invalid word may be further filtered to obtain a feature word, and a specific filtering policy may be preset according to a requirement, which is not limited in the embodiment of the present invention.

Moreover, the word frequency range corresponding to the high-frequency word may also be preset according to the requirement, and the embodiment of the present invention is not limited.

Sub-step 2113, constructing a multi-dimensional word vector of said third user data based on said feature words.

After the feature words are extracted, a multi-dimensional word vector of third user data of the reference user can be constructed and obtained based on the feature words.

Different vectorization models are adopted, and the form of the obtained multi-dimensional word vector can be different. For example, the multi-dimensional word vector may be in one-hot encoded form, or in TF-IDF (term frequency-inverse document frequency) form, and so on. The specific setting may be performed in advance according to the requirement, and the embodiment of the present invention is not limited.

And a substep 212 of obtaining the scene topic through a topic model based on the multi-dimensional word vector.

After the multidimensional word vector is obtained, a scene topic can be obtained through topic model mining based on the multidimensional word vector. The topic model may be any available topic model, such as a LDA (document topic Allocation) topic model, a Sentence LDA topic model, a Copula LDA topic model, and so on.

Step 220, defining a specific situation topic according to the preset situation judgment condition.

In addition, in the embodiment of the present invention, in order to avoid that the types of the scenario topics obtained by data mining are not comprehensive enough or accurate enough, a specific scenario topic may be defined according to a preset scenario judgment condition. The condition for determining the scene may be preset according to the requirement, and the embodiment of the present invention is not limited thereto. For example, it is possible to set different situation determination conditions in consideration of the actual intention and scene of the user, and to define the context of romantic appointments, define the context of dinner parties, and the like.

It should be noted that, in the embodiment of the present invention, the scenario topic may be obtained based on step 210 and/or step 220, which is not limited in the embodiment of the present invention.

Optionally, in this embodiment of the present invention, the scenario topic is characterized by a scenario topic word and/or a topic related word under the scenario topic word category.

In the embodiment of the present invention, in order to accurately characterize each scenario topic, it may be set that each scenario topic is characterized by at least one scenario topic word and/or at least one topic related word under a corresponding scenario topic word category.

For example, for a certain scene topic, the scene topic may be "appointment", and the category under the scene topic contains topic related words such as "romance", "candlelight dinner", "fresh flower", and so on.

Step 230, obtaining a feature value of each scenario topic word of the first user data according to a probability of the topic related word contained in each piece of first user data of the target user under each scenario topic word.

In the embodiment of the present invention, taking the LDA model as an example, the multidimensional word vector is input into the LDA model, so that a word-topic matrix can be obtained, where a word can be understood as a topic related word in the embodiment of the present invention, and a topic can be understood as a scenario topic word in the present invention. In addition, based on the topic model, the probability of each topic related word under the corresponding scene topic word can be obtained.

In the embodiment of the present invention, in order to obtain the degree of association of the target user to a certain scenario topic, and because the target user may correspond to at least one piece of first user data, and the content specifically included in different pieces of first user data corresponding to the same user may also be different, in the embodiment of the present invention, each piece of first user data of the target user may be taken as a unit, and the feature value of each piece of first user data for each scenario topic word is obtained according to the probability of the topic-related word included in each piece of first user data under each scenario topic word.

The corresponding relationship between the feature value of a certain first user data and the probability of the topic related word contained in the corresponding first user data under the corresponding topic word may be preset according to the requirement, and the embodiment of the present invention is not limited thereto.

Optionally, in an embodiment of the present invention, the step 230 may further include:

a substep 231, extracting topic related words in the first user data aiming at each piece of first user data;

similarly, since the first user data includes more contents, and at least one topic related word may exist therein, and may further include other unrelated words, in the embodiment of the present invention, in order to determine the feature value of the contextual topic word corresponding to each piece of first user data, the topic related word in each piece of first user data may be extracted first for each piece of first user data.

Moreover, in the embodiment of the present invention, in order to improve efficiency and accuracy of extracting topic related words, each piece of first user data may also be preprocessed. The preprocessing may include word segmentation, invalid word removal, and so on. The invalid words may also include the high frequency words and stop words, etc.

In the substep 232, for each context topic word, summing the probabilities of the topic related words under the context topic word to obtain a feature value of the first user data for the context topic word.

After obtaining the topic related words included in the first user data, in order to determine the feature value of the first user data corresponding to each contextual topic, for each contextual topic, the probabilities of the topic related words extracted from the corresponding first user data under the corresponding contextual topic words may be summed to obtain the feature value of the corresponding first user data for the corresponding contextual topic words.

Moreover, in the embodiment of the present invention, after the topic related words are extracted from the first user data, the contextual topic words corresponding to each extracted topic related word can be known, and then the contextual topic words corresponding to the corresponding first user data can be obtained. If the first user data does not contain all topic related words under a certain scenario topic word, the feature value of the first user data for the scenario topic word is zero. Therefore, in the embodiment of the present invention, the extracted probabilities of the topic related words corresponding to the corresponding topic terms under the corresponding topic terms can be summed only for each topic term corresponding to the corresponding first user data, so as to obtain the feature value of the corresponding first user data for the corresponding topic terms.

Assuming that Fuik represents a feature value of the target user u for the first user data a of the information i for the scenario topic word k, the feature value may be calculated as follows:

wherein n is the number of topic related words extracted from the first user data a and belonging to the classification of the scene topic words k; f. of_uiktIndicating the probability of the topic related word t under the scenario topic word k, and if the topic related word t is not contained under the scenario topic word k, f_uiktMay be 0.

Step 240, obtaining the association degree value of the target user to the scenario topic represented by the scenario topic word based on the feature value of the first user data under each scenario topic word.

In the embodiment of the present invention, after determining the feature value of each piece of first user data of the target user under each scenario topic word, the association degree value of the corresponding target user to each scenario topic may be further obtained based on the feature values of all pieces of first user data of the target user under each scenario topic word.

The correspondence between the characteristic value and the degree of association value may be preset according to a requirement, and the embodiment of the present invention is not limited thereto.

For example, an association degree value of a target user for a certain scenario topic may be set, which is a feature value average value of each piece of first user data of the target user for a scenario topic word of a corresponding scenario topic; or the association degree value of the target user for a certain scenario topic may be set, a weighted average value of the feature values of each piece of first user data of the target user for the scenario topic words of the corresponding scenario topic may be set, and the weight may be preset according to the requirement, and the like.

Optionally, in an embodiment of the present invention, the step 240 may further include:

substep 241, for each scenario topic word, obtaining a feature value and a value of all first user data of the target user for the scenario topic word;

as described above, each piece of first user data of the target user may characterize the current environment to some extent, and therefore, in the embodiment of the present invention, in order to determine the current association degree of the target user to each scenario topic, for each scenario topic, the feature value and the value of all the first user data of the target user for the scenario topic may be obtained separately for each scenario topic.

For example, for the scenario topic word k, it is assumed that the feature value of the first user data 1 for the scenario topic word k is n1, the feature value of the first user data 2 for the scenario topic word k is n2, and the feature value of the first user data 3 for the scenario topic word k is n 3. At this time, the sum of feature values of all the first user data of the target user for the scene topic word k may be obtained as n1+ n2+ n 3.

It should be noted that all the first user data may be all the first user data of the target user, which can be obtained, or all the first user data obtained according to the preset first user data obtaining condition, and so on. The first user data obtaining condition may be preset according to a requirement, and the embodiment of the present invention is not limited thereto. For example, the first user data acquisition condition may be set to acquire first user data whose time difference from the release time to the current time is within a preset time difference range, or acquire first user data for a preset information type, or the like.

And a substep 242, obtaining a ratio of the feature value and the value to the quantity of all the first user data, and obtaining an association degree value of the target user for the scenario topic represented by the scenario topic word.

In order to comprehensively consider the characterization effect of each first user data on the scenario where the target user is located, in the embodiment of the present invention, the average value of the feature values of each first user data for the same scenario topic may be used as the association degree value of the first user data for the corresponding scenario topic, and then the ratio of the feature value and the value of a certain scenario topic to the number of all first user data may be obtained at this time, so as to obtain the association degree value of the target user for the scenario topic characterized by the scenario topic.

For example, for the above scenario topic word k, the sum of feature values of all the first user data of the target user for the scenario topic word k is n1+ n2+ n3, and all the first user data at this time specifically include three pieces of first user data, so that the association degree of the scenario topic characterized by the target user for the scenario topic word k can be obtained as (n1+ n2+ n 3)/3.

Optionally, in this embodiment of the present invention, the correlation degree value includes a weighted sum of the short-term correlation degree value and/or the long-term correlation degree value; and the weight value of the short-term association degree value is greater than the weight value of the long-term association degree value.

In addition, in practical application, a user may be interested in a certain scenario only within a period of time, or may be interested in a certain scenario for a long time, so that the interest of the user can be divided into a short-term interest and a long-term relevance degree value according to the timeliness, and the relevance degree value of the corresponding target user to each scenario topic may also include a weighted sum of the short-term relevance degree value and/or the long-term relevance degree value, wherein the weights of the short-term relevance degree value and the long-term relevance degree value may be preset according to requirements, and the embodiment of the present invention is not limited. However, in practical application, the general short-term interest can better represent the current interest of the target user, so that the weight of the short-term relevance degree value can be set to be greater than the weight of the long-term relevance degree value, and the sum of the weight of the short-term relevance degree value and the weight of the long-term relevance degree value can be set to be 1.

Then the relevance value of the target user u for a certain scene topic at this time can be expressed as:

in the formula

Representing the short-term relevancy values of the target users for the respective contextual theme,

representing a long-term relevancy value of the target user for the respective contextual theme. Wherein alpha and beta are weights of the short-term correlation degree value and the long-term correlation degree value respectively.

The short-term relevance degree value may be a relevance degree value corresponding to first user data of a target user within a preset short-term period of time, and is usually a dynamic interest in a shorter time, while the long-term interest may be an interest within at least one preset longer historical period of time, and is mainly determined according to a user attribute and a long-term preference, for example, the user is a mother of a newborn and has a potential interest in a family category in a long term. The type of the user attribute corresponding to the specific long-term interest and the short-term association degree value of each preset shorter historical time period can be preset according to requirements, and the embodiment of the invention is not limited.

For example, the short-term relevance value may be set as a relevance value corresponding to first user data of a target user in a preset time period before the current time, and the long-term interest may be set as an interest of a basic attribute of the user, such as a preference of a young woman for beauty and make-up, a preference of a young man for sports and fitness, and the like, according to the interest level of the user in the current hour, the current day, or other short time periods, or the current user may repeatedly click and browse a hot pot merchant, and we may also determine that the user is potentially interested in a hot pot category. Or the long-term correlation degree value can be directly set to be the sum of the short-term correlation degree values in at least one preset historical time period.

Step 250, aiming at each piece of information to be recommended, each piece of second user data of each referenceable user aiming at the information to be recommended is obtained.

In the embodiment of the present invention, after determining the association degree value of the target user to each contextual theme, in order to recommend suitable information to be recommended to the target user, a matching degree value between each piece of information to be recommended and each contextual theme needs to be determined. Specifically, the second user data corresponding to a certain piece of information to be recommended may be referred to determine the matching degree between the information to be recommended and each scenario topic, so that, first, for each piece of information to be recommended, each piece of second user data of each referenceable user for the information to be recommended may be acquired. The referenceable users are also referenceable users corresponding to all the second user data.

For example, for information to be recommended i, assuming that the second user data of the referenceable user u1 for the information to be recommended i is s1, and the second user data of the referenceable user u2 for the information to be recommended i is s2, then the second user data s1 and s2 described above may be acquired for the information to be recommended i.

Step 260, for each scenario topic word, obtaining a matching degree value of the information to be recommended to the scenario topic represented by the scenario topic word according to the feature value of the second user data for the scenario topic word and the quantity of the second user data corresponding to the information to be recommended.

The corresponding relationship between the matching degree value and the characteristic value and the user data quantity can be preset according to requirements, and the embodiment of the invention is not limited. For example, the matching degree value of a certain piece of information to be recommended to a scene topic represented by a certain scene topic word may be set to be a ratio of the feature value and the value of all second user data corresponding to the corresponding information to be recommended to the corresponding scene topic word to the number of all second user data corresponding to the corresponding information to be recommended.

For example, assume that for a scenario topic word k, the second user data s1 for information i to be recommended has a feature value f for the scenario topic word_uikThen, the information i to be recommended represents the scene subject term k at this timeThe matching degree value of the scene theme is as follows:

T_ik＝∑_uf_uik/|C_i|

therein, sigma_uf_uikRepresenting the characteristic value and the value, | C, of all second user data corresponding to the information i to be recommended aiming at the corresponding scene subject term k_iAnd | represents the quantity of all second user data corresponding to the information i to be recommended.

In addition, it should be noted that, in the embodiment of the present invention, the matching degree between each piece of information to be recommended and each scene topic may also be predetermined, and since the second user data is continuously updated, a preset time period may also be used as a time interval, and the second user data of the current reference user may be periodically obtained to re-determine the matching degree between each piece of information to be recommended and each scene topic.

Moreover, if the association degree value of the target user to each scenario topic is determined currently, then M scenario topics with the highest association degree value may be selected, and then only the matching degree between each piece of information to be recommended and the M scenario topics may be obtained, without obtaining the matching degree between each piece of information to be recommended and each scenario topic.

And 270, performing normalization processing on the association degree value and the matching degree value.

In addition, in the embodiment of the present invention, in order to unify the degree of association of the context-related information of the item, the degree of association value and the matching degree value may be normalized. The specific normalization processing method may be preset according to the requirement, and the embodiment of the present invention is not limited.

Step 280, obtaining the similarity between the target user and the information to be recommended based on the association degree value and the matching degree value.

After the association degree value of the target user to each scenario topic and the matching degree value of each piece of information to be recommended and each scenario topic are obtained, the information to be recommended, which is interested by the target user, can be further screened from the information to be recommended. At this time, the similarity between the target user and each piece of information to be recommended may be further obtained based on the association degree value and the matching degree value. Specifically, the similarity between the target user and each piece of information to be recommended may be obtained by any available similarity determination method, and the embodiment of the present invention is not limited thereto. For example, euclidean Distance (euclidean Distance) Similarity, Manhattan Distance (Manhattan Distance) Similarity, Minkowski Distance (Minkowski Distance) Similarity, Cosine Similarity, and the like.

Optionally, in an embodiment of the present invention, the similarity includes a cosine similarity.

If the similarity is cosine similarity, the similarity between the target user u and the information i to be recommended is

In the formula I_uInterest level vector of target user, I_u＝[I_u1,I_u2,I_u3,...,I_uk]，T_iVector of degree of match, T, representing item i_i＝[T_i1,T_i2,T_i3,...,T_ik]，k∈[1，K]In which I_ukRepresents the degree of association value, T, of the target user u to the k-th scene subject_ikAnd the matching degree value of the information to be recommended to the kth scene subject is represented, and K represents the total number of the scene subjects.

Step 290, selecting a preset number of pieces of information to be recommended with the highest similarity to the target user as the recommendation information of the target user, and sending the recommendation information to the target user.

After the similarity between the target user and each piece of information to be recommended is determined, a preset number of pieces of information to be recommended with the highest similarity with the target user can be selected as the recommendation information of the target user, and the recommendation information is sent to the target user. The preset number may also be preset according to the requirement, and the embodiment of the present invention is not limited.

Optionally, in an embodiment of the present invention, the user data includes at least one of user original content data and user portrait data.

Step 2110, determining a target scene theme matched with the target user according to the association degree value.

And step 2120, displaying the information to be recommended according to the target scene theme.

Optionally, in this embodiment of the present invention, the step 2120 further includes: and displaying recommendation information related to the target scene theme in the information to be recommended according to the target scene theme, wherein the recommendation information comprises at least one of recommendation reason, picture information, video information and text information.

In addition, in the embodiment of the invention, the corresponding information to be recommended can be displayed in an individualized manner according to the situation where the target user is located, specifically, the target situation theme matched with the target user can be determined according to the determined association degree value, and further, the selected information to be recommended can be displayed according to the target situation theme. Specifically, the part of the information to be recommended related to the target scenario theme may be preferentially displayed, or if the current scenario of the target user is in a wireless network environment, the selected information to be recommended may be displayed in a manner of animation or high-definition pictures, or if the current scenario of the target user is in an outdoor environment, the selected information to be recommended may be displayed in a manner of voice, and so on.

Preferably, recommendation information related to the target contextual theme in the selected information to be recommended may be presented, where the recommendation information may include, but is not limited to, at least one of a reason for recommendation matching the target contextual theme, picture information, video information, and text information.

Moreover, in the embodiment of the invention, data mining can be carried out based on third user data of a referenceable user, and a specific scene theme is extracted; and/or defining a specific situation theme according to a preset situation judgment condition. Vectorizing third user data of the referenceable user to obtain a multi-dimensional word vector corresponding to the third user data; and obtaining the scene theme through a theme model based on the multi-dimensional word vector. Therefore, the comprehensiveness and the accuracy of the scene theme can be improved, and the accuracy of the recommendation information and the user attraction are further improved.

In addition, in the embodiment of the present invention, the scenario topic is characterized by a scenario topic word and/or a topic related word under the category of the scenario topic word. Moreover, the feature value of each scene subject term of the first user data can be obtained according to the probability of the subject related term contained in each piece of first user data of the target user under each scene subject term; and acquiring the association degree value of the target user to the scene topic represented by the scene topic word based on the characteristic value of the first user data under each scene topic word. Further, for each piece of first user data, extracting topic related words in the first user data; and for each scene subject term, summing the probability of the subject related term under the scene subject term to obtain the characteristic value of the first user data for the scene subject term. For each scene topic word, acquiring a characteristic value and a value of all first user data of the target user for the scene topic word; and acquiring the ratio of the characteristic value and the value to the quantity of all the first user data to obtain the association degree value of the target user for the scene topic represented by the scene topic word. And aiming at each piece of information to be recommended, each piece of second user data of each referenceable user aiming at the information to be recommended is obtained; and for each scene topic word, acquiring a matching degree value of the information to be recommended to the scene topic represented by the scene topic word according to the feature value of the second user data to the scene topic word and the quantity of the second user data corresponding to the information to be recommended. Acquiring the similarity between the target user and the information to be recommended based on the association degree value and the matching degree value; and selecting a preset amount of information to be recommended with the highest similarity to the target user as the recommendation information of the target user, and sending the recommendation information to the target user. Therefore, the efficiency and the accuracy of obtaining the recommendation information can be further improved.

Further, in the embodiment of the present invention, a target scenario topic matched with the target user may also be determined according to the association degree value; and displaying the information to be recommended according to the target scene theme. According to the target scene theme, recommending information related to the target scene theme in the information to be recommended is displayed; the recommendation information comprises at least one of recommendation reason, picture information, video information and character information. Therefore, personalized display of the information to be recommended is realized, and the user attraction of the recommended information is further improved.

For simplicity of explanation, the method embodiments are described as a series of acts or combinations, but those skilled in the art will appreciate that the embodiments are not limited by the order of acts described, as some steps may occur in other orders or concurrently with other steps in accordance with the embodiments of the invention. Further, those skilled in the art will appreciate that the embodiments described in the specification are presently preferred and that no particular act is required to implement the invention.

EXAMPLE III

The embodiment of the invention provides an information recommendation device.

Referring to fig. 3, a schematic structural diagram of an information recommendation apparatus in an embodiment of the present invention is shown.

The association degree determining module 310 is configured to determine, based on the first user data of the target user, association degree values of the target user and the respective scene topics.

The matching degree determining module 320 is configured to determine a matching degree value between the information to be recommended and the scene topic according to second user data corresponding to each piece of information to be recommended.

And the recommendation information matching module 330 is configured to determine information to be recommended matched with the target user based on the association degree value and the matching degree value, and send the information to be recommended to the target user.

Example four

The embodiment of the invention provides an information recommendation device.

Referring to fig. 4, a schematic structural diagram of an information recommendation apparatus in an embodiment of the present invention is shown.

And the scenario topic mining module 410 is configured to perform data mining based on third user data of the referenceable user to extract a specific scenario topic.

Optionally, in this embodiment of the present invention, the scenario topic mining module 410 further includes:

the vectorization processing submodule is used for vectorizing the third user data of the referent user to obtain a multi-dimensional word vector corresponding to the third user data;

and the scene topic mining submodule is used for obtaining the scene topic through a topic model based on the multi-dimensional word vector.

Optionally, in an embodiment of the present invention, the vectorization processing sub-module includes:

the word segmentation processing unit is used for carrying out word segmentation processing on the third user data of the referenceable user;

the feature word extraction unit is used for removing invalid words in the third user data after word segmentation processing and extracting feature words in the third user data, wherein the invalid words comprise at least one of stop words and high-frequency words;

and the multi-dimensional word vector construction unit is used for constructing a multi-dimensional word vector of the third user data based on the feature words.

The scenario theme definition module 420 is configured to define a specific scenario theme according to a preset scenario judgment condition.

The association degree determining module 430 is configured to determine, based on the first user data of the target user, association degree values of the target user and the respective scene topics.

The association degree determining module 430 may further include:

a feature value obtaining submodule 431, configured to obtain a feature value of each context topic word for the first user data according to a probability of a topic related word included in each piece of first user data of the target user under each context topic word;

the association degree determining sub-module 432 is configured to obtain, based on the feature values of the first user data under each scenario topic word, an association degree value of the target user to a scenario topic represented by the scenario topic word.

Optionally, in this embodiment of the present invention, the characteristic value obtaining submodule 431 further may include:

the topic related word extracting unit is used for extracting topic related words in the first user data aiming at each piece of first user data;

and the characteristic value acquisition unit is used for summing the probability of the theme related words under the scene theme words aiming at each scene theme word to obtain the characteristic value of the first user data aiming at the scene theme words.

Optionally, in this embodiment of the present invention, the association degree determining sub-module 432 further includes:

the characteristic value summation unit is used for acquiring the characteristic value and the value of all the first user data of the target user for each scene subject term;

and the association degree determining unit is used for acquiring the ratio of the characteristic value and the value to the number of all the first user data to obtain the association degree value of the target user for the scene topic represented by the scene topic word.

The matching degree determining module 440 is configured to determine a matching degree value between the information to be recommended and the scene topic according to the second user data corresponding to each piece of information to be recommended.

In this embodiment of the present invention, the matching degree determining module 440 further includes:

the user data obtaining sub-module 441 is configured to, for each piece of information to be recommended, obtain each piece of second user data of each referent user for the piece of information to be recommended;

the matching degree determining sub-module 442 is configured to, for each context topic word, obtain a matching degree value of the information to be recommended to a context topic represented by the context topic word according to the feature value of the second user data to the context topic word and the quantity of the second user data corresponding to the information to be recommended.

And the recommendation information matching module 450 is configured to determine information to be recommended matched with the target user based on the association degree value and the matching degree value, and send the information to be recommended to the target user.

In this embodiment of the present invention, the recommendation information matching module 450 further includes:

the normalization processing sub-module 451 is used for performing normalization processing on the association degree value and the matching degree value;

the similarity determining submodule 452 is configured to obtain a similarity between the target user and the information to be recommended based on the association degree value and the matching degree value;

and the recommendation information matching submodule 453 is configured to select a preset number of pieces of information to be recommended with the highest similarity to the target user as recommendation information of the target user, and send the recommendation information to the target user.

And a target scenario topic determination module 460, configured to determine a target scenario topic matched with the target user according to the association degree value.

And a recommendation information display module 470, configured to display the information to be recommended according to the target scenario topic.

Optionally, in an embodiment of the present invention, the recommendation information presentation module further may include:

the recommendation information display submodule is used for displaying recommendation information related to the target scene theme in the information to be recommended according to the target scene theme; the recommendation information comprises at least one of recommendation reason, picture information, video information and character information.

The embodiment of the invention also discloses an electronic device, which comprises:

The embodiment of the invention also discloses a readable storage medium, and when instructions in the storage medium are executed by a processor of the electronic equipment, the electronic equipment can execute the information recommendation method.

For the device embodiment, since it is basically similar to the method embodiment, the description is simple, and for the relevant points, refer to the partial description of the method embodiment.

The algorithms and displays presented herein are not inherently related to any particular computer, virtual machine, or other apparatus. Various general purpose systems may also be used with the teachings herein. The required structure for constructing such a system will be apparent from the description above. Moreover, the present invention is not directed to any particular programming language. It is appreciated that a variety of programming languages may be used to implement the teachings of the present invention as described herein, and any descriptions of specific languages are provided above to disclose the best mode of the invention.

In the description provided herein, numerous specific details are set forth. It is understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be interpreted as reflecting an intention that: that the invention as claimed requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.

Those skilled in the art will appreciate that the modules in the device in an embodiment may be adaptively changed and disposed in one or more devices different from the embodiment. The modules or units or components of the embodiments may be combined into one module or unit or component, and furthermore they may be divided into a plurality of sub-modules or sub-units or sub-components. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or elements of any method or apparatus so disclosed, may be combined in any combination, except combinations where at least some of such features and/or processes or elements are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.

Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features included in other embodiments, rather than other features, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments may be used in any combination.

The various component embodiments of the invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It will be appreciated by those skilled in the art that a microprocessor or Digital Signal Processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components of an information recommendation device according to embodiments of the present invention. The present invention may also be embodied as apparatus or device programs (e.g., computer programs and computer program products) for performing a portion or all of the methods described herein. Such programs implementing the present invention may be stored on computer-readable media or may be in the form of one or more signals. Such a signal may be downloaded from an internet website or provided on a carrier signal or in any other form.

It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The usage of the words first, second and third, etcetera do not indicate any ordering. These words may be interpreted as names.

Claims

1. An information recommendation method, comprising:

determining a matching degree value of the information to be recommended and the scene theme according to second user data corresponding to the information to be recommended, wherein the second user data corresponding to each information to be recommended comes from at least one referenceable user;

determining information to be recommended matched with the target user based on the association degree value and the matching degree value;

the step of determining the matching degree value of the information to be recommended and the scene theme according to the second user data corresponding to each piece of information to be recommended includes: and determining the matching degree value of the information to be recommended and each scene theme according to the second user data corresponding to the information to be recommended aiming at each piece of information to be recommended.

2. The method of claim 1, further comprising, prior to the step of determining the association degree value of the target user with each context topic based on the first user data of the target user:

performing data mining based on third user data of the referenceable user, and extracting a specific scene theme;

and/or defining a specific scene theme according to a preset scene judgment condition.

3. The method according to claim 2, wherein the step of performing data mining based on third user data of the referenceable user to extract a specific scene topic comprises:

vectorizing the third user data of the referenceable user to obtain a multi-dimensional word vector corresponding to the third user data;

and obtaining the scene theme through a theme model based on the multi-dimensional word vector.

4. The method according to claim 3, wherein the vectorizing the third user data of the referenceable user to obtain the multi-dimensional word vector corresponding to the third user data comprises:

performing word segmentation processing on the third user data of the referenceable user;

removing invalid words in the third user data after word segmentation processing, and extracting feature words in the third user data, wherein the invalid words comprise at least one of stop words and high-frequency words;

and constructing a multi-dimensional word vector of the third user data based on the feature words.

5. The method according to any of claims 1-3, wherein the contextual topic is characterized by contextual topic words and/or topic related words under the contextual topic word category.

6. The method of claim 5, wherein the step of determining the association degree value of the target user with each scene topic based on the first user data of the target user comprises:

according to the probability of the scenario related words contained in each piece of first user data of the target user under each scenario subject word and/or the subject related words under the scenario subject word category, obtaining the characteristic value of the first user data for each scenario subject word and/or the subject related words under the scenario subject word category;

and acquiring association degree values of the target user to the scene subject words and/or the scene subjects represented by the subject related words in the scene subject word category based on the characteristic values of the first user data in the scene subject words and/or the subject related words in the scene subject word category.

7. The method according to claim 6, wherein the step of obtaining the feature value of the first user data for each scenario topic word and/or topic related word under the scenario topic word category according to the probability of the scenario related word under each scenario topic word and/or topic related word under the scenario topic word category included in each first user data of the target user comprises:

for each piece of first user data, extracting topic related words in the first user data;

and summing the probabilities of the theme related words under the scenario theme words and/or the theme related words under the scenario theme word category aiming at each scenario theme word and/or the theme related words under the scenario theme word category to obtain the characteristic value of the first user data aiming at the scenario theme words and/or the theme related words under the scenario theme word category.

8. The method according to claim 6, wherein the step of obtaining the association degree value of the target user for the scenario topic word and/or the topic correlation word under the scenario topic word category based on the feature value of the first user data under each scenario topic word and/or topic correlation word under the scenario topic word category comprises:

for each scenario topic word and/or topic related words under the scenario topic word category, acquiring characteristic values and values of all first user data of the target user for the scenario topic word and/or topic related words under the scenario topic word category;

and acquiring the ratio of the characteristic value and the value to the quantity of all the first user data to obtain the association degree value of the target user for the scene topic word and/or the scene topic represented by the topic related word under the category of the scene topic word.

9. The method according to claim 6 or 8, wherein the correlation value comprises a weighted sum of short-term correlation values and/or long-term correlation values; wherein, the weight of the short-term correlation degree value is greater than the weight of the long-term correlation degree value; wherein the relevance degree value of the target user u for a certain scenario topic is represented as:

representing the short-term relevancy value of the target user for the contextual theme,

and representing the long-term relevance degree value of the target user for the corresponding scene theme, wherein alpha and beta are weights of the short-term relevance degree value and the long-term relevance degree value respectively.

10. The method according to claim 5, wherein the step of determining the matching degree value between the information to be recommended and the scene subject according to the second user data corresponding to each piece of information to be recommended comprises:

aiming at each piece of information to be recommended, each piece of second user data of each referenceable user aiming at the information to be recommended is obtained;

and for each scene topic word and/or the topic related words under the category of the scene topic word, obtaining a matching degree value of the information to be recommended to the scene topic word and/or the topic related words under the category of the scene topic word according to the second user data, the feature value of the scene topic word and/or the topic related words under the category of the scene topic word, and the quantity of the second user data corresponding to the information to be recommended.

11. The method according to claim 1, wherein the step of determining information to be recommended that matches the target user based on the degree of association value and the degree of matching value, and sending the information to be recommended to the target user comprises:

acquiring the similarity between the target user and the information to be recommended based on the association degree value and the matching degree value;

and selecting a preset amount of information to be recommended with the highest similarity to the target user as the recommendation information of the target user, and sending the recommendation information to the target user.

12. The method according to claim 11, further comprising, before the step of obtaining the similarity between the target user and the information to be recommended based on the association degree value and the matching degree value:

and carrying out normalization processing on the correlation degree value and the matching degree value.

13. The method according to claim 1, wherein after the step of determining information to be recommended that matches the target user based on the degree of association value and the degree of matching value, and sending the information to be recommended to the target user, the method further comprises:

determining a target scene theme matched with the target user according to the association degree value;

and displaying the information to be recommended according to the target scene theme.

14. The method according to claim 13, wherein the step of presenting the information to be recommended according to the target contextual theme comprises:

and displaying recommendation information related to the target scene theme in the information to be recommended according to the target scene theme, wherein the recommendation information comprises at least one of recommendation reason, picture information, video information and text information.

15. An information recommendation apparatus, comprising:

the matching degree determining module is used for determining a matching degree value between the information to be recommended and the scene theme according to second user data corresponding to the information to be recommended, wherein the second user data corresponding to each piece of information to be recommended is from at least one referenceable user;

the recommendation information matching module is used for determining information to be recommended matched with the target user based on the association degree value and the matching degree value and sending the information to be recommended to the target user;

wherein the matching degree determining module is specifically configured to: and determining the matching degree value of the information to be recommended and each scene theme according to the second user data corresponding to the information to be recommended aiming at each piece of information to be recommended.

16. An electronic device, comprising:

processor, memory and computer program stored on the memory and executable on the processor, characterized in that the processor implements the information recommendation method according to any one of claims 1-14 when executing the computer program.

17. A readable storage medium, wherein instructions in the storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the information recommendation method of any one of claims 1-14.