CN107368533B

CN107368533B - Content item recommendation method and device and electronic equipment

Info

Publication number: CN107368533B
Application number: CN201710450557.9A
Authority: CN
Inventors: 李雪; 宋华; 查强; 张鼎
Original assignee: Beijing QIYI Century Science and Technology Co Ltd
Current assignee: Beijing QIYI Century Science and Technology Co Ltd
Priority date: 2017-06-15
Filing date: 2017-06-15
Publication date: 2020-04-03
Anticipated expiration: 2037-06-15
Also published as: CN107368533A

Abstract

The embodiment of the invention provides a content item recommendation method, a content item recommendation device and electronic equipment, wherein the method comprises the following steps: determining a plurality of target items to be recommended from the items to be recommended corresponding to the target content items, wherein the similarity corresponding to the plurality of target items to be recommended is greater than the similarity corresponding to other items to be recommended; calculating the income capacity corresponding to each target item to be recommended; calculating an effective score corresponding to each target item to be recommended, wherein the effective score corresponding to any target item to be recommended is obtained by calculation according to the similarity and the profit capacity corresponding to the target item to be recommended; and determining recommended items corresponding to the target content items from the target items to be recommended according to the calculated effective scores, and generating a recommendation list comprising the recommended items, wherein the effective scores corresponding to the recommended items are all larger than those of other target items to be recommended. By applying the technical scheme provided by the embodiment of the invention, the effectiveness of the recommendation list can be improved.

Description

Content item recommendation method and device and electronic equipment

Technical Field

The present invention relates to the field of network information technologies, and in particular, to a content item recommendation method and apparatus, and an electronic device.

Background

With the development of network information technology, the recommendation system is widely applied, and users can quickly find out the content which interests themselves through the content recommended by the recommendation system, so that a great amount of time can be saved.

In the prior art, a recommendation system generates a recommendation list for a target content item according to the target content item, and displays the recommendation item to a user through the recommendation list after the user clicks the target content item. The recommendation list is all the content items that may be of interest to the user clicking on the target content item, where the content items may refer to video, audio, goods, applications, and so on that are available for consumption by the user. In the process of generating the recommendation list of the target content item, firstly, content information (such as a label, a title, and the like) and preference information of the user are obtained for the target content item and an item to be recommended of the target content item (for example, the preference information of the user on the content item is marked by utilizing explicit behaviors such as user purchase and approval and implicit behaviors such as browsing and watching); then, according to the obtained content information and preference information, the similarity between the target content item and each item to be recommended is calculated, a plurality of target items to be recommended are determined according to the similarity, and the plurality of target items to be recommended form a recommendation list of the target content item. For example, for all videos stored by the video portal website and used for the user to consume, the target content item may be any one of all stored videos, and the item to be recommended of the target content item may be a video other than the target content item in all stored videos.

However, the inventor finds that the prior art has at least the following problems in the process of implementing the invention: the recommendation method is based on the recommendation of the content items, only the similarity between the recommended items in the recommendation list and the target content items is considered, and the considered factors are single, so that the recommended content items cannot attract the user to continue clicking for consumption. That is, there is a problem in the related art that the recommendation list is low in effectiveness.

Disclosure of Invention

The embodiment of the invention aims to provide an item recommendation method, an item recommendation device and electronic equipment so as to improve the effectiveness of a recommendation list. The specific technical scheme is as follows:

in one aspect of the present invention, an embodiment of the present invention provides a content item recommendation method, including:

determining a plurality of target items to be recommended from items to be recommended corresponding to target content items, wherein the similarity corresponding to the target items to be recommended is greater than the similarity corresponding to other items to be recommended, and the similarity corresponding to any item is the similarity between the item and the target content items;

calculating the profit capacity corresponding to each target item to be recommended, wherein the profit capacity corresponding to any target item to be recommended represents the possibility that a user clicks a recommendation list of the target item to be recommended when the target item to be recommended is used as the recommendation item of the target content item;

calculating an effective score corresponding to each target item to be recommended, wherein the effective score corresponding to any target item to be recommended is obtained by calculation according to the similarity and the profit capacity corresponding to the target item to be recommended;

and determining recommended items corresponding to the target content items from the target items to be recommended according to the calculated effective scores, and generating a recommendation list comprising the recommended items, wherein the effective scores corresponding to the recommended items are all larger than the effective scores of other target items to be recommended.

Optionally, the step of calculating the profit-and-profit capacity corresponding to each target item to be recommended includes:

collecting a recommendation log of each target item to be recommended;

according to the recommendation log, under the condition that each target item to be recommended is used as the recommendation item of the target content item, the display times of a recommendation list for displaying the target item to be recommended after the target item to be recommended is clicked and the effective click times of the recommendation list for displaying the target item to be recommended are calculated, wherein the effective click times refer to the times of clicks meeting preset requirements;

calculating the ratio of the effective click times corresponding to each target item to be recommended to the corresponding display times of each target item to be recommended;

and determining the ratio corresponding to each target item to be recommended as the profit capacity of the corresponding target item to be recommended.

collecting a recommendation log of each target item to be recommended;

according to the recommendation log, aiming at each target item to be recommended, calculating the display times of an effective recommendation list of the target content item after the target content item is clicked and the effective click times of the target item to be recommended in all the effective recommendation lists, wherein the effective recommendation list is a recommendation list when the target item to be recommended is used as the recommendation item of the target content item, and the effective click times are the times of clicking meeting the preset requirements;

Optionally, the method further includes:

and for each target item to be recommended, determining a preset value as the ratio corresponding to the target item to be recommended under the condition that the calculation of the ratio corresponding to the target item to be recommended fails.

Optionally, the step of calculating the effective score corresponding to each target item to be recommended includes:

and for each target item to be recommended, carrying out weighted summation on the similarity and the profit capacity corresponding to the target item to be recommended to obtain an effective score corresponding to the target item to be recommended.

Optionally, the step of generating a recommendation list including the recommendation item includes:

and sorting the recommended items according to the effective scores of the recommended items, and forming a recommendation list of the target content item by the sorted recommended items.

In another aspect of the present invention, an embodiment of the present invention further provides a content item recommendation apparatus, where the apparatus includes:

the device comprises a first determining module, a second determining module and a third determining module, wherein the first determining module is used for determining a plurality of target items to be recommended from items to be recommended corresponding to target content items, the similarity corresponding to the target items to be recommended is larger than the similarity corresponding to other items to be recommended, and the similarity corresponding to any item is the similarity between the item and the target content items;

the first calculation module is used for calculating the profit capacity corresponding to each target item to be recommended, wherein the profit capacity corresponding to any target item to be recommended represents the possibility that a user clicks the recommendation list of the target item to be recommended when the target item to be recommended is used as the recommendation item of the target content item;

the second calculation module is used for calculating the effective score corresponding to each target item to be recommended, wherein the effective score corresponding to any target item to be recommended is obtained by calculation according to the similarity and the profit capacity corresponding to the target item to be recommended;

and the processing module is used for determining recommended items corresponding to the target content items from the target items to be recommended according to the calculated effective scores and generating a recommendation list comprising the recommended items, wherein the effective scores corresponding to the recommended items are all larger than the effective scores of other target items to be recommended.

Optionally, the first computing module includes:

the first collection unit is used for collecting a recommendation log of each target item to be recommended;

the first processing unit is used for calculating the display times of a recommendation list for displaying the target item to be recommended after the target item to be recommended is clicked and the effective click times of the recommendation list for displaying the target item to be recommended according to the recommendation log under the condition that the target item to be recommended is used as the recommendation item of the target content item, wherein the effective click times refer to the click times meeting the preset requirement;

the first calculating unit is used for calculating the ratio of the effective click times corresponding to each target item to be recommended to the corresponding display times of the target item to be recommended;

and the first determining unit is used for determining the ratio corresponding to each target item to be recommended as the profit capacity of the corresponding target item to be recommended.

Optionally, the first computing module includes:

the second collection unit is used for collecting a recommendation log of each target item to be recommended;

the second processing unit is used for calculating the display times of an effective recommendation list of the target content item after the target content item is clicked and the effective click times of the target to-be-recommended item in all the effective recommendation lists according to the recommendation log and aiming at each target to-be-recommended item, wherein the effective recommendation list is a recommendation list when the target to-be-recommended item is used as a recommendation item of the target content item, and the effective click times are click times meeting preset requirements;

the second calculation unit is used for calculating the ratio of the effective click times corresponding to each target item to be recommended to the corresponding display times of the target item to be recommended;

and the second determining unit is used for determining the ratio corresponding to each target item to be recommended as the profit capacity of the corresponding target item to be recommended.

Optionally, the apparatus further comprises:

and the second determining module is used for determining a preset value as the ratio corresponding to each target item to be recommended under the condition that the calculation of the ratio corresponding to the target item to be recommended fails.

Optionally, the second computing module includes:

and the score calculating unit is used for weighting and summing the similarity and the profit capacity corresponding to each target item to be recommended to obtain the effective score corresponding to the target item to be recommended.

Optionally, the processing module is configured to, when generating the recommendation list including the recommendation items, specifically, sort the recommendation items according to the validity score of each recommendation item, and combine the sorted recommendation items into the recommendation list of the target content item.

In another aspect of the present invention, an embodiment of the present invention further provides an electronic device, where the electronic device includes a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus;

a memory for storing a computer program;

a processor for implementing any of the above-described content item recommendation methods when executing a program stored on the memory.

In yet another aspect of the present invention, there is also provided a computer-readable storage medium having stored therein instructions, which when run on a computer, cause the computer to perform any of the above-described content item recommendation methods.

In yet another aspect of the present invention, there is also provided a computer program product comprising instructions which, when run on a computer, cause the computer to perform any of the content item recommendation methods described above.

According to the content item recommendation method, device and electronic equipment provided by the embodiment of the invention, firstly, a plurality of target items to be recommended are determined from the items to be recommended corresponding to the target content items, wherein the similarity corresponding to the plurality of target items to be recommended is greater than the similarity corresponding to other items to be recommended, and the similarity corresponding to any item is the similarity between the item and the target content items, so that each target item to be recommended and the target content items have certain similarity, and users are likely to be more interested in the target items to be recommended relative to the content items with lower similarity. And calculating the profit capacity corresponding to each target item to be recommended, wherein the profit capacity corresponding to any target item to be recommended represents the possibility that the user clicks the recommendation list of the target item to be recommended when the target item to be recommended is taken as the recommendation item of the target content item, and the higher the profit capacity can represent the higher the possibility that the user continuously consumes through the recommendation list. Then, calculating an effective score corresponding to each target item to be recommended, wherein the effective score corresponding to any target item to be recommended is obtained by calculation according to the similarity and the profit capacity corresponding to the target item to be recommended; and finally, according to the calculated effective scores, determining recommended items corresponding to the target content items from the target items to be recommended, and generating a recommendation list comprising the recommended items, wherein the effective scores corresponding to the recommended items are all larger than those of other target items to be recommended. The higher the validity score is, the higher the possibility that the representative user has a high degree of interest in the recommended item is, and the higher the possibility that the representative user is continuously consumed through the recommendation list is, that is, the higher the degree of attraction of the representative item to the user is, and therefore, the validity of the recommendation list can be improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below.

FIG. 1 is a flow chart illustrating a method for recommending content items according to an embodiment of the present invention;

FIG. 2 is a schematic diagram illustrating a first process of calculating the profit-making capability corresponding to each target item to be recommended;

FIG. 3 is a schematic diagram of a second flowchart of calculating the profit-making capability corresponding to each target item to be recommended;

FIG. 4 is a schematic structural diagram of a content item recommendation apparatus according to an embodiment of the present invention;

FIG. 5 is a diagram illustrating a first structure of a first computing module;

FIG. 6 is a diagram illustrating a second structure of the first computing module;

fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be described below with reference to the drawings in the embodiments of the present invention.

In the content item recommendation method in the prior art, only the similarity between the recommended item and the target content item in the recommendation list is considered, and the considered factors are single, so that the recommended content item cannot attract the user to continue clicking for consumption. That is to say, the prior art has a problem of low validity of a recommendation list, and in order to improve the validity of the recommendation list, embodiments of the present invention provide a content item recommendation method, apparatus, and electronic device.

A method for recommending a content item according to an embodiment of the present invention will be described first.

It should be noted that the target content item described in the embodiment of the present invention refers to a content item for which a recommendation list is to be generated.

In a specific application, if the target content item is a content item that is currently consumed by the user, a recommendation list of the target content item may be generated and recommended to the user by using the method, the apparatus, or the electronic device provided in the embodiments of the present invention when the user consumes the target content item, so as to display the recommended item in the recommendation list.

If the target content item is not the content item currently being consumed, a recommendation list of the target content item may be generated and stored by using the method, apparatus, or electronic device provided in the embodiment of the present invention, and specifically, if the recommendation list of the target content item already exists, the original recommendation list of the target content item may be updated by using the newly generated recommendation list; if no recommendation list for the target content item originally exists, the newly generated recommendation list is directly stored. And when the user clicks and consumes the target content item at the later stage, recommending the currently stored recommendation list of the target content item to the user so as to display the recommendation item in the recommendation list.

Moreover, the content item recommendation method provided by the embodiment of the invention can be applied to electronic equipment, and the electronic equipment can be terminal equipment or a server.

As shown in fig. 1, the content item recommendation method includes:

s101, determining a plurality of target items to be recommended from items to be recommended corresponding to the target content items, wherein the similarity corresponding to the plurality of target items to be recommended is larger than the similarity corresponding to other items to be recommended, and the similarity corresponding to any item is the similarity between the item and the target content items.

The other items to be recommended in the embodiment of the present invention refer to items to be recommended, except for a plurality of target recommended items, in the items to be recommended corresponding to the target content items. Each item to be recommended and the target content item respectively correspond to a similarity, the similarity represents the association degree between the item to be recommended and the target content item, and the greater the similarity, the higher the association degree between the item to be recommended and the target content item. The similarity can be calculated in an existing manner, and the embodiment of the invention is not described in detail.

The item to be recommended of the target content item refers to a content item which can be recommended to the user by the recommendation system after the user clicks the target content item and can be continuously clicked by the user, for example, for all audio stored by the audio portal site and used for being consumed by the user, the target content item may be any one of all stored audio, and the item to be recommended of the target content item may be part or all of other audio except the audio of the target content item in all stored audio.

It can be understood that, from the comparison between the items to be recommended of the target content item, if the item to be recommended with the greater similarity to the target content item is taken as the recommended item of the target content item, the probability of being consumed by clicking is greater; conversely, if the similarity between the item to be recommended and the target content item is smaller, the recommended item as the target content item is less likely to be clicked for consumption. Therefore, the target item to be recommended with larger similarity to the target content item is selected according to the step, and the effectiveness of the recommendation list consisting of the recommendation items is improved.

In an implementation manner of the embodiment of the present invention, the step may be performed as follows:

(1) obtaining content information (such as labels, titles and the like) aiming at the target content item and the item to be recommended of the target content item, and obtaining preference information of the user (such as marking the preference information of the user on the content item by utilizing explicit behaviors of purchase, approval and the like, and implicit behaviors of browsing, watching and the like of the user);

for example, the target content item is the tenth album of the "conututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututut:

watch 1

(2) Calculating the similarity between the target content item and each item to be recommended according to the obtained content information and preference information;

how to calculate the similarity belongs to the prior art, and is not described herein.

(3) And determining a plurality of target items to be recommended according to the similarity among the items to be recommended corresponding to the target content items, wherein the similarity between the target items to be recommended and the target content items is greater than the similarity between other items to be recommended and the target content items.

In specific implementation, all items to be recommended corresponding to the target content item may be sorted according to the magnitude of the corresponding similarity, and a certain number of items to be recommended are sequentially selected as the target items to be recommended. Of course, a similarity threshold may also be set, and a plurality of items to be recommended, of which the corresponding similarity is greater than the similarity threshold, are selected as a plurality of target items to be recommended from the items to be recommended corresponding to the target content items.

S102, calculating the profit capacity corresponding to each target item to be recommended, wherein the profit capacity corresponding to any target item to be recommended represents the possibility that a user clicks a recommendation list of the target item to be recommended when the target item to be recommended is used as a recommendation item of a target content item;

it will be appreciated that the purpose of recommending content items to a user via a recommendation list is to expect that the user can consume in multiple clicks, the more clicks the user consumes, the more revenue the provider receives accordingly. The profit capacity can be represented by a numerical value, the numerical value is large and represents the capacity, the profit capacity and the small and represents the weak profit capacity.

The target item to be recommended has strong profit capacity, when the representative target item to be recommended is used as the recommended item of the target content item, the probability that the user clicks the recommendation list of the target item to be recommended is higher, actually, the probability that the representative user clicks the target item to be recommended through the recommendation list of the target content item and clicks the content item of the recommendation list of the target item to be recommended is higher, that is, the probability that the user continuously consumes is higher. Therefore, the target item to be recommended has strong profit capacity, and when the recommendation list of the target content item includes the target item to be recommended, the probability that the user can continuously consume is higher, which indicates that the effectiveness of the recommendation list of the target content item is higher.

On the contrary, the profit capacity corresponding to the target item to be recommended is weak, and when the representative target item to be recommended is used as the recommended item of the target content item, the probability that the user clicks the recommendation list of the target item to be recommended is smaller, actually, the probability that the representative user clicks the target item to be recommended through the recommendation list of the target content item and clicks the content item of the recommendation list of the target item to be recommended is smaller, that is, the probability that the user continuously consumes is smaller. Therefore, the target item to be recommended has weak profit capacity, and when the recommendation list of the target content item includes the target item to be recommended, the probability that the user may continuously consume is smaller, which indicates that the recommendation list of the target content item has lower effectiveness.

Therefore, if the target item to be recommended with strong corresponding profit capacity is taken as the recommended item of the target content item, the effectiveness of the recommendation list of the target content item is improved.

In specific application, statistics can be carried out by aiming at the historical recommendation data of the target content item, and when the target item to be recommended is actually taken as the recommendation item of the target content item in the past, the user clicks the consumed data. And according to the statistical data, predicting and calculating the income capacity corresponding to the target item to be recommended.

For clarity of the scheme and clear layout, a specific implementation manner for calculating the profit capacity corresponding to each target item to be recommended is introduced subsequently.

S103, calculating an effective score corresponding to each target item to be recommended, wherein the effective score corresponding to any target item to be recommended is obtained by calculation according to the similarity and the profit capacity corresponding to the target item to be recommended;

the effective score is obtained by calculation according to the similarity and the profit capacity corresponding to the target item to be recommended, and the calculated effective score comprehensively considers the similarity and the profit capacity corresponding to the target item to be recommended, so that the calculated effective score can reflect the interest degree and the click possibility of the content item to the user, namely the attraction degree of the content item to the user.

And S104, determining recommended items corresponding to the target content items from the target items to be recommended according to the calculated effective scores, and generating a recommendation list comprising the recommended items, wherein the effective scores corresponding to the recommended items are all larger than those of other target items to be recommended.

It is understood that a higher effective score represents a higher likelihood of a higher similarity and a higher profitability, and that generating a recommendation list from a plurality of recommendation items having higher effective scores may improve the effectiveness of the recommendation list.

By applying the embodiment shown in fig. 1, through the determination of the target items to be recommended, it can be ensured that each target item to be recommended has a certain similarity with the target content items, and the target items to be recommended may be more interested by the user than the content items with lower similarities. Because the similarity can reflect the interest degree of the user to the target item to be recommended and the calculated income ability can reflect the possibility of continuous click consumption of the user, the effective score calculated according to the similarity and the income ability can comprehensively reflect the attraction degree of the corresponding content item to the user. Therefore, the recommendation list is generated according to the recommendation item with the larger effective score, and the effectiveness of the recommendation list can be improved.

In addition, by applying the technical scheme provided by the embodiment of the invention, the related recommendation can be changed into a line from a point, so that the deeper the user walks along the line of the related recommendation, the larger the excavation space is, and the continuity of one session is improved. When the method is applied to the continuous broadcasting of the short video, the video transmission capacity, the access depth and the stay time of the user can be improved through the cascade capacity of the video, and therefore obvious benefit increase can be obtained.

Optionally, in an implementation manner of the embodiment of the present invention, as shown in fig. 2, the step of calculating the profit-making capability corresponding to each target item to be recommended may include:

s201, collecting a recommendation log of each target item to be recommended.

The recommendation log in this embodiment may be used to record all content items clicked by the user and all displayed recommendation lists.

S202, according to the recommendation log, under the condition that each target item to be recommended is used as a recommendation item of a target content item, after the target item to be recommended is clicked, the display times of a recommendation list for displaying the target item to be recommended and the effective click times of the recommendation list for displaying the target item to be recommended are calculated, wherein the effective click times refer to the times of clicks meeting preset requirements.

It can be understood that, in the recommendation history recorded in the recommendation log, the target item to be recommended is once taken as the recommended item of the target content item and clicked by the user, the recommendation system may present the recommendation list of the target item to be recommended to the user, and the recommendation list of the target item to be recommended that is presented each time is not clicked by the user, so the presentation times and the effective click times mentioned herein may be different.

In addition, the preset requirement is a preset requirement, for example, any click of a content item of the recommendation list for the target item to be recommended is a valid click, i.e., a click that meets the preset requirement. For another example, when video recommendation is performed, a long click may be extracted as an effective click, that is, a click meeting a preset requirement, that is, a video of a recommendation list for a target item to be recommended, and as long as a physical duration of viewing after the click exceeds a certain threshold or a physical duration of viewing exceeds a certain proportion of a total duration of corresponding videos, the click is considered as a long click.

Example one, the first set of "conutus trails" is a target content item, and the second set of "conutus trails" and the full set of advance videos of "conutus trails" are target to-be-recommended items. The recommended log records for the first set of the "conutleaves biography" are: in the process of watching the first set of the "conutus tradition" by 50 users, the recommendation system displays a first recommendation list of the first set of the "conutus tradition" to the 50 users, and the first recommendation list comprises prejudice videos of a second set of the "conutus tradition" and a complete set of the "conutus tradition" in the first recommendation list. 28 of the 50 users click to watch the complete set forecast video of the Coconutututlet, the recommendation system continuously recommends a second recommendation list to the 28 users, and the second recommendation list comprises a second set and a third set of the Coconutlet; another 20 users click on the second set of conutus, and the recommendation system continues to recommend a third recommendation list to the 20 users, where the third recommendation list includes the third set and the fourth set of conutus. Of the 28 users who viewed the second recommendation list, 27 clicked the second set of "conutus trails" through the second recommendation list, and of the 20 users who viewed the third recommendation list, 2 clicked the third set of "conutus trails" through the third recommendation list. Wherein, in addition, it is specified that any click on the content item belongs to a click that satisfies a preset requirement.

According to the recommendation log of the first set of the conutleaves biography, the number of times that the recommendation list, namely the third recommendation list, of the second set of the conutleaves biography is displayed after the second set of the conutleaves biography is clicked is 20, and the number of times that the third recommendation list is clicked effectively is 2. The number of times of displaying the recommendation list, namely the second recommendation list, after the complete set preview video of the coconut and conutleaves biography is obtained and clicked is 28, and the effective click number of the second recommendation list is 27.

S203, calculating the ratio of the effective click times corresponding to each target item to be recommended to the corresponding display times of each target item to be recommended;

it can be understood that the larger the ratio calculated in this step is, the higher the probability that the recommendation list of the target item to be recommended is clicked after the target item to be recommended is clicked as the recommendation item of the target content item in the recommendation history is, and according to this trend, if the target item to be recommended is further used as the recommendation item of the target content item, the higher the probability that the user clicks the recommendation list of the target item to be recommended is.

On the contrary, after the target item to be recommended is clicked as the recommended item of the target content item, the probability that the recommendation list of the target item to be recommended is clicked is low, and according to the trend, if the target item to be recommended is used as the recommended item of the target content item, the probability that the user clicks the recommendation list of the target item to be recommended is also low.

Therefore, the ratio calculated in this step can well represent the possibility that the user clicks the recommendation list of the target item to be recommended when the target item to be recommended is taken as the recommendation item of the target content item.

Example two, continuing with the above example one, the number of times that the third recommendation list, which is the recommendation list after the second set of the conutututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututututut. The number of times of displaying the recommendation list, namely the second recommendation list, after the complete set preview video of the coconut and conutus tradition is clicked is 28, the effective click number of the second recommendation list is 27, and the calculated ratio is 27/28.

S204, determining the ratio corresponding to each target item to be recommended as the profit capacity of the corresponding target item to be recommended.

Since the ratio calculated in the previous step can well represent the possibility that the user clicks the recommendation list of the target item to be recommended when the target item to be recommended is used as the recommendation item of the target content item, the generated recommendation list can be more effective by using the ratio as the profit capacity of the corresponding target item to be recommended.

Example two, continuing with example two above, the ratio calculated for the second set of the conutus trails is 1/10, and the ratio calculated for the full set of the predictive video of the conutus trails is 27/28. The ratio of the latter is larger than that of the former, and the latter is used as a recommendation item of the first set of the Coconutus conutus, so that the continuous consumption of the user can be promoted, and the effectiveness of a recommendation list can be improved. It will be appreciated that typically the highlights of a tv series are not the first few episodes, and that if recommendations are made only in the order of episodes, the user may abandon the series in the past. If the user is recommended to watch the following highlight drama forecast in the early stage and then the continuous episode is recommended, the user can be attracted to continue consuming.

Optionally, in an implementation manner of the embodiment of the present invention, as shown in fig. 3, the step of calculating the profit-making capability corresponding to each target item to be recommended includes:

s301, collecting a recommendation log of each target item to be recommended;

S302, according to the recommendation log, aiming at each target item to be recommended, calculating the display times of an effective recommendation list of the target content item after the target content item is clicked and the effective click times of the target item to be recommended in all the effective recommendation lists, wherein the effective recommendation list is a recommendation list when the target item to be recommended is used as the recommendation item of the target content item, and the effective click times are the click times meeting the preset requirements;

it can be understood that, in the recommendation history recorded in the recommendation log, the target item to be recommended is once presented to the user in the form of a recommendation list of the target item to be recommended, and the user does not click the target item to be recommended in the recommendation list every time the user is presented, so the presentation times and the effective click times mentioned herein may be different.

In addition, the preset requirement is a preset requirement, for example, for a recommendation list containing target items to be recommended, which is shown by a target content item, as long as all the target items to be recommended in the corresponding recommendation list are clicked effectively, that is, clicks meeting the preset requirement are clicked. For another example, when video recommendation is performed, a long click may be extracted as an effective click, that is, a click meeting a preset requirement, that is, a recommendation list containing a target item to be recommended, which is displayed for a target content item, and as long as after the target item to be recommended is clicked, the physical viewing duration exceeds a certain threshold or the physical viewing duration exceeds the total duration of a corresponding video by a certain proportion, the click is considered as a long click.

S303, calculating the ratio of the effective click times corresponding to each target item to be recommended to the corresponding display times of each target item to be recommended;

it is understood that the larger the ratio calculated in this step is, the higher the probability that the target item to be recommended is clicked as the recommended item of the target content item in the recommendation history is. According to this trend, if the target item to be recommended is taken as the recommended item of the target content item again, the probability that the user clicks the target item to be recommended is higher. Since it is possible to continue clicking the content item of the recommendation list of the target item to be recommended only after the target item to be recommended is clicked, the higher the probability that the target item to be recommended is clicked as the recommendation item of the target content item is, the higher the probability that the user continues clicking the recommendation list of the target item to be recommended is.

Conversely, the smaller the ratio calculated in this step is, the lower the probability that the target item to be recommended is clicked as the recommended item of the target content item in the recommendation history is. According to this trend, if the target item to be recommended is taken as the recommended item of the target content item again, the probability that the user clicks the target item to be recommended is also lower. Since it is possible to continue clicking the content item of the recommendation list of the target item to be recommended only after the target item to be recommended is clicked, the lower the probability that the target item to be recommended is clicked as the recommendation item of the target content item is, the lower the probability that the user continues clicking the recommendation list of the target item to be recommended is.

Therefore, the ratio calculated in the step can well represent the possibility that the user clicks the recommendation list of the target item to be recommended when the target item to be recommended is taken as the recommendation item of the target content item.

S304, determining the ratio corresponding to each target item to be recommended as the profit capacity of the corresponding target item to be recommended.

Based on any one of the two specific implementation manners for calculating the profit potential corresponding to each target item to be recommended, the content item recommendation method further includes:

It should be noted that, if the ratio calculation fails, the profit capacity corresponding to the target item to be recommended cannot be correctly obtained. For example, in the early stage of application of the embodiment of the present invention, a recommendation log regarding the target content item has not been generated, and at this time, the ratio cannot be calculated by data recorded by the recommendation log. In addition, the step is specifically executed before the step of determining the ratio corresponding to each target item to be recommended as the profit capacity of the corresponding target item to be recommended, and finally the purpose of taking the preset value as the profit capacity of the target item to be recommended can be achieved. Through the setting of the preset value, which is equivalent to the initialization of the ratio, further, the influence of the preset value on the actual effectiveness of the recommendation list can be eliminated through the continuous application of the embodiment of the invention.

From the above analysis, through the setting of this step, when the embodiment of the present invention is applied to generating a recommendation list, the generation of the recommendation list can still be achieved even if the ratio calculation fails.

Optionally, in an implementation manner of the embodiment of the present invention, the step of calculating the effective score corresponding to each target item to be recommended includes:

The effective score is calculated by means of weighted summation, so that the process of obtaining the effective score is simpler. The used weight value can be preset, and can be specifically set according to the relative importance degree of the two parameters of the similarity and the profit capacity, and the relative importance degree of the two parameters can be determined according to the actual situation, so that the relative importance degree of the similarity and the profit capacity can be conveniently adjusted through the implementation of the step.

Optionally, in an implementation manner of the embodiment of the present invention, the step of generating a recommendation list including the recommendation item includes:

and sorting the recommendation items according to the effective scores of the recommendation items, and forming a recommendation list of the target content item by the sorted recommendation items.

According to the analysis process, the higher the effective score is, the higher the probability that the user clicks the corresponding recommended item is, and the recommended items with higher effective scores are arranged in the recommendation list in front of the recommendation list more easily to draw the attention of the user, so that the effectiveness of the recommendation list is more easily improved.

On the basis of the above method embodiment, an embodiment of the present invention further provides a content item recommendation apparatus, as shown in fig. 4, the apparatus including:

a first determining module 41, configured to determine a plurality of target items to be recommended from items to be recommended corresponding to a target content item, where the similarity corresponding to each of the plurality of target items to be recommended is greater than the similarity corresponding to other items to be recommended, and the similarity corresponding to any item is the similarity between the item and the target content item;

the first calculating module 42 is configured to calculate a profit capacity corresponding to each target item to be recommended, where the profit capacity corresponding to any target item to be recommended represents a possibility that the user clicks a recommendation list of the target item to be recommended when the target item to be recommended is used as a recommendation item of the target content item;

a second calculating module 43, configured to calculate an effective score corresponding to each target item to be recommended, where the effective score corresponding to any target item to be recommended is calculated according to the similarity and the profit capacity corresponding to the target item to be recommended;

and the processing module 44 is configured to determine, according to the calculated effective scores, recommended items corresponding to the target content items from the multiple target items to be recommended, and generate a recommendation list including the recommended items, where the effective scores corresponding to the recommended items are all greater than those of other target items to be recommended.

By applying the embodiment shown in fig. 4, through the determination of the target items to be recommended, it can be ensured that each target item to be recommended has a certain similarity with the target content items, and the target items to be recommended may be more interested by the user than the content items with lower similarities. Because the similarity can reflect the interest degree of the user to the target item to be recommended and the calculated income ability can reflect the possibility of continuous click consumption of the user, the effective score calculated according to the similarity and the income ability can comprehensively reflect the attraction degree of the corresponding content item to the user. Therefore, the recommendation list is generated according to the recommendation item with the larger effective score, and the effectiveness of the recommendation list can be improved.

In an implementation manner of the embodiment of the present invention, as shown in fig. 5, the first calculating module 42 includes:

a first collecting unit 51 for collecting a recommendation log of each target item to be recommended;

the first processing unit 52 is configured to calculate, according to the recommendation log, for each target item to be recommended, when the target item to be recommended is used as a recommendation item of a target content item, the number of times that the recommendation list of the target item to be recommended is displayed after the target item to be recommended is clicked, and the number of times that the recommendation list of the target item to be recommended is displayed is effectively clicked, where the number of times that the effective click is the number of times that a preset requirement is met;

the first calculating unit 53 is configured to calculate, for each target item to be recommended, a ratio of an effective click number corresponding to the target item to be recommended to a corresponding display number;

the first determining unit 54 is configured to determine a ratio corresponding to each target item to be recommended as the profit capacity of the corresponding target item to be recommended.

In an implementation manner of the embodiment of the present invention, as shown in fig. 6, the first calculating module 42 of the implementation manner includes:

a second collecting unit 61 for collecting a recommendation log of each target item to be recommended;

the second processing unit 62 is configured to calculate, according to the recommendation log, for each target item to be recommended, the number of times that an effective recommendation list of the target content item is displayed after the target content item is clicked, and the number of times that the target item to be recommended is clicked in all the effective recommendation lists, where an effective recommendation list is a recommendation list when the target item to be recommended is used as a recommendation item of the target content item, and the number of times that the effective click is clicked meets a preset requirement;

the second calculating unit 63 is configured to calculate, for each target item to be recommended, a ratio of an effective click number corresponding to the target item to be recommended to a corresponding display number;

the second determining unit 64 is configured to determine a ratio corresponding to each target item to be recommended as the profit capacity of the corresponding target item to be recommended.

In an implementation manner of the embodiment of the present invention, the content item recommendation apparatus further includes:

In an implementation manner of the embodiment of the present invention, the second calculating module 43 includes:

In an implementation manner of the embodiment of the present invention, when the processing module is configured to generate the recommendation list including the recommendation items, the processing module is specifically configured to sort the recommendation items according to the validity score of each recommendation item, and combine the sorted recommendation items into the recommendation list of the target content item.

Fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, where the electronic device includes a processor 71, a communication interface 72, a memory 73, and a communication bus 74, where the processor 71, the communication interface 72, and the memory 73 complete mutual communication through the communication bus 74,

a memory 73 for storing a computer program;

the processor 71, when executing the program stored in the memory 73, implements the following steps:

determining a plurality of target items to be recommended from items to be recommended corresponding to the target content items, wherein the similarity corresponding to the plurality of target items to be recommended is greater than the similarity corresponding to other items to be recommended, and the similarity corresponding to any item is the similarity between the item and the target content items;

calculating the profit capacity corresponding to each target item to be recommended, wherein the profit capacity corresponding to any target item to be recommended represents the possibility that a user clicks a recommendation list of the target item to be recommended when the target item to be recommended is used as a recommendation item of the target content item;

calculating an effective score corresponding to each target item to be recommended, wherein the effective score corresponding to any target item to be recommended is calculated according to the similarity and the profit capacity corresponding to the target item to be recommended;

and determining recommended items corresponding to the target content items from the target items to be recommended according to the calculated effective scores, and generating a recommendation list comprising the recommended items, wherein the effective scores corresponding to the recommended items are all larger than those of other target items to be recommended.

It should be noted that, for specific implementation and related explanation of each step of the method, reference may be made to the above-mentioned method embodiments, which are not described herein again.

The communication bus mentioned in the electronic device may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.

The communication interface is used for communication between the electronic equipment and other equipment.

The Memory may include a Random Access Memory (RAM) or a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.

The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a network Processor (Ne word Processor, NP), and the like; the integrated Circuit may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component.

In yet another embodiment provided by the present invention, there is also provided a computer-readable storage medium having stored therein instructions, which when run on a computer, cause the computer to perform the content item recommendation method of any of the above embodiments.

In a further embodiment provided by the present invention, there is also provided a computer program product comprising instructions which, when run on a computer, cause the computer to perform the method of content item recommendation described in any of the above embodiments.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

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.

All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.

The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims

1. A method of content item recommendation, the method comprising:

calculating the profit capacity corresponding to each target item to be recommended, wherein the profit capacity corresponding to any target item to be recommended represents the possibility that a user clicks a recommendation list of the target item to be recommended when the target item to be recommended is used as the recommendation item of the target content item; the recommendation list of the target item to be recommended is a recommendation list displayed after the target item to be recommended is clicked by a user;

2. The recommendation method according to claim 1, wherein the step of calculating the profitability corresponding to each target item to be recommended comprises:

collecting a recommendation log of each target item to be recommended;

3. The recommendation method according to claim 1, wherein the step of calculating the profitability corresponding to each target item to be recommended comprises:

collecting a recommendation log of each target item to be recommended;

4. The recommendation method according to claim 2 or 3, characterized in that the method further comprises:

5. The recommendation method according to claim 1, wherein the step of calculating the effective score corresponding to each target item to be recommended comprises:

6. The recommendation method according to claim 1, wherein the step of generating a recommendation list including the recommended items comprises:

7. An apparatus for recommending content items, the apparatus comprising:

the first calculation module is used for calculating the profit capacity corresponding to each target item to be recommended, wherein the profit capacity corresponding to any target item to be recommended represents the possibility that a user clicks the recommendation list of the target item to be recommended when the target item to be recommended is used as the recommendation item of the target content item; the recommendation list of the target item to be recommended is a recommendation list displayed after the target item to be recommended is clicked by a user;

8. The recommendation device of claim 7, wherein the first computing module comprises:

9. The recommendation device of claim 7, wherein the first computing module comprises:

10. The recommendation device according to claim 8 or 9, further comprising:

11. The recommendation device of claim 7, wherein the second calculation module comprises:

12. The recommendation device of claim 7,

the processing module is configured to, when generating a recommendation list including the recommendation items, specifically sort the recommendation items according to the validity scores of the recommendation items, and compose the sorted recommendation items into the recommendation list of the target content item.

13. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing mutual communication by the memory through the communication bus;

a memory for storing a computer program;

a processor for implementing the method steps of any of claims 1-6 when executing a program stored in the memory.