CN104462270A - Information recommendation method and device - Google Patents
Information recommendation method and device Download PDFInfo
- Publication number
- CN104462270A CN104462270A CN201410684996.2A CN201410684996A CN104462270A CN 104462270 A CN104462270 A CN 104462270A CN 201410684996 A CN201410684996 A CN 201410684996A CN 104462270 A CN104462270 A CN 104462270A
- Authority
- CN
- China
- Prior art keywords
- time
- recommendation list
- user equipment
- recommendation
- prediction result
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000004364 calculation method Methods 0.000 claims description 12
- 101100264195 Caenorhabditis elegans app-1 gene Proteins 0.000 description 11
- 238000007477 logistic regression Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 6
- 230000006399 behavior Effects 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/954—Navigation, e.g. using categorised browsing
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention provides an information recommendation method and device and relates to the technical field of Internet. By means of the information recommendation method and device, user equipment using a recommendation list can be predicted, the recommendation list of the predicted user equipment using the recommendation list can be updated, the number of users needing to update the recommendation list is reduced, computing resources are saved, and the recommendation list can be updated in time for the user equipment which will use the recommendation list. The method comprises the steps that a first prediction result is obtained at first time, wherein the first prediction result comprises the predicted user equipment which will use the recommendation list at second time and third time, the first time is earlier than the second time, and the second time is earlier than the third time; according to the first prediction result, the recommendation list is generated for the predicted user equipment which will use the recommendation list at the second time and the third time.
Description
Technical Field
The invention relates to the technical field of internet, in particular to an information recommendation method and device.
Background
The rapid development of the internet technology meets the requirements of users on information, but with the great increase of the information, the users can hardly obtain useful information from the information quickly. The recommendation system solves the problem, and can recommend information interested by the user to the user according to the requirement, the interest and the like of the user.
Conventional recommendation systems generate recommendation lists for users based on user characteristics, information characteristics, and the like. The conventional recommendation system generally generates a recommendation list for a user in a time period when the user uses the recommendation list less, that is, when a server is idle, large data is calculated, and the recommendation list is generated for all users using the recommendation list. In the prior art, a recommendation list is generated for a historically active user when the recommendation list is generated for the user. The historically active users are defined by adopting a statistical method, such as defining users who use the recommendation list for a preset number of times in the past month as the historically active users.
In the prior art, recommendation lists are generated for all history active users, but actually, not all history active users use the recommendation lists, so that a part of meaningless calculation processes of the recommendation lists exist in the recommendation system in the prior art, which causes resource waste, and the latest information cannot be fed back to the users in time because the history active users in the prior art have more users and the period for generating the recommendation lists is longer.
Disclosure of Invention
The embodiment of the invention provides an information recommendation method and device, which can predict user equipment using a recommendation list, update the recommendation list aiming at the predicted user equipment using the recommendation list, reduce the number of users needing to update the recommendation list, save computing resources and update the recommendation list for the user equipment using the recommendation list in time.
In order to achieve the above purpose, the embodiment of the invention adopts the following technical scheme:
in a first aspect, an embodiment of the present invention provides an information recommendation method, where the method includes: obtaining a first prediction result at a first time, wherein the first prediction result comprises a predicted user equipment which will use a recommendation list from a second time to a third time, the first time is earlier than the second time, and the second time is earlier than the third time; generating the recommendation list for the predicted user equipment that will use the recommendation list from the second time to the third time according to the first prediction result.
With reference to the first aspect, in a first possible implementation manner of the first aspect, before the obtaining the first prediction result at the first time, the method further includes: and acquiring a second prediction result, wherein the second prediction result comprises predicted user equipment which will use the recommendation list in a preset time period and use time, the preset time period is a period for acquiring the second prediction result, the use time is the time for the user equipment which will use the recommendation list to use the recommendation list in the preset time period, and the first time, the second time and the third time are in the preset time period.
With reference to the first possible implementation manner of the first aspect, in a second possible implementation manner of the first aspect, the obtaining a first prediction result at the first time specifically includes: and reading the user equipment which uses the recommendation list in the second time to the third time in the use time at the first time.
With reference to the first possible implementation manner of the first aspect, in a third possible implementation manner of the first aspect, the obtaining a second prediction result specifically includes: obtaining a prediction model, wherein the prediction model is used for calculating the probability that each user equipment will use a recommendation list within a preset time period; respectively bringing the parameters of each user equipment into the prediction model, and calculating the probability that each user equipment will use a recommendation list in a preset time period; and if the probability that the first user equipment will use the recommendation list in a preset time period is greater than the preset probability, setting the first user equipment as the user equipment which will use the recommendation list in the preset time period, wherein the first user equipment is any one of the user equipment.
With reference to the third possible implementation manner of the first aspect, in a fourth possible implementation manner of the first aspect, the prediction model is y (x) -1/1 + e- (β)0+β1x1+...+βnxn) (ii) a Wherein y (x) is a probability of using a recommendation list for the first user device; beta is a0、β1…βnParameters of the prediction model; x is the number of1…xnIs a parameter of the first user equipment.
With reference to the first aspect, in a fifth possible implementation manner of the first aspect, after the generating the recommendation list for the user equipment according to the first prediction result, the method further includes: sending, prior to or at the second time, the recommendation list to a user device predicted to use the recommendation list within the second time to the third time; or receiving a recommendation list request message sent by the user equipment which is predicted to use the recommendation list from the second time to the third time within a second time to a third time, and sending the recommendation list to the user equipment which is predicted to use the recommendation list from the second time to the third time.
In a second aspect, an embodiment of the present invention provides an apparatus for information recommendation, where the apparatus includes: a first obtaining module, configured to obtain a first prediction result at a first time, where the first prediction result includes a predicted user equipment that will use a recommendation list from a second time to a third time, the first time being earlier than the second time, and the second time being earlier than the third time; a generating module, configured to generate the recommendation list for the predicted user equipment that will use the recommendation list from the second time to the third time according to the first prediction result.
With reference to the second aspect, in a first possible implementation manner of the second aspect, the apparatus further includes: the second obtaining module is configured to obtain a second prediction result before the first obtaining module obtains the first prediction result at a first time, where the second prediction result includes a predicted user equipment that will use the recommendation list within a preset time period and a predicted usage time, the preset time period is a period for obtaining the second prediction result, the predicted usage time is a time for the user equipment that will use the recommendation list to use the recommendation list within the preset time period, and the first time, the second time, and the third time are within the preset time period.
With reference to the first possible implementation manner of the second aspect, in a second possible implementation manner of the second aspect, the first obtaining module is specifically configured to: and reading the user equipment which uses the recommendation list in the second time to the third time in the use time at the first time.
With reference to the first possible implementation manner of the second aspect, in a third possible implementation manner of the second aspect, the second obtaining module specifically includes: the obtaining sub-module is used for obtaining a prediction model, wherein the prediction model is used for calculating the probability that each user equipment will use the recommendation list within a preset time period; the calculation submodule is used for respectively bringing the parameters of each user equipment into the prediction model and calculating the probability that each user equipment will use the recommendation list in a preset time period; the setting sub-module is configured to set the first user equipment as the user equipment that will use the recommendation list within a preset time period if the probability that the first user equipment will use the recommendation list within the preset time period is greater than a preset probability, where the first user equipment is any one of the user equipments.
With reference to the third possible implementation manner of the second aspect, in a fourth possible implementation manner of the second aspect, the prediction model is y (x) -1/1 + e- (β)0+β1x1+...+βnxn) (ii) a Wherein y (x) is a probability of using a recommendation list for the first user device; beta is a0、β1…βnParameters of the prediction model; x is the number of1…xnIs a parameter of the first user equipment.
With reference to the second aspect, in a fifth possible implementation manner of the second aspect, the apparatus further includes: a receiving module, configured to receive a recommendation list request message sent by a user equipment that is predicted to use the recommendation list from the second time to the third time; a sending module, configured to send, after generating the recommendation list for the user equipment according to the first prediction result, the recommendation list to a user equipment predicted to use the recommendation list within the second time to the third time before the second time or at the second time; or, in a second time to a third time, after receiving a recommendation list request message sent by a user device predicted to use the recommendation list in the second time to the third time, the receiving module sends the recommendation list to the user device predicted to use the recommendation list in the second time to the third time.
The embodiment of the invention provides an information recommendation method and device, wherein the method comprises the following steps: obtaining a first prediction result at a first time, wherein the first prediction result comprises predicted user equipment which will use the recommendation list from a second time to a third time, the first time is earlier than the second time, and the second time is earlier than the third time; and generating a recommendation list for the predicted user equipment which will use the recommendation list in the second time to the third time according to the first prediction result.
Based on the description of the above embodiment, the present invention can predict the user devices using the recommendation list and update the recommendation list for the user devices using the recommendation list by acquiring the user devices that will use the recommendation list from the second time to the third time at the first time and generating the recommendation list for the user devices. According to the technical scheme, the recommendation list does not need to be updated for all the history active users, so that the number of users needing to update the recommendation list is reduced, computing resources are saved, and the recommendation list can be updated for the user equipment which will use the recommendation list in time.
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 of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a first flowchart illustrating an information recommendation method according to an embodiment of the present invention;
fig. 2 is a flowchart illustrating a second method for information recommendation according to an embodiment of the present invention;
fig. 3 is a third schematic flowchart of a method for information recommendation according to an embodiment of the present invention;
fig. 4 is a first schematic structural diagram of an information recommendation apparatus according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an information recommendation apparatus according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram three of an information recommendation apparatus according to an embodiment of the present invention.
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, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example one
An embodiment of the present invention provides an information recommendation method, and as shown in fig. 1, a flow diagram of the method is provided, where the method includes:
s101, the information recommendation device obtains a first prediction result at a first time.
Wherein the first prediction result comprises a predicted user equipment that will use the recommendation list from a second time to a third time, the first time being earlier than the second time, the second time being earlier than the third time.
Specifically, the first time is a time when the user equipment which will use the recommendation list starts to calculate the recommendation list from the second time to the third time, and since a certain period of time is required for calculating the recommendation list, the first time is earlier than the second time to ensure that the calculation process of the recommendation list is completed before the second time.
S102, the information recommending device generates a recommendation list for the predicted user equipment which will use the recommendation list in the second time to the third time according to the first prediction result.
Specifically, the information recommendation device generates a recommendation list for the user equipment according to the acquired user equipment which will use the recommendation list in the second time to the third time. The method is equivalent to that the information recommendation device only carries out calculation for updating the recommendation list on the predicted user equipment which will use the recommendation list, so that the calculation amount of the information recommendation device can be greatly reduced, the calculation time is reduced, and the calculation resources are saved.
Illustratively, the user equipment that will use the recommendation list between 13:00 of 2014-11-18 (i.e., the second time in the present scheme) and 14:00 of 2014-11-18 (i.e., the third time in the present scheme) performs the update of the recommendation list, the first time may be any time before 13:00 of 2014-11-18, optionally, 12:00 of 2014-11-18 is set, and the information recommendation device generates the recommendation list using the user equipment that will use the recommendation list within 13:00 of 2014-11-18 and 14:00 of 2014-11-18 at 12:00 of 2014-11-18 and 12:00 of 2014-11-18.
It should be noted that the setting of the first time, the second time, and the third time may be preset, and the present invention is not limited thereto.
Specifically, the second time and the third time may be set to a fixed interval, such as one hour, two hours, or other time duration; in another embodiment, the time between the second time and the third time may be a variable time, and at this time, the specific values of the second time and the third time may be determined according to the historical access amount of the user equipment.
For example, the update period of the recommendation list may be determined according to the frequency of using the recommendation list by the user equipment. For example, generally, when the frequency of using the recommendation list by the user equipment is higher at 21:00 to 23:00 evening, the period for updating the recommendation list for the user equipment can be set to be shorter correspondingly; and in the early morning, the frequency of using the recommendation list by the user equipment is low, and the period for updating the recommendation list for the user equipment can be set to be longer correspondingly, so that the computing resources required by the information recommendation device for generating the recommendation list for the user equipment can be reduced.
It is added that the process of recommendation list generation is completed before the second time.
The embodiment of the invention provides an information recommendation method, which comprises the steps of obtaining a first prediction result at a first time, wherein the first prediction result comprises predicted user equipment which will use a recommendation list from a second time to a third time, the first time is earlier than the second time, and the second time is earlier than the third time; and generating a recommendation list for the predicted user equipment which will use the recommendation list in the second time to the third time according to the first prediction result.
Based on the description of the above embodiment, the present invention can predict the user devices using the recommendation list and update the recommendation list for the user devices using the recommendation list by acquiring the user devices that will use the recommendation list from the second time to the third time at the first time and generating the recommendation list for the user devices. According to the technical scheme, the recommendation list does not need to be updated for all the history active users, so that the number of users needing to update the recommendation list is reduced, computing resources are saved, and the recommendation list can be updated for the user equipment which will use the recommendation list in time.
Example two
An embodiment of the present invention provides an information recommendation method, and as shown in fig. 2, a flow diagram of the method is provided, where the method includes:
s201, obtaining a second prediction result.
The second prediction result comprises predicted user equipment which will use the recommendation list within a preset time period and use time, the preset time period is a period for obtaining the second prediction result, the use time is the time for the user equipment which will use the recommendation list to use the recommendation list within the preset time period, and the first time, the second time and the third time are within the preset time period.
Specifically, the preset time period includes a first time, a second time, and a third time. The second prediction result comprises predicted user devices using the recommendation list at respective times within a preset time period, i.e. the second prediction result comprises detailed information of which predicted user devices will use the recommendation list at which times.
The preset time period is a period for obtaining the second prediction result, and optionally, the preset time period may be set to one day (i.e. 24 hours), that is, the program for calculating the second prediction result may be operated at any fixed time of 24 hours per day. Preferably, the program for calculating the second prediction result may be run once every morning at 3:00, that is, a time with a small access amount of the user equipment is selected to perform a large amount of calculations, so that resources can be fully utilized.
The specific S201 includes S201a-S201 c.
S201a, obtaining a prediction model.
The prediction model is used for calculating the probability that each user equipment will use the recommendation list within a preset time period.
Preferably, the prediction model is a logistic regression model y (x) 1/1+ e- (. beta.) (β)0+β1x1+...+βnxn) (ii) a Wherein y (x) is the probability of the first user device using the recommendation list; beta is a0、β1…βnParameters of the prediction model; x is the number of1…xnIs a parameter of the first user equipment.
It should be added that the prediction model y (x) 1/1+ e- (β)0+β1x1+...+βnxn) For the logistic regression model, the technical solution of the present invention may also use other prediction models to calculate the probability that each user equipment will use the recommendation list within a preset time period, such as a matrix model.
Exemplary parameters of the user equipment may include: user attributes such as user identification, gender, age, user tags, etc.; device attributes such as device identification, device model; application attributes such as application identification, application tag, application category, etc.; contextual attributes, such as user location information, etc.; time attributes such as a time period to which the information to be recommended belongs, whether the current time period belongs to a holiday or a workday, a week attribute (monday, tuesday), and the like; the information attribute to be recommended, such as an information identifier, a category to which the information belongs, an information tag and the like; the parameters of the user equipment also include the combination of user attributes, equipment attributes and attributes of information to be recommended, such as generating a combined attribute 'gender-information identifier' from the gender and the information identifier of the user.
Further, the prediction model is used for calculating the probability that each user equipment will use the recommendation list within a preset time period, that is, calculating the probability that the historically active user equipment marked in the database will use the recommendation list within the preset time period. The selection of the history active user may select the history active user in the last 3 months, or may select the history active user in the last one month, and specifically, the selection rule of the history active user may be preset, which is not limited by the present invention.
Illustratively, the logistic regression model y (x) 1/1+ e- (beta)0+β1x1+...+βnxn) For example, the prediction model is obtained as follows:
the logistic regression model is: y (x) 1/1+ e- (. beta.)0+β1x1+...+βnxn) Substituting the training data into the logistic regression model to obtain each parameter beta of the logistic regression model0、β1…βn。
It should be added that the training data is data for obtaining various parameters in the prediction model. For example, the embodiment of the present invention uses a logistic regression model to obtain the second prediction result, and the parameters in the logistic regression model are trained by the training data.
It should be noted that the training data is obtained from a result of prediction of a user device using the recommendation list in a past certain period of time and a result of actual use of the user device using the recommendation list in the past certain period of time. Here, the past certain time is taken as the time of two days in the past (between 1:00 in the morning of 2014-11-16 and 1:00 in the morning of 2014-11-18) as an example for explanation.
Specifically, the selection of the past certain period of time may be specifically set according to actual conditions.
For example, in the past two days (between 1:00 in the early morning of 2014-11-16 and 1:00 in the early morning of 2014-11-18), the prediction results of the information recommendation device on the user devices that will use the recommendation list in the two days are shown in Table 1:
TABLE 1
| userid1，app1,6，sunday |
| userid1，app1,18，sunday |
| userid2，app1,20，sunday |
| userid1，app1,6，monday |
| userid1，app1,18，monday |
| userid2，app1,21，monday |
| …… |
Specifically, the meanings of the predicted results of the user devices that will use the recommendation list in table 1 from 1:00 am 2014-11-16 to 1:00 am 2014-11-18 are illustrated by the first row of data in table 1, which is userid1, app1,6, sunday, where userid1 is the user identifier, app1 is the application identifier, and 6 and sunday are the time attributes of the user devices accessing the application. This line of data indicates that the user device userid1 will access the application app1 at 6 points on the sunday.
Illustratively, the results of the user device actually using the recommendation list between 1:00 in the early morning of 2014-11-16 and 1:00 in the early morning of 2014-11-18 are shown in Table 2:
TABLE 2
| userid1，app1，6，sunday |
| userid1，app1，18，sunday |
| userid2，app1，20，sunday |
| userid1，app1，18，monday |
| userid2，app1，21，monday |
| …… |
Specifically, in the last two days, i.e. in the early morning 1:00 of 2014-11-16 to the early morning 1:00 of 2014-11-18, the results of the recommendation list actually used by the user equipment are shown in table 2, the meaning of the data in table 2 is illustrated by taking the first row of data as an example, and the first row of data is: userid1, app1,6, sunday, where userid1 is the user identification, app1 is the application identification, and 6 and sunday are the time attributes for the user device to use the application. This line of data represents that the user device userid1 accessed the application app1 at 6 o' clock on a sunday.
Illustratively, training data is obtained based on the predicted results of user devices that will use the recommendation list between morning 1:00 at 2014-11-16 and morning 1:00 at 2014-11-18 (as shown in Table 1) and user devices that actually used the recommendation list between morning 1:00 at 2014-11-16 and morning 1:00 at 2014-11-18 (as shown in Table 2), as shown in Table 3:
TABLE 3
| user^userid1，app^app1,time^6，week^sunday，1 |
| user^userid1，app^app1,time^18，week^sunday，1 |
| user^userid2，app^app1,time^20，week^sunday，1 |
| user^userid1，app^app1,time^6，week^monday，0 |
| user^userid1，app^app1,time^18，week^monday，1 |
| user^userid2，app^app1,time^21，week^monday，1 |
| …… |
Specifically, each row of data in table 3 represents a piece of training data, and taking the first piece of training data, i.e., user ^ userid1, app ^ app1, time ^6, week ^ sunday, 1 as an example, combining table 1 and table 2, it can be known that the information recommendation apparatus predicts that the user equipment userid1 will visit the user equipment in the time period corresponding to 6 points of sundaysApplication app1, in fact, user device userid1 did access application app1 for a time period corresponding to 6 o' clock on sunday, and therefore, this piece of data is recorded as a positive example in the training data, labeled 1; conversely, the fourth piece of training data user ^ userid1, app ^ app1, time ^6, week ^ monday, 0 in table 3 indicates that the information recommendation apparatus predicts that the user device userid1 will access the application app1 in the time period corresponding to 6 points on Saturday, while the user device userid1 does not actually access the application app1 in the time period corresponding to 6 points on Saturday, and therefore, the piece of data is recorded as a negative example in the piece of training data, which is marked as 0. Training data is generated according to such a rule. Calculating a parameter beta in the prediction model according to the training data0、β1…βn。
It should be added that the attribute of the application is added to the training data in order to distinguish the probability that the user equipment uses the recommendation list by accessing different applications, so that the information recommendation apparatus generates different recommendation lists when the user equipment accesses different applications.
Optionally, the attribute of the application may not be added to the training data, so that the information recommendation device generates the same recommendation list when the user equipment accesses different applications.
It should be noted that for a user device that has not used a recommendation list before, there is no behavior attribute, but the unlabeled test data may be constructed according to other attributes in the parameters of the user device.
In actual use, the training data is fused with various data, including user attributes; device attributes; an application attribute; an information attribute; a context attribute; a time attribute; user behavior attributes, such as whether the user clicks on information displayed to him, etc.; combinations of user attributes, device attributes, application attributes, information attributes, context attributes, time attributes, user behavior attributes are also included.
S201b, respectively bringing the parameters of each user device into a prediction model, and calculating the probability that each user device will use the recommendation list within a preset time period.
The specific role of the prediction model here is to predict the probability that a certain user device will use the recommendation list at a certain time. Firstly, parameters of a prediction model are calculated according to training data, wherein the prediction model is y (x) 1/1+ e- (beta)0+β1x1+...+βnxn) (ii) a Wherein y (x) is the probability of the user equipment using the recommendation list, β0、β1…βnParameters of the prediction model; x is the number of1…xnIs a parameter of the first user equipment.
Further, the parameters of the user equipment in the training data are used as the values of each x in the prediction model.
Specifically, the method for generating parameters of a prediction model according to training data mainly includes offline batch generation and online real-time generation, wherein the offline batch generation refers to calculating the parameters of the prediction model after all the training data are obtained, and the algorithms mainly include an SGD (random gradient Descent) algorithm, an LBFGS (Limited-memoryBroyden-Fletcher-Goldfarb-Shanno, finite-field quasi-newton) algorithm; the online real-time generation means that a new piece of training data is generated, and the parameters of the prediction model are updated according to the training data, so that the prediction model can be updated in real time.
Specifically, label-free test data are constructed according to parameters of each user equipment, the label-free test data are brought into a prediction model, and the probability that each user equipment will use a recommendation list within a preset time period is calculated. The format of the unlabeled test data is shown in table 4:
TABLE 4
| userid1，app1，6，tuesday |
| userid1，app1，18，tuesday |
| …… |
Illustratively, according to the unlabeled test data in the table and the calculated prediction model, the probability that the user equipment userid1 uses the recommendation list at the time period corresponding to each time point is predicted to obtain a second prediction result. The first row of unlabeled data in table 4 is taken into the predictive model, resulting in a probability that the user device uses the recommendation list through application app1 for a time period corresponding to 6 points on tuesdays.
According to the method, the probability that the recommendation list is used by accessing different applications is calculated for the time period corresponding to each time point in the preset time period of each user equipment.
S201c, if the probability that the first user equipment will use the recommendation list within the preset time period is greater than the preset probability, setting the first user equipment as the user equipment that will use the recommendation list within the preset time period.
The first user equipment is any one of the user equipments.
Optionally, the user equipment that uses the recommendation list at a certain time with a probability greater than a preset probability is used as the user equipment that will use the recommendation list at the certain time. The predetermined probability may be preset, and the present invention is not limited thereto.
S202, at the first time, a first prediction result is obtained.
Wherein the first prediction result comprises a predicted user equipment that will use the recommendation list from a second time to a third time, the first time being earlier than the second time, the second time being earlier than the third time.
Specifically, at the first time, the user equipment which uses the recommendation list within the second time to the third time is read from the second prediction result.
And S203, generating a recommendation list for the predicted user equipment which will use the recommendation list in the second time to the third time according to the first prediction result.
Further, after the information recommendation device obtains the first prediction result, a recommendation list is generated for the user equipment in the first prediction result.
Specifically, the process of the information recommendation device generating the recommendation list for the user equipment which will use the recommendation list in the second time to the third time at the first time is as follows:
(1) parameters of the user equipment which will use the recommendation list in the second time to the third time are obtained, and the non-label test data is formed.
For example, the unlabeled test data configured according to the attribute of the user equipment and the attribute of the information may be: userid1, 20, itemid1, webgame; the meaning of the unlabeled test data is: the userid1 is the user identification, 20 is the age of the user, itemid1 is the identification of the information, and the online tour is the category of the information.
(2) And generating a recommendation list for the user equipment which will use the recommendation list in the second time to the third time according to the recommendation model and the unlabeled test data.
Illustratively, user equipment parameters for unlabeled test data are: userid1, 20, itemid1, webgame; the onboarding recommendation model calculates a probability that the user equipment uses the information within the second time to the third time.
And calculating the probability of using each piece of information by the user equipment according to the mode, and recommending the top n pieces of information with higher probability to the user equipment.
Then, the probability that each user device uses each piece of information is calculated for the user devices which will use the recommendation list in the second time to the third time, and the first n pieces of information with larger usage probability of each user device are recommended to the corresponding user devices. Wherein the size of n can be preset.
It should be noted that, in order to ensure the real-time performance of the information recommendation device for updating the recommendation list for the user equipment, the information recommendation device may set the period for updating the recommendation list to be shorter, for example, set the period for updating the recommendation list to be 1 hour, and when the user access amount is large, the period may be even shorter. According to the technical scheme, the recommendation list only needs to be updated for the predicted user equipment which will use the recommendation list in the second time to the third time, so that the short updating time of the recommendation list can be completely realized, and the real-time performance of updating and recommending the equipment can be ensured.
S204, the user equipment which uses the recommendation list in the second time to the third time obtains the updated recommendation list.
Specifically, S204 may be S204a or S204 b.
S204a, before or at the second time, sending the recommendation list to the user device predicted to use the recommendation list from the second time to the third time.
S204b, receiving the recommendation list request message from the user equipment predicted to use the recommendation list from the second time to the third time, and sending the recommendation list to the user equipment predicted to use the recommendation list from the second time to the third time.
It should be noted that the information recommendation apparatus will use the user equipment of the recommendation list to generate the recommendation list from the second time to the third time at the first time, and after the generation of the recommendation list is completed, the user equipment obtains the updated recommendation list in two ways, as shown in S204a and S204 b.
Specifically, S204a is that before or at the second time, the information recommendation apparatus sends the recommendation list to the user equipment that is predicted to use the recommendation list in the second time to the third time.
More specifically, in step S205b, the information recommendation apparatus receives the recommendation list request message from the user equipment that is predicted to use the recommendation list in the second time to the third time, and sends the recommendation list to the user equipment that is predicted to use the recommendation list in the second time to the third time.
The following describes the implementation of the technical solution of the present invention in a specific application scenario. The 24 hours of any day are marked as 0-24, assuming that the preset time period is 24 hours, the starting time of the preset time period is 3 points, the first time is 12 points, the second time is 13 points, the third time is 14 points, and the update cycle of the recommendation list is set to be a fixed cycle of 1 hour. At 3 points a prediction is made of which users will use the recommendation list at what time for the next 24 hours. The user who will use the recommendation list will make an update of the recommendation list from the second time 13 to the third time 14 in the future at the first time 12.
The first step is as follows: at point 3, the user devices that will use the recommendation list and the time at which each user device will use the recommendation list during the time period from the current point 3 to the next point 3 are calculated.
Specifically, the parameter β of the prediction model is calculated according to the training data, so that the probability that the user equipment in the prediction database uses the recommendation list at a certain time is predicted according to the obtained prediction model, and the user equipment with the probability greater than the preset probability is used as the user equipment that will use the recommendation list at a certain time.
The second step is that: and acquiring user equipment which uses the recommendation list within the second time to the third time at the first time and generating the recommendation list for the user equipment.
Specifically, the specific process of generating the recommendation list for the user equipment that will use the recommendation list in the second time to the third time is as follows:
and forming label-free test data according to the parameters of the user equipment, calculating the probability of using information of the user equipment according to the recommendation model and the label-free test data, and recommending the first n pieces of information with higher probability to the user.
The parameters of the above-described recommended model can be calculated by training data as shown in table 5.
TABLE 5
| userid1, 20, itemid1, webgame, 1 |
| userid1, 20, itemid2, health care, 0 |
| userid2, 60, itemid1, webgame, 1 |
| userid2, 60, itemid2, health care, 1 |
| …… |
The meaning of the data in table 5 is illustrated by taking the first row of data as an example, the first row of data is userid1, 20, itemid1, and webgame 1, where userid1 represents a user identifier, 20 represents a user age, itemid1 represents an information identifier, and the webgame is a category of information, and 1 represents a behavior characteristic of the user, specifically represents information recommended by the user, and the piece of data constitutes a positive example in the training data. Similarly, the second row of data is userid1, 20, itemid2, health care, 0, wherein userid1, 20, itemid2, health care, and the meaning of these data is the same as that of the first row of data, and 0 in this piece of data indicates that the user recommended information, but the user did not browse, this piece of data constitutes a negative example in the training data.
The parameters of the recommendation model are calculated based on these training data, and the recommendation model here may use a logistic regression model y (x) ═ 1/1+ e — (β)0+β1x1+...+βnxn) Where x represents a parameter of the user device and y (x) represents a probability that the user uses the recommendation list. In order to ensure real-time performance, the incremental learning optimization method can be used for training each parameter of the recommendation model, that is, the new training data is used to update the parameter of the old recommendation model, and the FTRL online learning algorithm as described above belongs to the incremental learning optimization method.
And then constructing unlabeled test data according to the parameters of the user equipment which will use the recommendation list within the second time to the third time, calculating the probability of using certain information by the user equipment according to the calculated recommendation model and the unlabeled test data, and recommending n pieces of information with higher probability of using the certain information by the user equipment to the user equipment.
It should be noted that the recommendation model for generating the recommendation list for the user equipment may also adopt other models besides the above logistic regression model, such as a matrix model.
Further, when the recommendation model used when generating the recommendation list for the user equipment is a logistic regression model in a linear model, the calculation time for predicting the probability that the user equipment uses a certain piece of information according to the recommendation model may be estimated, that is, the calculation resource required by the information recommendation apparatus for updating the recommendation list for the user equipment that will use the recommendation list may be determined according to the data of the determined user equipment that will use the recommendation list and the time for updating the recommendation list for each user equipment, and then the resource is applied by the cluster management tool, so that the resource may be effectively utilized. The computing resources may be, but are not limited to, CPU resources and/or memory resources.
For example, if one CPU can update the recommendation list of 100 user devices in one minute, and the recommendation list of 1000 user devices needs to be updated within 1 minute, 10 CPUs need to be applied to the system; 1 user equipment needs 1M memory to complete the calculation of the recommendation list, and then 1000 user equipments will apply for 1000M memory.
The embodiment of the invention provides an information recommendation method, which comprises the steps of obtaining a first prediction result at a first time, wherein the first prediction result comprises predicted user equipment which will use a recommendation list from a second time to a third time, the first time is earlier than the second time, and the second time is earlier than the third time; and generating a recommendation list for the predicted user equipment which will use the recommendation list in the second time to the third time according to the first prediction result.
Based on the description of the above embodiment, the present invention can predict the user devices using the recommendation list and update the recommendation list for the user devices using the recommendation list by acquiring the user devices that will use the recommendation list from the second time to the third time at the first time and generating the recommendation list for the user devices. According to the technical scheme, the recommendation list does not need to be updated for all the history active users, so that the number of users needing to update the recommendation list is reduced, computing resources are saved, and the recommendation list can be updated for the user equipment which will use the recommendation list in time.
EXAMPLE III
An embodiment of the present invention provides an information recommendation apparatus, as shown in fig. 4, which is a schematic structural diagram of the apparatus, and includes:
a first obtaining module 10, configured to obtain a first prediction result at a first time, where the first prediction result includes a predicted user equipment that will use the recommendation list from a second time to a third time, the first time is earlier than the second time, and the second time is earlier than the third time.
And the generating module 11 is configured to generate a recommendation list for the predicted user equipment that will use the recommendation list from the second time to the third time according to the first prediction result.
As shown in fig. 5, the apparatus further includes:
the second obtaining module 12 is configured to obtain a second prediction result before obtaining the first prediction result at the first time, where the second prediction result includes a predicted user equipment that will use the recommendation list within a preset time period and a predicted usage time, the preset time period is a period of obtaining the second prediction result, the predicted usage time is a time when the user equipment that will use the recommendation list uses the recommendation list within the preset time period, and the first time, the second time, and the third time are within the preset time period.
The first obtaining module 10 is specifically configured to read, at the first time, the user equipment whose usage time is from the second time to the third time that will use the recommendation list.
The second obtaining module 12 specifically includes:
the obtaining sub-module 120 is configured to obtain a prediction model, where the prediction model is used to calculate a probability that each user equipment will use the recommendation list within a preset time period.
The calculating submodule 121 is configured to bring the parameters of each user equipment into the prediction model, and calculate a probability that each user equipment will use the recommendation list within a preset time period.
The setting sub-module 122 is configured to set the first user equipment as the user equipment that will use the recommendation list within the preset time period if the probability that the first user equipment will use the recommendation list within the preset time period is greater than the preset probability, where the first user equipment is any one of the user equipments.
The prediction model is y (x) 1/1+ e- (. beta.)0+β1x1+...+βnxn) (ii) a Wherein y (x) is the probability of the first user device using the recommendation list; beta is a0、β1…βnParameters of the prediction model; x is the number of1…xnIs a parameter of the first user equipment.
The device also includes:
a receiving module 13, configured to receive a recommendation list request message sent by a user equipment that is predicted to use the recommendation list from the second time to the third time.
A sending module 14, configured to, after generating the recommendation list for the user equipment according to the first prediction result, send the recommendation list to the predicted user equipment that will use the recommendation list in the second time to the third time before the second time or at the second time; alternatively, the receiving module 13 receives the recommendation list request message sent by the user equipment predicted to use the recommendation list from the second time to the third time, and sends the recommendation list to the user equipment predicted to use the recommendation list from the second time to the third time.
The embodiment of the invention provides an information recommendation device, which comprises a first obtaining module, a second obtaining module and a recommendation module, wherein the first obtaining module is used for obtaining a first prediction result at a first time, the first prediction result comprises predicted user equipment which will use a recommendation list within a second time to a third time, the first time is earlier than the second time, and the second time is earlier than the third time; and the generation module is used for generating a recommendation list for the predicted user equipment which uses the recommendation list from the second time to the third time according to the first prediction result.
Based on the description of the above embodiment, the present invention can predict the user devices using the recommendation list and update the recommendation list for the user devices using the recommendation list by acquiring the user devices that will use the recommendation list from the second time to the third time at the first time and generating the recommendation list for the user devices. According to the technical scheme, the recommendation list does not need to be updated for all the history active users, so that the number of users needing to update the recommendation list is reduced, computing resources are saved, and the recommendation list can be updated for the user equipment which will use the recommendation list in time.
Example four
An embodiment of the present invention provides an information recommendation apparatus, as shown in fig. 6, including:
a processor 20 configured to obtain a first prediction result at a first time, wherein the first prediction result includes predicted user equipment that will use the recommendation list from a second time to a third time, the first time being earlier than the second time, and the second time being earlier than the third time; the processor 20 is further configured to generate a recommendation list for the predicted user equipment that will use the recommendation list from the second time to the third time based on the first prediction result.
The processor 20 is further configured to obtain a second prediction result before obtaining the first prediction result at the first time, where the second prediction result includes a predicted user equipment that will use the recommendation list within a preset time period and a predicted usage time, the preset time period is a period of obtaining the second prediction result, the predicted usage time is a time when the user equipment that will use the recommendation list uses the recommendation list within the preset time period, and the first time, the second time, and the third time are within the preset time period.
The processor 20 is specifically configured to read, at the first time, the user equipment whose usage time is from the second time to the third time, which will use the recommendation list.
The processor 20 is specifically configured to obtain a prediction model, where the prediction model is used to calculate a probability that each user equipment will use the recommendation list within a preset time period; the system comprises a prediction model, a recommendation list and a user equipment, wherein the prediction model is used for respectively bringing parameters of each user equipment into the prediction model and calculating the probability that each user equipment will use the recommendation list in a preset time period; and if the probability that the first user equipment will use the recommendation list in the preset time period is greater than the preset probability, setting the first user equipment as the user equipment that will use the recommendation list in the preset time period, wherein the first user equipment is any one of the user equipment.
The prediction model is y (x) 1/1+ e- (. beta.)0+β1x1+...+βnxn) (ii) a Wherein y (x) is the probability of the first user device using the recommendation list; beta is a0、β1…βnParameters of the prediction model; x is the number of1…xnIs a parameter of the first user equipment.
A receiver 21, configured to receive a recommendation list request message sent by a user equipment that is predicted to use the recommendation list from the second time to the third time.
A transmitter 22, configured to, after generating the recommendation list for the user equipment according to the first prediction result, transmit the recommendation list to the user equipment predicted to use the recommendation list in the second time to the third time before the second time or at the second time; alternatively, the receiver 21 receives the recommendation list request message transmitted from the user device predicted to use the recommendation list from the second time to the third time, and transmits the recommendation list to the user device predicted to use the recommendation list from the second time to the third time.
And a memory 23 for storing the attributes of the user device, including the attributes of the user and the attributes of the information.
The embodiment of the invention provides an information recommendation device, wherein a processor is used for obtaining a first prediction result at a first time, the first prediction result comprises predicted user equipment which will use a recommendation list from a second time to a third time, the first time is earlier than the second time, and the second time is earlier than the third time; and generating a recommendation list for the predicted user equipment which will use the recommendation list in the second time to the third time according to the first prediction result.
Based on the description of the above embodiment, the present invention can predict the user devices using the recommendation list and update the recommendation list for the user devices using the recommendation list by acquiring the user devices that will use the recommendation list from the second time to the third time at the first time and generating the recommendation list for the user devices. According to the technical scheme, the recommendation list does not need to be updated for all the history active users, so that the number of users needing to update the recommendation list is reduced, computing resources are saved, and the recommendation list can be updated for the user equipment which will use the recommendation list in time.
Those of ordinary skill in the art will understand that: all or part of the steps for implementing the method embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer readable storage medium, and when executed, the program performs the steps including the method embodiments; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.
Claims (12)
1. A method for information recommendation, comprising:
obtaining a first prediction result at a first time, wherein the first prediction result comprises a predicted user equipment which will use a recommendation list from a second time to a third time, the first time is earlier than the second time, and the second time is earlier than the third time;
generating the recommendation list for the predicted user equipment that will use the recommendation list from the second time to the third time according to the first prediction result.
2. The method of information recommendation according to claim 1, wherein said obtaining a first prediction result at a first time is preceded by:
and acquiring a second prediction result, wherein the second prediction result comprises predicted user equipment which will use the recommendation list in a preset time period and use time, the preset time period is a period for acquiring the second prediction result, the use time is the time for the user equipment which will use the recommendation list to use the recommendation list in the preset time period, and the first time, the second time and the third time are in the preset time period.
3. The information recommendation method according to claim 2, wherein the obtaining a first prediction result at a first time specifically comprises:
and reading the user equipment which uses the recommendation list in the second time to the third time in the use time at the first time.
4. The information recommendation method according to claim 2, wherein the obtaining of the second prediction result specifically includes:
obtaining a prediction model, wherein the prediction model is used for calculating the probability that each user equipment will use a recommendation list within a preset time period;
respectively bringing the parameters of each user equipment into the prediction model, and calculating the probability that each user equipment will use a recommendation list in a preset time period;
and if the probability that the first user equipment will use the recommendation list in a preset time period is greater than the preset probability, setting the first user equipment as the user equipment which will use the recommendation list in the preset time period, wherein the first user equipment is any one of the user equipment.
5. The method of claim 4, wherein the prediction model is <math>
<mrow>
<mi>y</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mn>1</mn>
<mo>/</mo>
<mn>1</mn>
<mo>+</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>β</mi>
<mn>0</mn>
</msub>
<mo>+</mo>
<msub>
<mi>β</mi>
<mn>1</mn>
</msub>
<msub>
<mi>x</mi>
<mn>1</mn>
</msub>
<mo>+</mo>
<mo>.</mo>
<mo>.</mo>
<mo>.</mo>
<mo>+</mo>
<msub>
<mi>β</mi>
<mi>n</mi>
</msub>
<msub>
<mi>x</mi>
<mi>n</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>;</mo>
</mrow>
</math>
Wherein y (x) is a probability of using a recommendation list for the first user device; beta is a0、β1…βnParameters of the prediction model; x is the number of1…xnIs a parameter of the first user equipment.
6. The method of claim 1, wherein after generating the recommendation list for the user device according to the first prediction result, the method further comprises:
sending, prior to or at the second time, the recommendation list to a user device predicted to use the recommendation list within the second time to the third time;
or,
and receiving a recommendation list request message sent by the user equipment which is predicted to use the recommendation list from the second time to the third time within a second time to a third time, and sending the recommendation list to the user equipment which is predicted to use the recommendation list from the second time to the third time.
7. An apparatus for information recommendation, comprising:
a first obtaining module, configured to obtain a first prediction result at a first time, where the first prediction result includes a predicted user equipment that will use a recommendation list from a second time to a third time, the first time being earlier than the second time, and the second time being earlier than the third time;
a generating module, configured to generate the recommendation list for the predicted user equipment that will use the recommendation list from the second time to the third time according to the first prediction result.
8. The information recommendation device according to claim 7, further comprising:
the second obtaining module is configured to obtain a second prediction result before the first obtaining module obtains the first prediction result at a first time, where the second prediction result includes a predicted user equipment that will use the recommendation list within a preset time period and a predicted usage time, the preset time period is a period for obtaining the second prediction result, the predicted usage time is a time for the user equipment that will use the recommendation list to use the recommendation list within the preset time period, and the first time, the second time, and the third time are within the preset time period.
9. The information recommendation device according to claim 8, wherein the first obtaining module is specifically configured to:
and reading the user equipment which uses the recommendation list in the second time to the third time in the use time at the first time.
10. The information recommendation device according to claim 8, wherein the second obtaining module specifically includes:
the obtaining sub-module is used for obtaining a prediction model, wherein the prediction model is used for calculating the probability that each user equipment will use the recommendation list within a preset time period;
the calculation submodule is used for respectively bringing the parameters of each user equipment into the prediction model and calculating the probability that each user equipment will use the recommendation list in a preset time period;
the setting sub-module is configured to set the first user equipment as the user equipment that will use the recommendation list within a preset time period if the probability that the first user equipment will use the recommendation list within the preset time period is greater than a preset probability, where the first user equipment is any one of the user equipments.
11. The apparatus for information recommendation according to claim 10, wherein said prediction model is <math>
<mrow>
<mi>y</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mn>1</mn>
<mo>/</mo>
<mn>1</mn>
<mo>+</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>β</mi>
<mn>0</mn>
</msub>
<mo>+</mo>
<msub>
<mi>β</mi>
<mn>1</mn>
</msub>
<msub>
<mi>x</mi>
<mn>1</mn>
</msub>
<mo>+</mo>
<mo>.</mo>
<mo>.</mo>
<mo>.</mo>
<mo>+</mo>
<msub>
<mi>β</mi>
<mi>n</mi>
</msub>
<msub>
<mi>x</mi>
<mi>n</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>;</mo>
</mrow>
</math>
Wherein y (x) is a probability of using a recommendation list for the first user device; beta is a0、β1…βnParameters of the prediction model; x is the number of1…xnIs a parameter of the first user equipment.
12. The apparatus for information recommendation according to claim 7, further comprising:
a receiving module, configured to receive a recommendation list request message sent by a user equipment that is predicted to use the recommendation list from the second time to the third time;
a sending module, configured to send, after generating the recommendation list for the user equipment according to the first prediction result, the recommendation list to a user equipment predicted to use the recommendation list within the second time to the third time before the second time or at the second time; or, in a second time to a third time, after receiving a recommendation list request message sent by a user device predicted to use the recommendation list in the second time to the third time, the receiving module sends the recommendation list to the user device predicted to use the recommendation list in the second time to the third time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410684996.2A CN104462270B (en) | 2014-11-24 | 2014-11-24 | A kind of method and device of information recommendation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410684996.2A CN104462270B (en) | 2014-11-24 | 2014-11-24 | A kind of method and device of information recommendation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104462270A true CN104462270A (en) | 2015-03-25 |
| CN104462270B CN104462270B (en) | 2018-09-21 |
Family
ID=52908306
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410684996.2A Active CN104462270B (en) | 2014-11-24 | 2014-11-24 | A kind of method and device of information recommendation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104462270B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104850655A (en) * | 2015-06-04 | 2015-08-19 | 郑州悉知信息技术有限公司 | Data statistical method and device |
| CN105045831A (en) * | 2015-06-30 | 2015-11-11 | 北京奇艺世纪科技有限公司 | Message push method and apparatus |
| CN106227844A (en) * | 2016-07-27 | 2016-12-14 | 深圳市金立通信设备有限公司 | The method of a kind of application recommendation and terminal |
| CN107329882A (en) * | 2017-06-12 | 2017-11-07 | 腾讯科技（深圳）有限公司 | Obtain method and apparatus, storage medium and the electronic installation of online data |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101286217A (en) * | 2007-04-12 | 2008-10-15 | 株式会社东芝 | Information evaluation system and method for evaluating information |
| CN101556622A (en) * | 2008-04-11 | 2009-10-14 | 纳珀企业有限责任公司 | Providing expected desirability information prior to sending a recommendation |
| CN101814068A (en) * | 2009-02-24 | 2010-08-25 | 日电（中国）有限公司 | Rating prediction based project recommending method for time-sequence control and system thereof |
| CN102495837A (en) * | 2011-11-01 | 2012-06-13 | 中国科学院计算技术研究所 | Training method and system for digital information recommending and forecasting model |
| CN103870453A (en) * | 2012-12-07 | 2014-06-18 | 盛乐信息技术（上海）有限公司 | Method and method for recommending data |
-
2014
- 2014-11-24 CN CN201410684996.2A patent/CN104462270B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101286217A (en) * | 2007-04-12 | 2008-10-15 | 株式会社东芝 | Information evaluation system and method for evaluating information |
| CN101556622A (en) * | 2008-04-11 | 2009-10-14 | 纳珀企业有限责任公司 | Providing expected desirability information prior to sending a recommendation |
| CN101814068A (en) * | 2009-02-24 | 2010-08-25 | 日电（中国）有限公司 | Rating prediction based project recommending method for time-sequence control and system thereof |
| CN102495837A (en) * | 2011-11-01 | 2012-06-13 | 中国科学院计算技术研究所 | Training method and system for digital information recommending and forecasting model |
| CN103870453A (en) * | 2012-12-07 | 2014-06-18 | 盛乐信息技术（上海）有限公司 | Method and method for recommending data |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104850655A (en) * | 2015-06-04 | 2015-08-19 | 郑州悉知信息技术有限公司 | Data statistical method and device |
| CN105045831A (en) * | 2015-06-30 | 2015-11-11 | 北京奇艺世纪科技有限公司 | Message push method and apparatus |
| CN105045831B (en) * | 2015-06-30 | 2018-04-13 | 北京奇艺世纪科技有限公司 | A kind of information push method and device |
| CN106227844A (en) * | 2016-07-27 | 2016-12-14 | 深圳市金立通信设备有限公司 | The method of a kind of application recommendation and terminal |
| CN107329882A (en) * | 2017-06-12 | 2017-11-07 | 腾讯科技（深圳）有限公司 | Obtain method and apparatus, storage medium and the electronic installation of online data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104462270B (en) | 2018-09-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10679169B2 (en) | Cross-domain multi-attribute hashed and weighted dynamic process prioritization | |
| CN107683486B (en) | Personally influential changes to user events | |
| US11188950B2 (en) | Audience expansion for online social network content | |
| RU2721176C2 (en) | Systems and methods for predicting user behavior based on location data | |
| US9262716B2 (en) | Content response prediction | |
| US11836654B2 (en) | On-demand resource scheduling | |
| EP3796134B1 (en) | User behavior recognition method and behavior recognition system | |
| US20170178253A1 (en) | User data store for online advertisement events | |
| US20160063560A1 (en) | Accelerating engagement of potential buyers based on big data analytics | |
| Benedito et al. | Optimal manufacturing policy in a reverse logistic system with dependent stochastic returns and limited capacities | |
| Choudhury et al. | A batch arrival unreliable server delaying repair queue with two phases of service and Bernoulli vacation under multiple vacation policy | |
| TW201923675A (en) | Information recommendation method, device and apparatus | |
| CN112231533A (en) | Data processing method, device, equipment and storage medium | |
| CN104462270B (en) | A kind of method and device of information recommendation | |
| CN107463580B (en) | Click rate estimation model training method and device and click rate estimation method and device | |
| US10212121B2 (en) | Intelligent scheduling for employee activation | |
| CN111080417A (en) | Processing method for improving booking smoothness rate, model training method and system | |
| Grasman et al. | Setting basestock levels in multi-product systems with setups and random yield | |
| WO2022068280A1 (en) | Data processing method and apparatus, device, and storage medium | |
| JP2016524227A (en) | Application ranking calculation device and usage information collection device | |
| JP2019145006A (en) | Selection device, selection method and selection program | |
| Hao et al. | Look-ahead control of conveyor-serviced production station by using potential-based online policy iteration | |
| Khan et al. | Periodic review inventory model with normally distributed fuzzy random variable demand | |
| US20100082405A1 (en) | Multi-period-ahead Forecasting | |
| US10621622B1 (en) | Adaptive sequencing of notifications in a client server architecture |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200214 Address after: 518129 Bantian HUAWEI headquarters office building, Longgang District, Guangdong, Shenzhen Patentee after: HUAWEI TECHNOLOGIES Co.,Ltd. Address before: 210012 HUAWEI Nanjing base, 101 software Avenue, Yuhuatai District, Jiangsu, Nanjing Patentee before: Huawei Technologies Co.,Ltd. |