CN109241451B - Content combination recommendation method and device and readable storage medium - Google Patents
Content combination recommendation method and device and readable storage medium Download PDFInfo
- Publication number
- CN109241451B CN109241451B CN201811324196.4A CN201811324196A CN109241451B CN 109241451 B CN109241451 B CN 109241451B CN 201811324196 A CN201811324196 A CN 201811324196A CN 109241451 B CN109241451 B CN 109241451B
- Authority
- CN
- China
- Prior art keywords
- content
- recommended
- combination
- user
- combinations
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The invention provides a content combination recommendation method, a content combination recommendation device and a readable storage medium, and relates to the technical field of Internet, wherein the recommendation method comprises the following steps: acquiring a prepared recommended content set and a preset recommended content number; calculating a plurality of content combinations according to the prepared recommended content set and the number of recommended contents, wherein each content combination comprises the content of the number of recommended contents; determining a recommended content combination from the plurality of content combinations; and pushing the recommended content combination. The plurality of combined contents are obtained by combining according to the collected contents, and the recommended content combination is determined from the plurality of combined contents, so that the displayed contents are more comprehensive, and the user experience is improved.
Description
Technical Field
The invention relates to the technical field of internet, in particular to a content combination recommendation method and device and a readable storage medium.
Background
At present, in a browser recommendation system, a browser main interface can only display part of contents, the number of the contents which can be displayed is only a few, in order to attract more users to consume the contents on the main interface, the most suitable user and the content which can attract the specific user most need to be selected from a lot of contents, and in the prior art, a recommended content combination is obtained by simply sequencing through a recommendation algorithm of a single content, the diversity of characteristics among the contents is lacked, and the like, so that the user cannot be attracted comprehensively.
Disclosure of Invention
In order to overcome the above-mentioned deficiencies in the prior art, the present invention provides a content recommendation method and apparatus for improving the above-mentioned problems.
In order to achieve the above object, the technical solutions provided by the embodiments of the present invention are as follows:
in a first aspect, an embodiment of the present invention provides a content recommendation method and apparatus, including: acquiring a prepared recommended content set and a preset recommended content number; calculating to obtain a plurality of content combinations according to the prepared recommended content set and the number of recommended contents; wherein each content combination comprises the content of the recommended content number; determining a recommended content combination from the plurality of content combinations; and pushing the recommended content combination.
With reference to the first aspect, in some possible implementations, the obtaining of the preliminary recommended content set includes: acquiring contents in a preset database by adopting at least two content acquisition modes to obtain a content set corresponding to each content acquisition mode; and extracting contents in the content set corresponding to each content acquisition mode through a preset rule to obtain the prepared recommended content set.
With reference to the first aspect, in some possible implementation manners, the acquiring, by using at least two content acquisition manners, content in a preset database to obtain a content set corresponding to each content acquisition manner includes: adopting the acquisition rules respectively corresponding to the at least two content acquisition modes to carry out relevancy sorting on the contents in the database to obtain a content set which corresponds to each content acquisition mode and is sorted according to relevancy; correspondingly, the extracting the content in the content set corresponding to each content acquisition mode through a preset rule to obtain the prepared recommended content set includes: and acquiring the preset selection number of contents with the highest correlation in each content set to obtain the prepared recommended content set.
With reference to the first aspect, in some possible implementations, the calculating, according to the preliminary recommended content set and the number of recommended contents, a plurality of content combinations includes: and randomly combining each content in the preliminary recommended content set by taking the number of the recommended contents as the number of each content combination to obtain a plurality of content combinations.
With reference to the first aspect, in some possible implementations, before the randomly combining each content in the preliminary recommended content set to obtain a plurality of content combinations, the method further includes: removing the same content in the set of preliminary recommended content.
With reference to the first aspect, in some possible implementations, the determining a recommended content combination from the plurality of content combinations includes: acquiring a preset content feature set of each content in each content combination; merging the preset content feature set of each content in each content combination to obtain a combined feature set of each content combination; acquiring a judgment feature set of a user, wherein the types of the features included in the combined feature set and the types of the features included in the judgment feature set both comprise: at least one feature or a combination of features of the article-related features, the user habit features and the user scene features; and comparing the judgment feature set with each combined feature set to determine the recommended content combination in the plurality of content combinations.
With reference to the first aspect, in some possible implementations, the acquiring a determination feature set of a user includes: acquiring a historical content combination displayed for a user; according to the historical content combination, determining that the historical content combination selected by the user is a positive case judgment set, and determining that the historical content not selected by the user is a negative case judgment set; the set of decision features includes the set of positive case decisions and the set of negative case decisions.
With reference to the first aspect, in some possible implementations, the comparing the determination feature set with each of the combined feature sets to determine the recommended content combination in the plurality of content combinations includes: collecting the feature set of the content combination in the positive case judgment set as a positive correlation feature set, and collecting the feature set of the content combination in the negative case judgment set as a negative correlation feature set; training a preset initial incidence relation model according to the positive correlation characteristic set and the negative correlation characteristic set to generate an incidence relation model; and determining the recommended content combination in the plurality of content combinations according to the incidence relation model and the combined feature set corresponding to the plurality of content combinations.
In a second aspect, an embodiment of the present invention further provides a content combination recommendation apparatus, where the apparatus includes: the device comprises an acquisition unit, a processing unit and a sending unit. The acquisition unit is used for acquiring a prepared recommended content set and the number of preset recommended contents; the processing unit is used for calculating a plurality of content combinations according to the prepared recommended content set and the number of recommended contents and determining the recommended content combination from the plurality of content combinations; the sending unit is used for pushing the recommended content combination to the user.
In a third aspect, an embodiment of the present invention further provides a readable storage medium, where a computer program is stored in the readable storage medium, and when the computer program runs on a computer, the computer is caused to execute the content combination recommendation method according to the first aspect or any one of the embodiments of the first aspect.
The beneficial effects of the invention include:
by obtaining the preliminary recommended content set, combining the contents in the preliminary content set to obtain a plurality of content combinations, and obtaining the recommended content combination from the plurality of content combinations, a better content combination strategy can be selected compared with the current single content comparison sorting, so that the recommended content combination is more attractive to users.
In order to make the aforementioned objects, features and advantages of the present invention more comprehensible, embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.
Fig. 1 is a schematic flowchart of a content combination recommendation method according to embodiment 1 of the present invention;
fig. 2 is a schematic flowchart of one implementation of obtaining a preliminary recommended content set according to embodiment 1 of the present invention;
fig. 3 is a schematic flowchart of determining a recommended content combination according to embodiment 1 of the present invention;
fig. 4 is a functional block diagram of a content combination recommendation device according to embodiment 2 of the present invention;
fig. 5 is a block diagram of an electronic device according to embodiment 3 of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.
In the description of the embodiments of the present invention, it should be noted that the indication of the orientation or the positional relationship is based on the orientation or the positional relationship shown in the drawings, or the orientation or the positional relationship which is usually placed when the product of the present invention is used, or the orientation or the positional relationship which is usually understood by those skilled in the art, or the orientation or the positional relationship which is usually placed when the product of the present invention is used, and is only for the convenience of describing the present invention and simplifying the description, but does not indicate or imply that the indicated device or element must have a specific orientation, be constructed in a specific orientation, and be operated, and therefore, cannot be understood as limiting the present invention.
Example 1
Referring to fig. 1, fig. 1 is a flowchart illustrating a combined content recommendation method according to an embodiment of the present invention. The content recommendation algorithm comprises:
step S101: acquiring a prepared recommended content set and a preset recommended content number;
step S102: calculating to obtain a plurality of content combinations according to the prepared recommended content set and the number of recommended contents;
step S103: determining a recommended content combination from the plurality of content combinations;
step S104: and pushing the recommended content combination.
It should be noted that, the terminal currently has two forms, namely a computer end and a mobile phone end, and because the two forms are completely different in size, when a user uses different terminals to browse a page or stays in an application main interface, a corresponding appropriate recommended content number can be preset according to the terminal used by the user, so that each content combination contains an appropriate number of contents.
By obtaining the preliminary recommended content set, combining the contents in the preliminary content set to obtain a plurality of content combinations, and obtaining the recommended content combinations from the obtained certain number of combination strategies, compared with the current single content, the method can select a better content combination strategy, and can select so that the recommended content combinations are more attractive to users.
The following describes in detail implementation procedures of steps of the combined content recommendation method according to the embodiment of the present invention.
Referring to fig. 2, fig. 2 is a flowchart illustrating an implementation of obtaining a preliminary recommended content set according to embodiment 1 of the present invention.
Optionally, step S101 includes:
step S201: acquiring contents in a preset database by adopting at least two content acquisition modes to obtain a content set corresponding to each content acquisition mode;
step S202: and extracting contents in the content set corresponding to each content acquisition mode through a preset rule to obtain the prepared recommended content set.
In this embodiment, three recall manners based on collaborative filtering, a topic model, and an association rule of a user are used as content acquisition manners, and the acquisition algorithms of the three content acquisition manners are different, for example: the collaborative filtering based on the user is to take a specific user as a theme and pay attention to the social attributes of the user, namely, the user is more emphatically recommended with contents liked by other users having similar hobbies with the user; the theme model algorithm focuses on the theme of each article content, and carries out classification selection aiming at the favorite theme of a specific user; since all three ways are well known to those skilled in the art, the present embodiment will not be described in detail herein. And respectively adopting three content acquisition modes, and classifying the contents or extracting a certain amount of contents from a preset database of recommended contents according to the characteristics concerned by each acquisition mode to be used as a content set, thereby obtaining three incompletely identical sequenced content sets. In other embodiments, for example, a collaborative filtering algorithm based on content may also be used, which is not limited in this embodiment of the present invention.
In the process of collecting the content in the preset database in step S201, the present invention further proposes to sort the content so as to extract different content sets to obtain a pre-recommended content set in step S202.
Optionally, step S201 includes: adopting the acquisition rules respectively corresponding to the at least two content acquisition modes to carry out relevancy sorting on the contents in the database to obtain a content set which corresponds to each content acquisition mode and is sorted according to relevancy;
correspondingly, after acquiring the content set, step S202 includes: and acquiring the preset selection number of contents with the highest correlation in each content set to obtain the prepared recommended content set.
It should be noted that, in the multiple content acquisition manners, the process of acquiring the content may be understood as a process of calculating and comparing the relevance once, and when the content with higher relevance is calculated by using a manner such as a keyword, the content may be extracted according to the result of the ranking of the relevance by using the ranking of the computed relevance, that is, in this embodiment, two articles with the highest relevance acquired in each acquisition manner are selected, and in other embodiments, one or more articles with the highest relevance may also be selected, which is not limited in this embodiment.
In this embodiment, after content sets with different attention characteristics are obtained by using multiple content acquisition manners, a pre-recommended content set is formed by combining the content with high correlation degree acquired by each content acquisition manner. In other embodiments, content classification may be performed first, and then, according to the category selected by the user, the content in the multiple preference categories of the user is used as a content set, so as to obtain different content sets.
After the preliminary recommended content set is acquired, step S102 is executed to combine the contents in the preliminary recommended content set into a plurality of content combinations.
Optionally, step S102 includes: and randomly combining each content in the preliminary recommended content set by taking the number of the recommended contents as the number of each content combination to obtain a plurality of content combinations.
The random combination is exemplified: assuming that three content sets A, B and C are acquired by three content acquisition methods, 3 pieces of content are extracted from each content set to obtain the set of preliminary recommended content:
[A1，A2，A3，B1，B2，B3，C1，C2，C3]
making the nine contents into C9 3The most possible content combination is obtained by the random combination, so that a recommended content combination which is more suitable and preferred by the user can be determined for the user.
In each content collection method, since the same content database is preset, although the feature attention of each content collection method is different, each content set may select the same content, for example: content A1And content B1Are identical or are the same article.
Thus, in randomly combining each piece of content in the set of preliminary recommended content, a plurality of content combinations are obtained: and removing the same content in the preliminary recommended content set, and randomly combining each content in the preliminary recommended content set to obtain a plurality of content combinations.
And performing re-ranking processing on the obtained preliminary recommended content set, removing the contents with the same content or the same article and the like, and preventing two same contents from appearing in the recommended content combination.
Referring to fig. 3, fig. 3 is a schematic flowchart of determining a recommended content combination according to embodiment 1 of the present invention. Further, after obtaining a plurality of content combinations, a recommended content combination needs to be determined according to a certain rule.
Optionally, step S103 includes:
step S301: acquiring a preset content feature set of each content in each content combination;
step S302: merging the preset content feature set of each content in each content combination to obtain a combined feature set of each content combination;
step S303: acquiring a judgment feature set of a user, wherein the types of the features contained in the preset content feature set are the same as the types of the features contained in the judgment feature set;
step S304: and comparing the judgment feature set with each combined feature set, and determining and calculating to obtain the recommended content combination.
In the existing machine learning, the algorithm model for ranking and scoring includes: an LR model (Logistic Regression), a GBDT model (Gradient Boosting Decision Tree model), and the like, in this embodiment, the GBDT model is used to score each content combination, and in the GBDT model, the determination features of the user in the determination feature set of the user are selected as each scoring node, and then the combined feature set of each content combination is imported into the model to be scored and sorted, so as to obtain the recommended content combination.
It should be noted that the categories of the features include content aspects, user aspects, scenario aspects, and device aspects, such as: the content aspects include article types: education, medical, etc., user aspects including: gender, hobbies, occupation, etc., situational aspects including: current time, working day, holiday, etc.; the equipment aspect comprises: android or apple, a computer end or a mobile phone end and the like, and various features can be used in combination, and the embodiment does not limit the application.
Further, in this embodiment, the step S303 of determining the feature set acquisition manner by the user includes:
acquiring a historical content combination displayed for a user; according to the historical content combination, determining that the historical content combination selected by the user is a positive case judgment set, and determining that the historical content not selected by the user is a negative case judgment set; the set of decision features includes the set of positive case decisions and the set of negative case decisions.
By acquiring historical content combinations displayed for a user in the past, taking the historical content combinations as a training set of a GBDT model and adopting a positive and negative case common judgment mode, a plurality of content combinations are screened from two aspects, the content which makes the user feel dislike is also avoided being selected, and the user experience is improved. In other embodiments, the positive or negative example may also be used alone for determination, which is not limited in this embodiment.
It should be noted that, if only the historical information of a specific user is collected to obtain a determination feature set, the content of the training set is often too small, in other embodiments, a feature set of a user concerned or liked by the user may also be collected, and the feature set is incorporated into the training set of the GBDT model to expand the training set.
After the specific decision feature set is determined, the combined feature set of each content combination needs to be substituted into the GBDT model for scoring and sorting to obtain recommended content combinations.
Optionally, step S304 includes:
collecting the feature set of the content combination in the positive case judgment set as a positive correlation feature set, and collecting the feature set of the content combination in the negative case judgment set as a negative correlation feature set; training a preset initial incidence relation model according to the positive correlation characteristic set and the negative correlation characteristic set to generate an incidence relation model; and determining the recommended content combination in the plurality of content combinations according to the incidence relation model and the combined feature set corresponding to the plurality of content combinations.
And training an initial GBDT model by utilizing a positive correlation characteristic set and a negative correlation characteristic set, wherein the positive correlation characteristic set and the negative correlation characteristic set are taken as output quantities of the GBDT model, the characteristics in the positive correlation characteristic set and the negative correlation characteristic set are taken as input quantities of the GBDT model, and the selection of the GBDT model for a combined characteristic set in a plurality of content combinations is optimized, so that the optimal content combination recommendation is obtained.
Optionally, the first implementation manner of step S304 is: and collecting the features of the content combinations in the positive case judgment set as positive correlation features, judging whether each feature in the combination feature set of each content combination is the same as each positive correlation feature, and if so, obtaining a first score by the content combinations. And the positive and direct scores collect the characteristics of the content combinations in the negative example judgment set as negative correlation characteristics, judge whether each characteristic in the combination characteristic set of each content combination is the same as each negative correlation characteristic, and if so, the content combinations obtain second scores. Calculating a final score of a corresponding content combination according to the first score and the second score of each content combination; and determining the content combination corresponding to the final score as a recommended content combination.
Optionally, a second implementation manner of step S304 is: collecting feature sets of the content combinations in the normal case judgment set as positive correlation feature sets, comparing the combined feature set of each content combination with the positive correlation feature sets, and acquiring first similarity of the combined feature sets and the positive correlation feature sets; obtaining a first user click probability of the content combination according to a first preset user click probability of the positive correlation characteristic set and the first similarity;
collecting the feature sets of the content combinations in the negative case judgment set as negative correlation feature sets, comparing the combined feature set of each content combination with the negative correlation feature set, and acquiring first similarity of the combined feature set and the negative correlation feature set; according to the second preset user click probability of the negative correlation characteristic set, it can be understood that the second preset user click probability is lower than the first preset user click probability, and the second user click probability of the content combination is obtained according to the second similarity and the second preset user click probability of the negative correlation characteristic set;
and obtaining the user click probability of the content combination by combining the first user click probability and the second user click probability, for example, by adding the first user click probability and the second user click probability, and further comparing the user click probabilities of the content combinations to obtain the recommended content combination.
In other embodiments, in step S304, a recommended content combination may be obtained in other manners according to the first similarity and the second similarity, which are obtained by comparing a feature combination policy of the content combination with a sample, and this embodiment is not limited thereto.
In this embodiment, whether the positive and negative correlation features are the same as each feature in the combined feature set is compared, and then according to the comparison result, each node scores the content combination, and finally the scoring results of each node are algebraically added to obtain the final score. It should be noted that, in other embodiments, similar features may also be classified into one type for determination in the same semantic determination manner, which is not limited in this embodiment.
Further, in this embodiment, the same positive correlation characteristic is used to obtain one score, and the same negative correlation characteristic is used to obtain one minus score; in other embodiments, the relevant score may also be given according to the degree of correlation between the two features, which is not limited by this embodiment.
The working principle of embodiment 1 of the present invention is explained as one possible case: assuming that three content recall strategies of collaborative filtering, a theme model and an association rule are adopted as the content acquisition mode in the invention, a content set A, a content set B and a content set C are obtained, the similarity weight of the contents among users is obtained according to a scoring ordering algorithm based on the collaborative filtering of the users, such as Euclidean measurement, so as to obtain the recommended ordering in the content set A, and the recommended ordering in the content set B and the content set C is obtained according to the scoring ordering algorithm in the theme model and the association rule in the same way.
Further, assume that the first three pieces of content in the content collection in the present embodiment are collected into the preliminary recommended content set, and thus, the preliminary recommended content set includes 9 pieces of content. Then, for the contents in the preliminary recommended content setPerforming content duplication elimination, removing the same content in 9 pieces of content to obtain 6 pieces of content, and assuming that the recommended content combination comprises 3 pieces of content; therefore, after the pre-recommended content set is obtained, the pre-recommended contents in the pre-recommended content set are randomly combined, namely according to C6 3Resulting in 20 content combinations.
Then, a preset content feature set of each piece of preliminary recommended content is obtained, so that the preset content feature sets of the three pieces of content of each combined content are combined to form a combined feature set of the combined content, in this embodiment, preference features, usage habit features, and scene features used by the user at this time are obtained as determination feature sets, and the preference features of the specific user include: favorite article categories, time for browsing articles, etc., user usage habit features include: the network type often used on the internet, the average stay time of a page in a week, and the like, and the scene characteristics used by the user at the moment comprise: the current time and date on the user terminal, etc. And further, judging whether each judgment feature in the judgment feature set is a positive correlation feature or a negative correlation feature from a preset database of the user according to a historical content combination displayed for the user before.
In this embodiment, the GBDT model is used to rank each combination: taking each feature in the determination feature set as each node in the GBDT model to obtain a decision tree of the GBDT model, and scoring each combination content by comparing the combination feature set of each combination content with the feature in each node, for example: when a node is characterized by: the combined feature set of one combined content comprises the feature of the food, the combined content is scored for 1, the combined feature set of the other combined content does not comprise the feature of the food, scoring is not performed, after the combined feature set traverses all the nodes, the scoring condition of each combined content at each node is integrated to obtain the final score of each combined content, the combined content corresponding to the highest score is obtained through sequencing, the combined content is used as recommended combined content and is further pushed to a user terminal interface, and the user is attracted to select the recommended combined content.
Selecting two pieces of content which are optimal under the attention feature from the three content sets through the content sets which are obtained according to the three acquisition modes and aim at the three attention features, randomly combining the two pieces of content to obtain a plurality of content combinations, and sequencing and scoring the combined features of the plurality of combinations according to user historical data to obtain an optimal recommended combination; by scoring and sequencing the combined features, compared with the existing single attention attribute sequencing, attention is paid to the matching of features among the contents, the comprehensiveness of content recommendation can be ensured, and the content isolation under the existing recommendation algorithm is reduced.
Example 2
Referring to fig. 4, fig. 4 is a functional block diagram of a content combination recommendation device 40 according to embodiment 2 of the present invention.
The content combination recommendation apparatus 40 includes: acquisition section 401, processing section 402, and transmission section 403. The obtaining unit 401 is configured to obtain a set of pre-recommended content and a preset number of recommended content; the processing unit 402 is configured to calculate a plurality of content combinations according to the preliminary recommended content set and the number of recommended contents, and determine a recommended content combination from the plurality of content combinations; the sending unit 403 is configured to push the recommended content combination to the user.
Optionally, the obtaining unit 401 is further configured to collect content in a preset database by using at least two content collection manners, so as to obtain a content set corresponding to each content collection manner; and extracting contents in the content set corresponding to each content acquisition mode through a preset rule to obtain the prepared recommended content set.
Optionally, the processing unit 402 is further configured to perform relevancy sorting on the contents in the database by using the collection rules respectively corresponding to the at least two content collection manners, so as to obtain a content set corresponding to each content collection manner and sorted according to the relevancy;
correspondingly, the obtaining unit 401 is further configured to obtain a preset number of selections, and the processing unit 402 further collects the content with the preset number of selections with the highest correlation in each content set to obtain the preliminary recommended content set.
Optionally, the processing unit 402 is further configured to randomly combine each content in the preliminary recommended content set to obtain a plurality of content combinations, with the number of recommended content as the number of each content combination.
Optionally, the processing unit 402 is further configured to remove the same content in the set of preliminary recommended content, and randomly combine each content in the set of preliminary recommended content to obtain a plurality of content combinations.
Optionally, the obtaining unit 401 is further configured to obtain a preset content feature set of each content in each content combination, and obtain a determination feature set of a user, where it needs to be noted that the preset content feature set and the determination feature set include the same feature type.
Correspondingly, the processing unit 402 is further configured to merge the preset content feature sets of each content in each content combination to obtain a combined feature set of each content combination, further compare the determination feature set with each combined feature set, and determine to calculate the recommended content combination.
Optionally, the obtaining unit 401 is further configured to obtain a history content combination displayed for the user; correspondingly, the processing unit 402 determines, according to the historical content combination, that the historical content combination selected by the user is a positive example judgment set, and determines that the historical content not selected by the user is a negative example judgment set, where it should be noted that the positive example judgment set and the negative example judgment set constitute a judgment feature set.
Further, the processing unit 402 collects the features of the content combinations in the positive example judgment set as positive correlation features, judges whether each feature in the combination feature set of each content combination is the same as each positive correlation feature, and if so, the content combinations obtain a first score;
meanwhile, whether each feature in the combined feature set of each content combination is the same as each negative correlation feature or not is judged, if yes, the content combination obtains a second score, and the first score is higher than the second score; calculating a score finally obtained by each content combination according to the first score and the second score of each content combination;
finally, it is determined that the content combination corresponding to the highest score is the recommended content combination, and then the sending unit 403 pushes the recommended content combination to the user.
The content combination recommendation apparatus 40 in this embodiment and the content combination recommendation method shown in fig. 1 are based on the same concept, and through the foregoing detailed description of the content combination recommendation method and various variations thereof, those skilled in the art can clearly understand the implementation process of the content combination recommendation apparatus 40 in this embodiment, so for brevity of the description, details are not repeated herein.
Example 3
Referring to fig. 5, fig. 5 is a block diagram of an electronic device 50 according to embodiment 3 of the present invention. The electronic device 50 includes: a memory 51 and a processor 52.
The memory 51 and the processor 52 are electrically connected to each other directly or indirectly to enable data transmission or interaction. For example, the components may be electrically connected to each other via one or more communication buses or signal lines. The preset database includes at least one software function module that may be stored in the memory 51 in the form of software or firmware (firmware) or solidified in an Operating System (OS) of the electronic device 50. The processor 52 is used to execute executable modules stored in the memory 51.
The Memory 51 may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Read-Only Memory (EPROM), an electrically Erasable Read-Only Memory (EEPROM), and the like. The memory 51 is used for storing a program, and the processor 52 executes the program after receiving an execution instruction, and a method executed by the electronic device 50 defined by a flow disclosed in any embodiment of the invention described later may be applied to the processor 52, or implemented by the processor 52.
Example 4
Embodiment 4 of the present invention provides a readable storage medium, in which a computer program is stored, and when the computer program runs on a computer, the computer is caused to execute the content combination recommendation method described in embodiment 1. The present invention will not be described in detail herein.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, the functional modules in the embodiments of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
The functions, if implemented in the form of software functional modules 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 instructions for causing 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: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes. It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes will occur to those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (8)
1. A content combination recommendation method is characterized by comprising the following steps:
acquiring a prepared recommended content set and a preset recommended content number; wherein the number of recommended contents is set according to a terminal used by a user;
calculating to obtain a plurality of content combinations according to the prepared recommended content set and the number of recommended contents; wherein each content combination comprises the content of the recommended content number; wherein the calculating according to the preliminary recommended content set and the number of recommended contents to obtain a plurality of content combinations comprises: randomly combining each content in the preliminary recommended content set to obtain a plurality of content combinations by taking the number of the recommended content as the number of each content combination;
acquiring a preset content feature set of each content in each content combination;
merging the preset content feature set of each content in each content combination to obtain a combined feature set of each content combination;
acquiring a judgment feature set of a user, wherein the types of the features included in the combined feature set and the types of the features included in the judgment feature set both comprise: at least one feature or a combination of features of the article-related features, the user habit features and the user scene features;
comparing the decision feature set with each of the combined feature sets to determine the recommended content combination of the plurality of content combinations;
and pushing the recommended content combination.
2. The recommendation method according to claim 1, wherein said obtaining a preliminary recommended content set comprises:
acquiring contents in a preset database by adopting at least two content acquisition modes to obtain a content set corresponding to each content acquisition mode;
and extracting contents in the content set corresponding to each content acquisition mode through a preset rule to obtain the prepared recommended content set.
3. The recommendation method according to claim 2, wherein the acquiring the content in the preset database by using at least two content acquisition manners to obtain the content set corresponding to each content acquisition manner comprises:
adopting the acquisition rules respectively corresponding to the at least two content acquisition modes to carry out relevancy sorting on the contents in the database to obtain a content set which corresponds to each content acquisition mode and is sorted according to relevancy;
correspondingly, the extracting the content in the content set corresponding to each content acquisition mode through a preset rule to obtain the prepared recommended content set includes:
and acquiring the preset selection number of contents with the highest correlation in each content set to obtain the prepared recommended content set.
4. The recommendation method according to claim 1, wherein before said randomly combining each content in the set of preliminary recommended content into a plurality of content combinations, the method further comprises:
removing the same content in the set of preliminary recommended content.
5. The recommendation method according to claim 1, wherein the obtaining of the decision feature set of the user comprises:
acquiring a historical content combination displayed for a user;
according to the historical content combination, determining that the historical content combination selected by the user is a positive case judgment set, and determining that the historical content not selected by the user is a negative case judgment set; the set of decision features includes the set of positive case decisions and the set of negative case decisions.
6. The recommendation method of claim 5, wherein said comparing said decision feature set to each of said combined feature sets to determine said recommended content combination of said plurality of content combinations comprises:
collecting the feature set of the content combination in the positive case judgment set as a positive correlation feature set, and collecting the feature set of the content combination in the negative case judgment set as a negative correlation feature set;
training a preset initial incidence relation model according to the positive correlation characteristic set and the negative correlation characteristic set to generate an incidence relation model;
and determining the recommended content combination in the plurality of content combinations according to the incidence relation model and the combined feature set corresponding to the plurality of content combinations.
7. A content combination recommendation apparatus, characterized in that the apparatus comprises:
the device comprises an acquisition unit, a display unit and a control unit, wherein the acquisition unit is used for acquiring a prepared recommended content set and the number of preset recommended contents; wherein the number of recommended contents is set according to a terminal used by a user;
the processing unit is used for calculating a plurality of content combinations according to the prepared recommended content set and the number of recommended contents and determining the recommended content combination in the content combinations; the processing unit is further configured to randomly combine each content in the preliminary recommended content set to obtain a plurality of content combinations, with the number of the recommended content as the number of each content combination;
the acquiring unit is further configured to acquire a preset content feature set of each piece of content in each content combination;
the processing unit is further configured to merge the preset content feature set of each content in each content combination to obtain a combined feature set of each content combination;
the acquiring unit is further configured to acquire a determination feature set of a user, where the types of the features included in the combined feature set and the types of the features included in the determination feature set both include: at least one feature or a combination of features of the article-related features, the user habit features and the user scene features;
the processing unit is further configured to compare the determination feature set with each of the combined feature sets, and determine the recommended content combination in the plurality of content combinations;
and the sending unit is used for pushing the recommended content combination to the user.
8. A readable storage medium, in which a computer program is stored, which, when run on a computer, causes the computer to execute a content combination recommendation method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811324196.4A CN109241451B (en) | 2018-11-08 | 2018-11-08 | Content combination recommendation method and device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811324196.4A CN109241451B (en) | 2018-11-08 | 2018-11-08 | Content combination recommendation method and device and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109241451A CN109241451A (en) | 2019-01-18 |
| CN109241451B true CN109241451B (en) | 2021-07-16 |
Family
ID=65077533
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811324196.4A Active CN109241451B (en) | 2018-11-08 | 2018-11-08 | Content combination recommendation method and device and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109241451B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111597429A (en) | 2019-02-21 | 2020-08-28 | 北京京东尚科信息技术有限公司 | Network resource pushing method and device and storage medium |
| CN110543598B (en) * | 2019-09-06 | 2022-03-25 | 腾讯科技（深圳）有限公司 | Information recommendation method and device and terminal |
| CN111626767B (en) * | 2020-04-29 | 2023-09-08 | 拉扎斯网络科技（上海）有限公司 | Resource data issuing method, device and equipment |
| CN113254843B (en) * | 2021-06-29 | 2021-10-01 | 腾讯科技（深圳）有限公司 | Information pushing method and device and storage medium |
| CN113806634B (en) * | 2021-09-17 | 2023-05-30 | 中国联合网络通信集团有限公司 | Service package recommending method, device and server |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740468A (en) * | 2016-03-07 | 2016-07-06 | 达而观信息科技（上海）有限公司 | Individuation recommendation method and system combined with content publisher information |
| CN107743249A (en) * | 2017-11-27 | 2018-02-27 | 四川长虹电器股份有限公司 | A kind of CTR predictor methods based on Model Fusion |
| CN107944063A (en) * | 2018-01-16 | 2018-04-20 | 马上消费金融股份有限公司 | It is a kind of that method and system are recommended based on the news of topic model and groups of users |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108304556B (en) * | 2018-02-06 | 2019-06-07 | 中国传媒大学 | The personalized recommendation method combined based on content with collaborative filtering |
-
2018
- 2018-11-08 CN CN201811324196.4A patent/CN109241451B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740468A (en) * | 2016-03-07 | 2016-07-06 | 达而观信息科技（上海）有限公司 | Individuation recommendation method and system combined with content publisher information |
| CN107743249A (en) * | 2017-11-27 | 2018-02-27 | 四川长虹电器股份有限公司 | A kind of CTR predictor methods based on Model Fusion |
| CN107944063A (en) * | 2018-01-16 | 2018-04-20 | 马上消费金融股份有限公司 | It is a kind of that method and system are recommended based on the news of topic model and groups of users |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109241451A (en) | 2019-01-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109241451B (en) | Content combination recommendation method and device and readable storage medium | |
| CN109685631B (en) | Personalized recommendation method based on big data user behavior analysis | |
| WO2020048084A1 (en) | Resource recommendation method and apparatus, computer device, and computer-readable storage medium | |
| CN107369075B (en) | Commodity display method and device and electronic equipment | |
| US10505884B2 (en) | Entity classification and/or relationship identification | |
| CN107341268B (en) | Hot searching ranking method and system | |
| WO2018041168A1 (en) | Information pushing method, storage medium and server | |
| CN110543598B (en) | Information recommendation method and device and terminal | |
| US8825672B1 (en) | System and method for determining originality of data content | |
| US20140074828A1 (en) | Systems and methods for cataloging consumer preferences in creative content | |
| CN110163703B (en) | Classification model establishing method, file pushing method and server | |
| TW201911080A (en) | Search method, search server and search system | |
| CN110197404B (en) | Personalized long-tail commodity recommendation method and system capable of reducing popularity deviation | |
| CN111259173A (en) | Search information recommendation method and device | |
| CN113254777B (en) | Information recommendation method and device, electronic equipment and storage medium | |
| CN111310011A (en) | Information pushing method and device, electronic equipment and storage medium | |
| CN108763369B (en) | Video searching method and device | |
| TWI645348B (en) | System and method for automatically summarizing images and comments within commodity-related web articles | |
| CN112199582A (en) | Content recommendation method, device, equipment and medium | |
| CN115129994A (en) | Commodity recommendation method and device, electronic equipment and readable storage medium | |
| US9020863B2 (en) | Information processing device, information processing method, and program | |
| CN111177564B (en) | Product recommendation method and device | |
| CN110598126B (en) | Cross-social network user identity recognition method based on behavior habits | |
| CN112288510A (en) | Article recommendation method, device, equipment and storage medium | |
| CN113535939A (en) | Text processing method and device, electronic equipment and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |