CN105160008B - Method and device for positioning recommended user - Google Patents
Method and device for positioning recommended user Download PDFInfo
- Publication number
- CN105160008B CN105160008B CN201510604634.2A CN201510604634A CN105160008B CN 105160008 B CN105160008 B CN 105160008B CN 201510604634 A CN201510604634 A CN 201510604634A CN 105160008 B CN105160008 B CN 105160008B
- Authority
- CN
- China
- Prior art keywords
- user
- channels
- specified
- data
- training
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/7867—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, title and artist information, manually generated time, location and usage information, user ratings
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Abstract
The invention discloses a method and a device for positioning a recommended user, which collect user transaction data and user video watching data in a video system; extracting a specified characteristic data set of a training user and a specified characteristic data set of a testing user from the collected user transaction data and user video watching data; training each designated feature in a designated feature data set of a training user according to a training algorithm to obtain a weight value of each designated feature; determining probability prediction data paid by each test user according to the specified feature data set of the test user and the weight value of each specified feature obtained by training; according to the probability prediction data of the payment of each test user, the recommended users are positioned, users with payment tendency can be more accurately selected from the test set users, and the recommended users can be more accurately positioned.
Description
Technical Field
The invention relates to the technical field of videos, in particular to a method and a device for positioning a recommended user.
Background
Video generally refers to various techniques for capturing, recording, processing, storing, transmitting, and reproducing a series of still images as electrical signals. When the continuous image changes more than 24 frames (frames) per second, the human eye cannot distinguish a single still image according to the principle of persistence of vision, and therefore, the continuous image appears as a smooth continuous visual effect when played, and such a continuous image is called a video.
One of the important ways of improving the revenue of the current video website is to promote the user to consume, such as pushing advertisements to the user, but the current recommendation system generally has the characteristics of wide broadcasting network and low conversion rate, and how to more accurately locate the user with higher payment tendency is a problem that needs to be solved urgently in the industry.
Disclosure of Invention
In view of the foregoing problems, the present application provides a method and an apparatus for locating a recommending user, so as to more accurately locate the recommending user.
In order to solve the technical problem, the following technical scheme is adopted in the application:
a method of locating a recommending user, comprising:
collecting user transaction data and user video watching data in a video system;
extracting a specified characteristic data set of a training user and a specified characteristic data set of a testing user from the collected user transaction data and user video watching data;
training each designated feature in a designated feature data set of a training user according to a training algorithm to obtain a weight value of each designated feature;
determining probability prediction data paid by each test user according to the specified feature data set of the test user and the weight value of each specified feature obtained by training;
and positioning the recommended users according to the probability prediction data paid by each test user.
Wherein, gather video system user transaction data and user's video and watch data and include:
acquiring a log of a member transaction table to obtain user transaction data of a video system;
and acquiring a user video watching log to obtain user transaction data of the video system.
Wherein the member transaction form log and the user video viewing log comprise logs corresponding to a personal computer and a wireless terminal.
The method for determining probability prediction data paid by each test user according to the specified feature data set of the test user and the weight value of each specified feature obtained by training comprises the following steps:
sorting the weight values of the obtained specified characteristics;
determining key designated features in a designated sorting range according to a sorting result;
and determining probability prediction data paid by each test user according to the weight of the key specified features and the specified feature data set of the test user.
The positioning of the recommended users according to the probability prediction data of the payment of each test user comprises the following steps:
determining a first threshold curve of a recommended user according to the probability prediction data paid by each test user and the recommendation accuracy within a positioning time period;
determining a second threshold curve of the recommended user according to the probability prediction data paid by each test user and the recommendation efficiency in the positioning time period;
determining a threshold value of a recommended user according to the first threshold value curve and the second threshold value curve;
and positioning the recommended users according to the determined threshold value of the recommended users.
The appointed characteristic data set of the training user comprises positive sample data and negative sample data, the positive sample data is the appointed characteristic data set of the user who pays at the appointed time point, and the negative sample data is the appointed characteristic data set of the user who pays at the appointed time point.
Wherein the specified time points are the day of positive and negative sample collection.
Wherein the number of negative sample data is three times the number of positive sample data.
Wherein the user who pays the purchase is a user who purchases a member.
Wherein the training algorithm is an L2 regular logistic regression training algorithm.
Wherein the specified characteristics are one or more of the following:
movie channels, drama channels, car channels, fun channels, animation channels, entertainment channels, fashion channels, parent-child channels, game channels, creative channels, advertisement channels, music channels, information channels, sports channels, living channels, travel channels, science channels, education channels, entertainment channels, documentary channels, other channels, android devices, iphone devices, ipad devices, ipod devices, other devices, members, non-members, pay videos, free videos, complete viewing, and trial viewing.
The present application further provides an apparatus for locating a recommended user, which includes:
the acquisition module is used for acquiring user transaction data and user video watching data in the video system;
the extraction module is used for extracting a specified characteristic data set of a training user and a specified characteristic data set of a testing user from the collected user transaction data and user video watching data;
the training module is used for training each specified feature in the specified feature data set of the training user according to a training algorithm to obtain a weight value of each specified feature;
the determining module is used for determining probability prediction data paid by each test user according to the specified feature data set of the test user and the weight value of each specified feature obtained by training;
and the positioning module is used for positioning the recommended users according to the probability prediction data paid by each test user.
Wherein, the collection module includes:
the first acquisition submodule is used for acquiring a member transaction table log to obtain user transaction data of the video system;
and the second acquisition submodule is used for acquiring the video watching log of the user to obtain the user transaction data of the video system.
Wherein the member transaction form log and the user video viewing log comprise logs corresponding to a personal computer and a wireless terminal.
Wherein the determining module comprises:
the ordering submodule is used for ordering the weighted values of the obtained specified characteristics;
the key specified characteristic determining submodule is used for determining key specified characteristics in a specified sequencing range according to a sequencing result;
and the probability prediction data determining submodule is used for determining probability prediction data paid by each test user according to the weight of the key specified characteristics and the specified characteristic data set of the test user.
Wherein, the orientation module includes:
the first threshold curve determining submodule is used for determining a first threshold curve of a recommended user according to the probability prediction data paid by each test user and the recommendation accuracy within the positioning time period;
the second threshold curve determining submodule determines a second threshold curve of the recommended user according to the probability prediction data paid by each test user and the recommendation efficiency in the positioning time period;
the threshold value determining submodule determines a threshold value of a recommended user according to the first threshold value curve and the second threshold value curve;
and the recommended user positioning submodule is used for positioning the recommended user according to the determined threshold value of the recommended user.
The appointed characteristic data set of the training user comprises positive sample data and negative sample data, the positive sample data is the appointed characteristic data set of the user who pays at the appointed time point, and the negative sample data is the appointed characteristic data set of the user who pays at the appointed time point.
Wherein the specified time points are the day of positive and negative sample collection.
Wherein the number of negative sample data is three times the number of positive sample data.
Wherein the user who pays the purchase is a user who purchases a member.
Wherein the training algorithm is an L2 regular logistic regression training algorithm.
Wherein the specified characteristics are one or more of the following:
movie channels, drama channels, car channels, fun channels, animation channels, entertainment channels, fashion channels, parent-child channels, game channels, creative channels, advertisement channels, music channels, information channels, sports channels, living channels, travel channels, science channels, education channels, entertainment channels, documentary channels, other channels, android devices, iphone devices, ipad devices, ipod devices, other devices, members, non-members, pay videos, free videos, complete viewing, and trial viewing.
According to the method and the device for positioning the recommended users, user transaction data and user video watching data in a video system are collected; extracting a specified characteristic data set of a training user and a specified characteristic data set of a testing user from the collected user transaction data and user video watching data; training each designated feature in a designated feature data set of a training user according to a training algorithm to obtain a weight value of each designated feature; determining probability prediction data paid by each test user according to the specified feature data set of the test user and the weight value of each specified feature obtained by training; and positioning the recommended users according to probability prediction data paid by each test user, wherein the designated feature data set is a set of data with frequent user use behaviors on designated features, the weight values of the designated features can be accurately obtained for the designated feature data set of the training users through a training algorithm, the test set users are further determined according to the weight values of the designated features, users with payment tendency can be more accurately selected from the test set users, and the recommended users can be more accurately positioned.
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, it is obvious that the drawings in the following description are only some embodiments described in the present invention, and other drawings can be obtained by those skilled in the art according to the drawings.
FIG. 1 is a flowchart of one embodiment of a method for locating a referring user according to the present application. (ii) a
FIG. 2 is a schematic diagram of the overall composition of an apparatus for locating recommended users according to the present application;
FIG. 3 is a schematic diagram of the acquisition module of FIG. 2 according to one embodiment;
FIG. 4 is a schematic diagram illustrating the composition of one embodiment of the determination module of FIG. 2;
FIG. 5 is a schematic diagram of an embodiment of the orientation module of FIG. 2.
Detailed Description
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 only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments that can be derived by one of ordinary skill in the art from the embodiments given herein are intended to be within the scope of the present invention.
Please refer to fig. 1, which is a flowchart illustrating a method for locating a recommended user according to an embodiment of the present invention. In this embodiment, the method for locating a recommended user mainly includes the following steps:
step S101, collecting user transaction data and user video watching data in a video system;
in particular, the collection of the user transaction data and the user video viewing data of the video system can be realized in various ways, and as a preferred embodiment, for example, the following ways can be adopted:
acquiring a log of a member transaction table to obtain user transaction data of a video system; and acquiring a user video watching log to obtain user transaction data of the video system.
It should be noted that the log of the member transaction table and the log of the user video viewing may include logs corresponding to a personal computer and a wireless terminal, so that the obtained data is more accurate for analyzing the user viewing behavior.
Step S102, extracting a specified characteristic data set of a training user and a specified characteristic data set of a testing user from the collected user transaction data and user video watching data;
for a user who pays for a purchase (e.g., a user who purchases a member), some features of the usage behavior features of the user in the video website are generally strongly associated with the purchase of the user, in this embodiment, the specified feature is a feature associated with the purchase behavior of the user, the specified feature dataset of the user is a set of data of the frequency of the usage behavior of the user on the specified feature, for example, the user who pays for the purchase relates to watching a movie channel, the movie channel is a specified feature, the frequency of the usage behavior of the user on the specified feature is the frequency of the user watching the movie channel within a statistical time period, the extracted specified feature dataset is the frequency data of the usage behavior of the user on the specified feature within the statistical time period, for example, the statistical time period is 60 days, the extracted feature dataset of the movie channel needs to be stored for the number of watching the movie channel within 60 days by the category of the user, as one particular example, the specified characteristics associated with the user's pay-for-purchase behavior may include one or more of the following characteristics, for example:
a movie channel, a drama channel, an automobile channel, a fun channel, an animation channel, a variety channel, a fashion channel, a parent-child channel, a game channel, an original channel, an advertisement channel, a music channel, an information channel, a sports channel, a living channel, a travel channel, a science channel, an education channel, an entertainment channel, a documentary channel, other channels, an android device, an iphone device, an ipad device, an ipod device, other devices, members, non-members, pay videos, free videos, complete watching and trial watching, etc., and more specified features may be added as the number of video watching channels increases or the number of mobile devices used increases and the number of related identities and services increases, which are merely exemplified and not particularly limited to the above features.
In addition, the specified feature data set of the user extracted in this step is divided into a specified feature data set of a training user and a specified feature data set of a testing user, wherein the specified feature data set of the training user comprises positive sample data and negative sample data, the negative sample data is the specified feature data set of the user who does not pay for purchasing at a specified time point, when the specific implementation is implemented, the proportion between the positive sample and the negative sample can be adjusted according to the actual situation, for example, the quantity of the negative sample data is three times that of the positive sample data, in addition, as a specific embodiment, the positive sample data is the specified feature data set of the user who pays for purchasing at a specified time point, for example, the probability of purchasing the member by the user at the next day is predicted, the specified time point is the same day, so that the training and probability prediction data have the same timeliness, wherein the quantity of the positive sample is the number of purchasing the member on the, the number of people is 1.5 ten thousand; meanwhile, the number of people who do not purchase the members is far more than that who purchase the members every day, so that the number of people with negative samples is more than that of people with positive samples, and 5 ten thousand of people can be selected as the negative samples, namely, the number of the negative samples is about 3 times that of the people with positive samples.
Step S103, training each designated feature in the designated feature data set of the training user according to a training algorithm to obtain a weight value of each designated feature;
in the present application, various existing training algorithms may be used for training, which is not limited herein, and by way of example only, for example, the training algorithm may use an L2 regular logistic regression training algorithm, and the L2 regular logistic regression training algorithm is also referred to as an L2 regular logistic algorithm, which is widely used in statistics, in this embodiment, the weight value of each designated feature may be obtained by training the designated feature data set extracted in the step S102 to obtain the designated feature data set of the training user, for example, the weight value of each designated feature may be obtained by using the 1.5 ten thousand positive sample data and the 5 ten thousand negative sample data as input data and training the training algorithm, for example, the L2 regular logistic regression training algorithm, if the total weight score is 100, the weight of the movie channel in the designated feature is 8, the weight of the tv drama channel is 10, and the weights of other designated features may also be obtained, this is merely an example and will not be described in detail.
Step S104, determining probability prediction data paid by each test user according to the specified feature data set of the test user and the weight value of each specified feature obtained by training;
in a specific implementation, the weighted values of the designated features obtained in step S103 are different in size, and some weighted values of the designated features may be smaller and may not be used for prediction, that is, in this step, all weighted values of the designated features may be used for prediction, and a part of weighted values of the designated features having a larger weight may also be used for prediction, for example, one way is as follows:
sorting the weight values of the obtained specified characteristics;
determining key specified features in a specified sequencing range according to the sequencing result, for example, using the specified features in the sequencing range ranked in the top ten as the key specified features;
and determining probability prediction data paid by each test user according to the weight of the key specified features and the specified feature data set of the test user.
And S105, positioning the recommended users according to the probability prediction data paid by each test user.
In the specific implementation, the recommended users can be located according to the size of the probability prediction data according to the probability prediction data paid by each test user, but a high conversion rate needs to be achieved under the condition of a low coverage rate, for this reason, the accuracy rate is assumed to be the ratio of the predicted correct number of people/the actual number of purchased people, namely the predicted correct number of people to the actual number of paid people on the same day, and the efficiency is assumed to be the predicted correct number of people/the predicted number of purchased people, namely the predicted correct number of paid people and the predicted number of people who will purchase members, in the embodiment, the recommended users can be located according to the probability prediction data paid by each test user in the following mode, that is:
determining a first threshold curve of a recommended user according to the probability prediction data paid by each test user and the recommendation accuracy within a positioning time period;
determining a second threshold curve of the recommended user according to the probability prediction data paid by each test user and the recommendation efficiency in the positioning time period;
determining a threshold value of a recommended user according to the first threshold value curve and the second threshold value curve;
and positioning the recommended users according to the determined threshold value of the recommended users.
According to the embodiment, video delivery to fewer people can be realized, the characteristic of improving efficiency is achieved, the prediction effect can be further tested through an actual delivery test, statistics is carried out on data every day, and further verification or adjustment can be carried out, and details are not repeated.
Please refer to fig. 2, which is a schematic diagram illustrating an embodiment of an apparatus for locating a recommended user according to the present invention, and mainly includes:
the collection module 1, in this embodiment, the collection module 1 is mainly used for collecting user transaction data and user video viewing data in a video system, and when the data is specifically implemented, the collection of the user transaction data and the user video viewing data in the video system can be implemented in various ways, and as a specific embodiment, referring to fig. 3, the collection module may include:
the first obtaining submodule 11 is used for obtaining a member transaction table log to obtain user transaction data of the video system;
and the second obtaining submodule 12 is used for obtaining the video watching log of the user to obtain the user transaction data of the video system.
As mentioned above, the member transaction list log and the user video viewing log comprise logs corresponding to a personal computer and a wireless terminal, so that the obtained data can be more accurately analyzed for the user viewing behavior.
An extracting module 2, in this embodiment, the extracting module 2 is mainly configured to extract a specified feature data set of a training user and a specified feature data set of a testing user from the collected user transaction data and user video viewing data, and when the specific implementation is implemented, the specified feature in this embodiment may be one or more of the following features:
movie channels, drama channels, car channels, fun channels, animation channels, entertainment channels, fashion channels, parent-child channels, game channels, creative channels, advertisement channels, music channels, information channels, sports channels, living channels, travel channels, science channels, education channels, entertainment channels, documentary channels, other channels, android devices, iphone devices, ipad devices, ipod devices, other devices, members, non-members, pay videos, free videos, complete viewing, and trial viewing.
In addition, it should be noted that, the extraction is mainly based on the user ID, and the classification is performed according to the specified features, and then the user behavior frequency data corresponding to each specified feature is collected, for example, it is determined that the user ID is the user 100, that is, the data of various viewing behaviors of the user 100, such as the frequency of viewing movie channels, the frequency of viewing tv channels, and the like, can be collected from the user video viewing log, and then the training user and the test user respectively form corresponding specified data sets.
In addition, in order to implement training, the specified feature data set of the training user may include positive sample data and negative sample data, the positive sample data may be a specified feature data set of a user who pays for purchase at a specified time point, the negative sample data is a specified feature data set of a user who does not pay for purchase at a specified time point, the number of the general negative sample data is greater than the number of the positive sample data, for example, the number of the negative sample data is three times or another ratio of the number of the positive sample data, which is not specifically limited herein, and in this embodiment, the specified time point is also not specifically limited herein, for example, the specified time point may be the current day, the user who pays for purchase may purchase a member, and the like, and in practice, the specified time point may also be other cases, which is only an example and is not specifically limited herein.
The training module 3, in this embodiment, the training module 3 is mainly configured to train each specified feature in the specified feature data set of the training user according to a training algorithm to obtain a weight value of each specified feature, as described above, various existing training algorithms may be used for training in the present application, which is not limited herein, and for example, the training algorithm may use an L2 regular logistic regression training algorithm.
A determining module 4, in this embodiment, the determining module 4 is mainly configured to determine probability prediction data paid by each test user according to the specified feature data set of the test user and the weight value of each specified feature obtained through training; in particular, for example, referring to fig. 4, as a specific embodiment, the determining module may include:
a sorting submodule 41 configured to sort the weight values of the obtained respective designated features;
a key specified characteristic determining submodule 42, configured to determine, according to the sorting result, a key specified characteristic within a specified sorting range;
and a probability prediction data determining submodule 43, configured to determine probability prediction data paid by each test user according to the weight of the key specified feature and the specified feature data set of the test user.
A positioning module 5, in this embodiment, the positioning module 5 is mainly used for positioning a recommended user according to probability prediction data paid by each test user, and when the specific implementation is implemented, for example, referring to fig. 5, the positioning module may include:
a first threshold curve determining submodule 51, configured to determine a first threshold curve of a recommended user according to the recommendation accuracy rate according to probability prediction data paid by each test user within a positioning time period;
a second threshold curve determining submodule 52, configured to determine a second threshold curve of the recommended user according to the recommendation efficiency according to the probability prediction data paid by each test user within the positioning time period;
a threshold determining submodule 53, configured to determine a threshold of the recommended user according to the first threshold curve and the second threshold curve;
and a recommended user positioning sub-module 54, configured to position the recommended user according to the determined threshold of the recommended user.
In the description provided above, numerous specific details are set forth. It is understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.
Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be interpreted as reflecting an intention that: that the invention as claimed requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims.
Claims (20)
1. A method for locating a recommending user, comprising:
collecting user transaction data and user video watching data in a video system;
extracting a specified characteristic data set of a training user and a specified characteristic data set of a testing user from the collected user transaction data and user video watching data;
training each designated feature in a designated feature data set of a training user according to a training algorithm to obtain a weight value of each designated feature;
determining probability prediction data paid by each test user according to the specified feature data set of the test user and the weight value of each specified feature obtained by training;
positioning the recommended users according to the probability prediction data paid by each test user;
the positioning of the recommended users according to the probability prediction data of the payment of each test user comprises the following steps:
determining a first threshold curve of a recommended user according to the probability prediction data paid by each test user and the recommendation accuracy within a positioning time period;
determining a second threshold curve of the recommended user according to the probability prediction data paid by each test user and the recommendation efficiency in the positioning time period;
determining a threshold value of a recommended user according to the first threshold value curve and the second threshold value curve;
and positioning the recommended users according to the determined threshold value of the recommended users.
2. The method of claim 1, wherein capturing video system user transaction data and user video viewing data comprises:
acquiring a log of a member transaction table to obtain user transaction data of a video system;
and acquiring a user video watching log to obtain user transaction data of the video system.
3. The method of claim 2, wherein the member transaction form log and the user video viewing log comprise pc and wireless end corresponding logs.
4. The method of claim 1, wherein determining probability prediction data for each test user to pay based on the specified feature data set of the test user and the trained weight values of each specified feature comprises:
sorting the weight values of the obtained specified characteristics;
determining key designated features in a designated sorting range according to a sorting result;
and determining probability prediction data paid by each test user according to the weight of the key specified features and the specified feature data set of the test user.
5. The method of claim 1, wherein the set of specified characteristics of the trained user comprises positive sample data for a set of specified characteristics of a user who has paid for purchases at a specified point in time and negative sample data for a set of specified characteristics of a user who has not paid for purchases at a specified point in time.
6. The method of claim 5, wherein the specified time points are the day of positive and negative sample collection.
7. The method of claim 5, wherein the number of negative sample data is three times the number of positive sample data.
8. The method of claim 5, wherein the user who pays for the purchase is a purchasing member user.
9. The method of any one of claims 1-8, wherein the training algorithm is an L2 regular logistic regression training algorithm.
10. The method according to any of claims 1-8, wherein the specified characteristics are one or more of the following:
movie channels, drama channels, car channels, fun channels, animation channels, entertainment channels, fashion channels, parent-child channels, game channels, creative channels, advertisement channels, music channels, information channels, sports channels, living channels, travel channels, science channels, education channels, entertainment channels, documentary channels, android devices, iphone devices, ipad devices, ipod devices, members, non-members, pay videos, free videos, complete viewing and viewing.
11. An apparatus for locating a recommending user, comprising:
the acquisition module is used for acquiring user transaction data and user video watching data in the video system;
the extraction module is used for extracting a specified characteristic data set of a training user and a specified characteristic data set of a testing user from the collected user transaction data and user video watching data;
the training module is used for training each specified feature in the specified feature data set of the training user according to a training algorithm to obtain a weight value of each specified feature;
the determining module is used for determining probability prediction data paid by each test user according to the specified feature data set of the test user and the weight value of each specified feature obtained by training;
the positioning module is used for positioning the recommended users according to the probability prediction data paid by each test user;
wherein the positioning module comprises:
the first threshold curve determining submodule is used for determining a first threshold curve of a recommended user according to the probability prediction data paid by each test user and the recommendation accuracy rate in the positioning time period;
the second threshold curve determining submodule determines a second threshold curve of the recommended user according to the probability prediction data paid by each test user and the recommendation efficiency in the positioning time period;
the threshold value determining submodule determines a threshold value of a recommended user according to the first threshold value curve and the second threshold value curve;
and the recommended user positioning submodule is used for positioning the recommended user according to the determined threshold value of the recommended user.
12. The apparatus of claim 11, wherein the acquisition module comprises:
the first acquisition submodule is used for acquiring a member transaction table log to obtain user transaction data of the video system;
and the second acquisition submodule is used for acquiring the video watching log of the user to obtain the user transaction data of the video system.
13. The apparatus of claim 12, wherein the member transaction form log and the user video viewing log comprise pc and wireless end corresponding logs.
14. The apparatus of claim 11, wherein the means for determining comprises:
the ordering submodule is used for ordering the weighted values of the obtained specified characteristics;
the key specified characteristic determining submodule is used for determining key specified characteristics in a specified sequencing range according to a sequencing result;
and the probability prediction data determining submodule is used for determining probability prediction data paid by each test user according to the weight of the key specified characteristics and the specified characteristic data set of the test user.
15. The apparatus of claim 11, wherein the set of specified characteristics of the trained user comprises positive sample data for a set of specified characteristics of a user who has paid for purchases at a specified point in time and negative sample data for a set of specified characteristics of a user who has not paid for purchases at a specified point in time.
16. The apparatus of claim 15, wherein the specified time points are the day of positive and negative sample collection.
17. The apparatus of claim 15, wherein the amount of negative sample data is three times the amount of positive sample data.
18. The apparatus of claim 15, wherein the user who pays for the purchase is a purchasing member user.
19. The apparatus of any one of claims 11-18, wherein the training algorithm is an L2 regular logistic regression training algorithm.
20. The apparatus according to any of claims 11-18, wherein the specified characteristics are one or more of the following:
movie channels, drama channels, car channels, fun channels, animation channels, entertainment channels, fashion channels, parent-child channels, game channels, creative channels, advertisement channels, music channels, information channels, sports channels, living channels, travel channels, science channels, education channels, entertainment channels, documentary channels, android devices, iphone devices, ipad devices, ipod devices, members, non-members, pay videos, free videos, complete viewing and viewing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510604634.2A CN105160008B (en) | 2015-09-21 | 2015-09-21 | Method and device for positioning recommended user |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510604634.2A CN105160008B (en) | 2015-09-21 | 2015-09-21 | Method and device for positioning recommended user |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105160008A CN105160008A (en) | 2015-12-16 |
| CN105160008B true CN105160008B (en) | 2020-03-31 |
Family
ID=54800864
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510604634.2A Active CN105160008B (en) | 2015-09-21 | 2015-09-21 | Method and device for positioning recommended user |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105160008B (en) |
Families Citing this family (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105719164A (en) * | 2016-01-21 | 2016-06-29 | 海信集团有限公司 | Paid multimedia resource recommending method and paid multimedia resource recommending device |
| CN106204106A (en) * | 2016-06-28 | 2016-12-07 | 武汉斗鱼网络科技有限公司 | A kind of specific user's recognition methods and system |
| CN106168980B (en) * | 2016-07-26 | 2020-07-28 | 阿里巴巴（中国）有限公司 | Multimedia resource recommendation sequencing method and device |
| CN108509459B (en) * | 2017-02-28 | 2022-05-20 | 腾讯科技（北京）有限公司 | Data processing method and device |
| CN107016569A (en) * | 2017-03-21 | 2017-08-04 | 聚好看科技股份有限公司 | The targeted customer's account acquisition methods and device of a kind of networking products |
| CN107679920A (en) * | 2017-10-20 | 2018-02-09 | 北京奇艺世纪科技有限公司 | The put-on method and device of a kind of advertisement |
| CN110033294A (en) * | 2018-01-12 | 2019-07-19 | 腾讯科技（深圳）有限公司 | A kind of determination method of business score value, business score value determining device and medium |
| CN108416649A (en) * | 2018-02-05 | 2018-08-17 | 北京三快在线科技有限公司 | Search result ordering method, device, electronic equipment and storage medium |
| CN108683949B (en) * | 2018-05-18 | 2021-11-02 | 北京奇艺世纪科技有限公司 | Method and device for extracting potential users of live broadcast platform |
| CN110555747A (en) * | 2018-06-01 | 2019-12-10 | 北京京东尚科信息技术有限公司 | method and device for determining target user |
| CN109121006B (en) * | 2018-08-02 | 2021-02-02 | 中国联合网络通信集团有限公司 | Marketing method and platform based on live broadcast watching user |
| CN108965938B (en) * | 2018-08-03 | 2020-03-20 | 山东大学 | Method and system for predicting potential pay users in smart television |
| CN111382307B (en) * | 2018-12-27 | 2023-10-17 | 深圳Tcl新技术有限公司 | Video recommendation method, system and storage medium based on deep neural network |
| CN111309939B (en) * | 2020-02-03 | 2023-02-28 | 天津智融创新科技发展有限公司 | Video recommendation sorting method and device |
| CN112131479A (en) * | 2020-09-30 | 2020-12-25 | 深圳前海微众银行股份有限公司 | Data processing method, device, equipment and storage medium |
| CN114513700A (en) * | 2020-11-16 | 2022-05-17 | 中国移动通信集团有限公司 | Method, device and equipment for recommending video and storage medium |
| CN112866760B (en) * | 2021-01-18 | 2023-08-22 | 青岛聚看云科技有限公司 | Content display method, display equipment and server |
| CN113139856B (en) * | 2021-05-14 | 2023-04-07 | 青岛聚看云科技有限公司 | Movie and television member package recommendation method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104598521A (en) * | 2014-12-12 | 2015-05-06 | 北京京东尚科信息技术有限公司 | Method and device for processing user behavior data |
| CN104778173A (en) * | 2014-01-10 | 2015-07-15 | 腾讯科技（深圳）有限公司 | Determination method, device and equipment of objective user |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080201201A1 (en) * | 2006-09-25 | 2008-08-21 | Sms.Ac | Methods and systems for finding, tagging, rating and suggesting content provided by networked application pods |
| US20080282305A1 (en) * | 2007-04-21 | 2008-11-13 | Carpe Media | Media Distribution System, Apparatus, Method and Software |
| CN104090888B (en) * | 2013-12-10 | 2016-05-11 | 深圳市腾讯计算机系统有限公司 | A kind of analytical method of user behavior data and device |
-
2015
- 2015-09-21 CN CN201510604634.2A patent/CN105160008B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104778173A (en) * | 2014-01-10 | 2015-07-15 | 腾讯科技（深圳）有限公司 | Determination method, device and equipment of objective user |
| CN104598521A (en) * | 2014-12-12 | 2015-05-06 | 北京京东尚科信息技术有限公司 | Method and device for processing user behavior data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105160008A (en) | 2015-12-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105160008B (en) | Method and device for positioning recommended user | |
| CN103763585B (en) | A kind of user's characteristic information acquisition methods, device and terminal device | |
| CN107105318B (en) | Video hotspot segment extraction method, user equipment and server | |
| CN103559206B (en) | A kind of information recommendation method and system | |
| CN106326391B (en) | Multimedia resource recommendation method and device | |
| CN107483982B (en) | Anchor recommendation method and device | |
| JP6502506B2 (en) | TV viewer measurement method and apparatus | |
| CN110475155B (en) | Live video hot state identification method, device, equipment and readable medium | |
| CN106604051A (en) | Live channel recommending method and device | |
| US20170134181A1 (en) | Method and device for data push | |
| CN107894998B (en) | Video recommendation method and device | |
| US20170169040A1 (en) | Method and electronic device for recommending video | |
| CN103984741B (en) | Customer attribute information extracting method and system thereof | |
| CN104506894A (en) | Method and device for evaluating multi-media resources | |
| CN109429103B (en) | Method and device for recommending information, computer readable storage medium and terminal equipment | |
| CN104967690B (en) | A kind of information-pushing method and device | |
| US20110167445A1 (en) | Audiovisual content channelization system | |
| CN103442270B (en) | A kind of method and device for the viewing-data for gathering user | |
| CN110366043A (en) | Method for processing video frequency, device, electronic equipment and readable medium | |
| CN109493146A (en) | Advertisement placement method and device | |
| CN103942275A (en) | Video identification method and device | |
| CN105184611A (en) | Advertising effect quantification method and display system | |
| CN105843876B (en) | Quality evaluation method and device for multimedia resources | |
| CN105956061B (en) | Method and device for determining similarity between users | |
| CN106056406A (en) | Method and device for generating program key word map |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 100080 Beijing Haidian District city Haidian street A Sinosteel International Plaza No. 8 block 5 layer A, C Patentee after: Youku network technology (Beijing) Co.,Ltd. Address before: 100080, A, 5, block 8, Sinosteel International Plaza, Haidian Avenue, Haidian District, Beijing Patentee before: 1VERGE INTERNET TECHNOLOGY (BEIJING) Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200619 Address after: 310052 room 508, floor 5, building 4, No. 699, Wangshang Road, Changhe street, Binjiang District, Hangzhou City, Zhejiang Province Patentee after: Alibaba (China) Co.,Ltd. Address before: 100080 Beijing Haidian District city Haidian street A Sinosteel International Plaza No. 8 block 5 layer A, C Patentee before: Youku network technology (Beijing) Co.,Ltd. |