CN108389066B - Content distribution method, device, system and computer readable storage medium - Google Patents
Content distribution method, device, system and computer readable storage medium Download PDFInfo
- Publication number
- CN108389066B CN108389066B CN201710063920.1A CN201710063920A CN108389066B CN 108389066 B CN108389066 B CN 108389066B CN 201710063920 A CN201710063920 A CN 201710063920A CN 108389066 B CN108389066 B CN 108389066B
- Authority
- CN
- China
- Prior art keywords
- content
- user group
- distributed
- preference
- label
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/55—Push-based network services
Abstract
The invention provides a content distribution method and a content distribution system. The method comprises the following steps: determining a content tag and a corresponding confidence for each content to be distributed; for each user group, determining a user group label and a corresponding confidence level; determining the preference degree of each user group to each content to be distributed according to the relevance and the corresponding confidence degree of the user group label of each user group and the content label of each content to be distributed; and distributing contents to each user group according to the preference degrees so that the total preference degree reaches the maximum.
Description
Technical Field
The present invention relates to content push technology, and in particular, to a content distribution method, device, and system based on dynamic programming, and a computer-readable storage medium.
Background
Personalized content and services have become an important competitive power for internet products at present. The application of personalized recommendation technology is seldom available no matter the content recommendation of music and news products or the operation of e-commerce websites. In the field of e-commerce website operations, there are several different content (e.g., activities, merchandise, advertisements, etc.) available for the resource slots that are presented to the user. The commonly employed approach is to present the highest quality content to the full number of users. Although the method is the most direct and universal method and has a good overall effect, the method does not take the characteristics of the user into consideration, cannot show the most appropriate content for the user in a targeted manner, and can cause waste of a large amount of unselected content. Therefore, for different users, the content most suitable for them should be distributed according to their preference, so as to optimize the overall display effect (e.g. click rate/conversion rate/amount of incoming order, etc.).
In order to achieve targeted presentation of appropriate content to users, another approach is to divide users into a number of user groups, evaluate the preference scores of the user groups for the respective content, and then manually distribute the content to each user group by an operator. Although this method can achieve the objective of targeted distribution of content to user groups, it consumes operational labor, often requiring the operation to be quite familiar with the partitioning of individual content and user groups. Under the conditions of large internal capacity, more user groups and high updating frequency, the method has low efficiency. In addition, there is another method that can assign content to users in a targeted manner, which introduces heuristic algorithms to distribute the content reasonably so that a large amount of content and user groups can be handled. However, the heuristic algorithm can only find a locally optimal solution, not an optimal distribution scheme, and cannot achieve an optimal distribution effect.
Therefore, a method and a system for content distribution based on dynamic programming are needed, so that the overall display effect can be optimized.
Disclosure of Invention
An aspect of the present disclosure is to address at least the above problems and/or disadvantages and to provide at least the advantages described below.
According to a first aspect of the present invention, there is provided a content distribution method including: determining a content tag and a corresponding confidence for each content to be distributed; for each user group, determining a user group label and a corresponding confidence level; determining the preference degree of each user group to each content to be distributed according to the relevance and the corresponding confidence degree of the user group label of each user group and the content label of each content to be distributed; and distributing contents to each user group according to the preference degrees so that the total preference degree reaches the maximum.
Preferably, said determining content tags and corresponding confidences comprises: the content attribute is determined as a content tag.
Preferably, the content attribute includes one or more of a commodity set, commodity classification information, store information, and brand information.
Preferably, the method further comprises: storing content tags and corresponding confidences with the content as a hive table.
Preferably, the method further comprises extracting a maximum number of distributed user groups allowed by the content.
Preferably, the determining the user group label and the corresponding confidence comprises: and determining the user group label according to the historical behavior record of the user group.
Preferably, the method further comprises: storing a user group label and corresponding confidence with the user group as a hive table.
Preferably, the determining the preference of each user group for each to-be-distributed content according to the relevance and the respective corresponding confidence of the user group label of each user group and the content label of each to-be-distributed content includes: if one or more content tags of the content to be distributed are associated with one or more user group tags of the user group, calculating products of the respective confidences of the content tags and the respective confidences of the user group tags for each associated content tag and user group tag group, respectively, and summing the products calculated for each associated content tag and user group tag group, determining the sum as a base preference score of the user group for the content; and if any content tag of the content to be distributed is not associated with any user group tag of the user group, recording the basic preference score of the user group for the content as 0; and determining the preference degree of each user group to each content to be distributed based on the basic preference scores.
Preferably, the method further comprises storing the base preference score for each user group for each content to be distributed as a hive table.
Preferably, the method further comprises evaluating the quality score of the content, wherein the determining the preference of each user group for each content to be distributed according to the relevance of the user group label of each user group and the content label of each content to be distributed and the respective corresponding confidence comprises: calculating a product of a base preference score and a quality score of the content, the product being determined as a final preference score for the content for the group of users; and determining the preference degree of each user group to each content to be distributed based on the final preference score.
Preferably, the allocating content to each user group according to the preference degree so that the total preference degree reaches the maximum may include: acquiring the number of distributable user groups for each content; traversing each content distribution scheme for the user groups and determining a total preference for each content distribution scheme based on the acquired number of assignable user groups, with respect to each of the user groups, so as to determine and store the content distribution scheme for the user groups in a case where the total preference is maximized; arranging each user group in a reverse order; and polling each user group which is arranged in the reverse order, and acquiring the stored content distribution scheme corresponding to each user group aiming at each user group so as to distribute the content to each user group according to the preference degree, so that the total preference degree reaches the maximum.
According to a second aspect of the present invention, there is provided a content distribution apparatus, which may include: a content tagging module configured to: determining a content tag and a corresponding confidence for each content to be distributed; a user group tagging module configured to: for each user group, determining a user group label and a corresponding confidence level; a preference calculation module configured to: determining the preference degree of each user group to each content to be distributed according to the relevance and the corresponding confidence degree of the user group label of each user group determined by the user group label module and the content label of each content to be distributed determined by the content label module; and a content distribution module configured to: the content is distributed to each user group according to the preference determined by the preference calculation module so that the total preference is maximized.
Preferably, the preference degree calculation module is further configured to: if one or more content tags of the content to be distributed are associated with one or more user group tags of the user group, calculating products of the respective confidences of the content tags and the respective confidences of the user group tags for each associated content tag and user group tag, respectively, and summing the products calculated for each associated content tag and user group tag together, the sum being determined as a base preference score for the user group for the content; and if any content tag of the content to be distributed is not associated with any user group tag of the user group, recording the basic preference score of the user group for the content as 0; and determining the preference degree of each user group to each content to be distributed based on the basic preference scores.
Preferably, the content distribution module is further configured to: acquiring the number of distributable user groups for each content; traversing each content distribution scheme for the user groups and determining a total preference for each content distribution scheme based on the acquired number of distributable user groups relative to each user group in the user groups to determine and store the content distribution scheme for the user groups under the condition that the total preference is maximum; arranging each user group in a reverse order; and polling each user group which is arranged in the reverse order, and acquiring the stored content distribution scheme corresponding to each user group aiming at each user group so as to distribute the content to each user group according to the preference degree, so that the total preference degree reaches the maximum.
According to a third aspect of the present invention, there is provided a content distribution system comprising: a memory configured to store content tags and respective confidences for content to be distributed and user group tags and respective confidences for each user group; and a processor connected to the memory via a wired or wireless connection and configured to: determining a content tag and a corresponding confidence for each content to be distributed; for each user group, determining a user group label and a corresponding confidence level; determining the preference degree of each user group to each content to be distributed according to the relevance and the corresponding confidence degree of the user group label of each user group and the content label of each content to be distributed; and distributing contents to each user group according to the preference degrees so that the total preference degree reaches the maximum.
According to a third aspect of the present invention, there is provided a content distribution apparatus comprising: a memory; and a processor coupled to the memory, the processor configured to perform the content distribution method according to the first aspect of the invention based on instructions stored in the memory.
According to a fourth aspect of the present invention, there is provided a computer readable storage medium storing computer instructions which, when executed by a processor, implement the content distribution method according to the first aspect of the present invention.
Drawings
The above and other aspects, features and advantages of example embodiments of the present disclosure will become more apparent from the following description when taken in conjunction with the accompanying drawings, in which:
fig. 1 shows a block diagram of an example hardware arrangement of a content distribution system according to an example embodiment of the present invention.
Fig. 2 illustrates an operation sequence diagram of a content distribution method according to an exemplary embodiment of the present invention.
Fig. 3 shows a flow chart of a content distribution method according to an example embodiment of the present invention.
Fig. 4 shows a flowchart of a specific algorithm for determining a content distribution scheme for each user group according to an exemplary embodiment of the present invention.
Fig. 5 shows a block diagram of a content distribution apparatus according to an example embodiment of the present invention.
Detailed Description
Example implementations of the present invention are described below with reference to the accompanying drawings. The invention provides a content distribution method and a content distribution system based on dynamic programming, which can distribute the most appropriate content for a user from a plurality of contents to obtain the theoretically globally optimal distribution effect.
Fig. 1 shows a block diagram of an example hardware arrangement 100 of a content distribution system according to an example embodiment of the present invention.
Fig. 1 is a block diagram illustrating an example hardware arrangement 100 of a content distribution system according to an embodiment of the present disclosure. The hardware arrangement 100 comprises a memory 110 and a processor 120.
The memory 110 may include memory in the form of non-volatile or volatile memory, such as electrically erasable programmable read-only memory (EEPROM), flash memory, and/or a hard drive. The memory 110 may be configured to store information for the content to be distributed and the respective user groups, e.g. content tags and respective confidence levels for the content to be distributed and user group tags and respective confidence levels for each user group, etc.
Furthermore, the memory 110 may also comprise a computer program 111, which computer program 111 comprises code/computer readable instructions, which when executed by the processor 120 in the arrangement 100, cause the hardware arrangement 100 and/or the device comprising the hardware arrangement 100 to perform the procedures of the content distribution method, e.g. as described in the present invention, and any variations thereof. Further, the computer program 111 may be configured as computer program code having, for example, an architecture of computer program modules 111A-111C.
A processor 120 (e.g., a microprocessor, Digital Signal Processor (DSP), etc.). Processor 120 may be a single processing unit or multiple processing units for performing different actions of the processes described herein. The processor 120 determines, for each content to be assigned, a content tag and a corresponding confidence level by loading one or more instruction codes on the memory; for each user group, determining a user group label and a corresponding confidence level; determining the preference degree of each user group to each content to be distributed according to the relevance and the corresponding confidence degree of the user group label of each user group and the content label of each content to be distributed; and distributing contents to each user group according to the preference degrees so that the total preference degree reaches the maximum.
The processor 120 may be a single CPU (central processing unit), but may also include two or more processing units. For example, the processor 120 may include a general purpose microprocessor, an instruction set processor and/or related chip sets and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)). The processor may also include on-board memory for caching purposes. The computer program may be carried by a computer program product connected to the processor. The computer program product may comprise a computer readable medium having a computer program stored thereon. For example, the computer program product may be a flash memory, a Random Access Memory (RAM), a Read Only Memory (ROM), an EEPROM, and the above-mentioned computer program modules may in alternative embodiments be distributed in the form of a memory within the UE to the different computer program products.
Furthermore, the arrangement 100 may also comprise an input unit 102 for receiving signals from other entities, and an output unit 104 for providing signals to other entities. The input unit 102 and the output unit 104 may be arranged as a single entity or as separate entities. In an exemplary embodiment of the present invention, both the input unit and the output unit may be implemented as a touch display, and the content distribution system performs the content distribution method according to an exemplary embodiment of the present invention in response to a command to display a web page being input through the touch display. After the content is assigned in a targeted manner to the individual users, it is output via the touch display.
Although the code means in the embodiment disclosed above in connection with fig. 1 are implemented as computer program modules which, when executed in the processor 120, cause the hardware arrangement 100 to perform the operations of the content distribution method according to the invention, at least one of the code means may, in alternative embodiments, be implemented at least partly as hardware circuits.
Fig. 2 illustrates an operation sequence diagram of a content distribution method according to an exemplary embodiment of the present invention. Specifically, the content distribution method includes four parts, and first, the content is marked in operation S1. Next, in operation S3, a user group is marked. Then, at S5, the preference degree of each user group for each content is calculated. Finally at S7, content is assigned for each user group according to the content assignment logic such that the overall preference for the global is highest. That is, to have sufficient knowledge of the content to be distributed, the content is tagged. And then, the preference degree of the user group to each content is calculated by combining the user group label. And then, distributing the most matched content for each user group by using a dynamic planning method in operational research. Alternatively, the method may further include an effect evaluation step (S9) to compare the method with the original scheme to evaluate the effect.
A content distribution method according to an exemplary embodiment of the present invention is described in detail below with reference to fig. 3. For example, the technical scheme provided by the invention can be based on a hive database and a hadoop framework, and calculates the distribution result by using a redis external cache, and stores the result in a memory for a server interface to use. Hadoop is a distributed system infrastructure. A user can develop a distributed program without knowing the distributed underlying details. The power of the cluster is fully utilized to carry out high-speed operation and storage. The Hive database is a database tool based on Hadoop, can map structured data files into a database table, provides a simple sql query function, and can convert sql statements into MapReduce tasks for operation. The method has the advantages that the learning cost is low, simple MapReduce statistics can be quickly realized through SQL-like statements, special MapReduce application does not need to be developed, and the method is very suitable for statistical analysis of a database. Redis is an open source log-type and Key-Value database which is written by using ANSI C language, supports network, can be based on memory and can also be persistent, and provides API of multiple languages. It should be noted that the above implementation framework is merely exemplary, and the present invention is not limited thereto.
Fig. 3 shows a flow diagram of a content distribution method 300 according to an example embodiment of the present invention.
According to the content distribution method of the exemplary embodiment of the present invention, in S301, for each content to be distributed, a content tag and a corresponding confidence are determined. In one embodiment, the content tag may be generated by generalizing the content attributes of the content to be distributed. In the e-commerce website operation, the content to be distributed is usually a sales promotion activity, a dedicated commodity, and the like, and content attributes such as a commodity set, commodity classification information, store information, brand information, and the like can be extracted from the content as a content tag. One content may contain a plurality of tags, and one tag may correspond to a plurality of contents. The respective confidence of a content tag may represent the degree of association of the content with the respective content tag, e.g. may be represented by 0 or 1, i.e. if the content has the content tag, its confidence may be considered to be 1, otherwise its confidence may be considered to be zero. It should be noted that the confidence level may also be expressed in other ways to represent the degree of correlation of the content and the content tag. Further, the content tags and corresponding confidences with the content may be stored in memory, for example as a hive table, for example, as shown in table 1 below:
table 1 content tag hive table fields
Furthermore, in addition to extracting the content tags and corresponding confidence levels of the content to be distributed, the maximum number of distributed user groups allowed by the content may also be determined and stored in a memory as well, for example, as a hive table as shown in table 2:
table 2 maximum number of distributed subscriber groups hive table field
On the other hand, in addition to extracting the content tags and the corresponding confidences of the content to be distributed, an evaluation may also be made of the content quality, i.e. the quality score of the content is evaluated. For example, the scores are formed by means of logistic regression by inquiring historical sales of the goods and historical performance of the activities. The score may also be stored in memory, for example as a hive table as shown in fig. 3:
table 3 content quality hive table fields
At S303, for each user group, a user group label and corresponding confidence are determined. The respective confidence levels of the user tags are similar to the respective confidence levels of the content tags and are used to indicate the degree of association of the user group with the corresponding user group tags. For example, a user group tag and corresponding confidence may be determined from historical behavior records of a user group, where the user group tag may be associated with a content tag, and the association may be stored in memory. The association relationship may include identical or covering, etc. For example, if a user tag is fresh and a content tag of a content includes fruit, the user tag and the content tag are considered to be associated. According to one implementation scheme of the invention, on the basis of the existing user group label, the content related label is perfected for the user group aiming at the newly added content label. It should be clear that one user group may contain multiple tags, and one tag may correspond to multiple user groups. Further, the user group labels and corresponding confidences may be stored in memory, for example as a hive table, as shown in table 4:
table 4 user group tag hive table field
Next, in S305, a preference degree of each user group for each content to be distributed is determined according to the relevance and the corresponding confidence of the user group tag of each user group and the content tag of each content to be distributed. Specifically, if one or more content tags of the content to be distributed are associated with one or more user group tags of the user group, products of the respective confidences of the content tags and the respective confidences of the user group tags are calculated for each associated content tag and user group tag group, respectively, and the products calculated for each associated content tag and user group tag group are added and summed, the sum being determined as a base preference score of the user group for the content. And if any content label of the content to be distributed is not associated with any user group label of the user group, recording the basic preference score of the user group to the content as 0. And then determining the preference degree of each user group to each content to be distributed based on the basic preference scores. Further, after the base preference score is calculated, a product of the base preference score and the quality score of the content may be calculated, and the product may be determined as a final preference score of the group of users for the content. And finally, determining the preference degree of each user group to each content to be distributed based on the final preference score. Further, the base preference score or the final preference score for each user group for each content to be distributed may also be stored as a hive table, as shown in table 5:
TABLE 5 user group to content preference score Table field
In S307, content is distributed to each user group according to the preference degrees, so that the total preference degree reaches the maximum. For example, to ensure that the maximum overall preference is finally obtained globally, the allocation algorithm may be designed as follows:
first, assuming that the content list is N and the user group list is M, the maximum number of user groups allocated to the content is obtained from table 2, for example, the maximum number of user groups allocated to the content j is c (j).
Next, a preference score matrix is obtained from Table 5, e.g., the preference score of the user group i for content j can be represented as w (i, content)j)。
Subsequently, an n-dimensional matrix f (g1, g2... gn) is defined, representing the highest total preference score that can be obtained with g1 user groups assigned to the 1 st content, g2 user groups … … assigned to the 2 nd content, and gn user groups assigned to the nth content.
Then, for the ith user group, there are at most n +1 different allocation schemes: content n is distributed without distributing any content, distribution content 1, distribution content 2 … …, i.e., derived:
f(g1，g2...gn)＝max(f(g1，g2...gn)，f(g1-1，g2...gn)+w(i，content1)，f(g1，g2-1...gn)+w(i，content2)…f(g1，g2...gn-1)+w(i，contentn))
by the above equation it can be determined how to assign content to the user group i to ensure that the highest global preference score is achieved.
Furthermore, an s (g1, g2... gn) matrix is defined for each user group. For any user group i, s (g1, g2... gn) indicates what allocation strategy the user group i should adopt in the case that the first content is allocated to g1 user groups and the second content is allocated to g2 user groups … … to allocate the nth content to gn user groups. That is, the value of s (g1, g2... gn) represents what strategy was used to calculate f (g1, g2... gn) in the derivation process described above: when the user group is not allocated, namely max in the f derivation formula takes the first item and is recorded with 0; when the user group is assigned as content 1, namely max takes the second item in the above f derivation formula, and takes 1; when the user group is assigned as content j, i.e. max takes the j +1 th item in the above f-derived expression, note j. The s matrix should be kept in an external buffer.
In this way, the user group list is traversed in a reverse order, an s matrix is taken for each user group to perform backward extrapolation, that is, on the premise of ensuring the global highest total preference score, the allocation policy of each user group is determined, and the code of the allocation policy can be expressed as:
the specific algorithm for determining the content distribution scheme for each user group includes: acquiring the number of distributable user groups for each content; traversing each content distribution scheme for the user groups and determining a total preference for each content distribution scheme based on the acquired number of assignable user groups, with respect to each of the user groups, so as to determine and store the content distribution scheme for the user groups in a case where the total preference is maximized; arranging each user group in a reverse order; and polling each user group which is arranged in the reverse order, and acquiring the stored content distribution scheme corresponding to each user group aiming at each user group so as to distribute the content to each user group according to the preference degree, so that the total preference degree reaches the maximum. Fig. 4 shows a flow chart of an algorithm for polling each user group in reverse order. In step S401, the number of user groups to which each content can be allocated, i.e., c (1), c (2).. c (n), is first obtained. Next, in step S403, the user group list is arranged in reverse order. Next, in step S405, it is determined whether the present flow has traversed all user groups. If all user groups have been traversed, this indicates that the content distribution has been completed, and therefore the method of distributing content according to the present invention ends. If all the user groups are not traversed, step S407 is performed. In step S407, the next user group is searched from the user group list. In step S409, the S-matrix for the user group is acquired from the external cache. In step S411, the value of S (c (1), c (2).. c (n)), that is, the content number j assigned to the user group is output. In step S413, when the content number j is not equal to 0, that is, when the content j is assigned to the user group, the number of assignable user groups of the corresponding content is decreased by one and returns to operation S405.
The above algorithm is only an example way to find the global optimal solution, and those skilled in the art can design other schemes to solve the global optimal solution.
Furthermore, the present invention can also be realized as a content distribution apparatus. Specifically, the content distribution apparatus 500 may include: a content tagging module 510 configured to: determining a content tag and a corresponding confidence for each content to be distributed; a user group tagging module 520 configured to: for each user group, determining a user group label and a corresponding confidence level; a preference calculation module 530 configured to: determining the preference of each user group for each content to be distributed according to the relevance and the corresponding confidence of the user group label of each user group determined by the user group label module 520 and the content label of each content to be distributed determined by the content label module 510; and a content distribution module 540 configured to: content is distributed to each user group according to the preference determined by the preference calculation module 530 such that the total preference is maximized.
In one embodiment, the preference calculation module 530 is further configured to: if one or more content tags of the content to be distributed are associated with one or more user group tags of the user group, calculating products of the respective confidences of the content tags and the respective confidences of the user group tags for each associated content tag and user group tag, respectively, and summing the products calculated for each associated content tag and user group tag together, the sum being determined as a base preference score for the user group for the content; and if any content tag of the content to be distributed is not associated with any user group tag of the user group, recording the basic preference score of the user group for the content as 0; and determining the preference degree of each user group to each content to be distributed based on the basic preference scores.
Moreover, in another embodiment, the content distribution module 540 is further configured to: acquiring the number of distributable user groups for each content; traversing each content distribution scheme for the user groups and determining a total preference for each content distribution scheme based on the acquired number of assignable user groups, with respect to each of the user groups, so as to determine and store the content distribution scheme for the user groups in a case where the total preference is maximized; arranging each user group in a reverse order; and polling each user group which is arranged in the reverse order, and acquiring the stored content distribution scheme corresponding to each user group aiming at each user group so as to distribute the content to each user group according to the preference degree, so that the total preference degree reaches the maximum.
In summary, a content identification method and a content identification system based on dynamic programming are described, the method can process a large number of labels of the content and the user group based on a big data technology framework, so that the preference degree of the user group for each content can be more accurately evaluated, and under the constraint that each content needs to be distributed with different numbers of user groups, a globally optimal distribution strategy can be calculated, thereby improving the overall personalized effect.
With the techniques presented in this disclosure, an evaluation of its effect may be performed, for example, by performing an on-line ABTEST. For example, the users in each user group can be randomly divided into two groups AB. For each user group, group A distributes content according to the method provided by the scheme, group B distributes content according to the heuristic method, the test is continued for a week, feedback data are observed, and personalized effects are compared. According to the test, the technology disclosed by the invention can be used for more accurately evaluating the preference degree of the user group to each content, and improving the overall preference score, so that the optimal distribution is realized.
It should be noted that the above solution is only one specific implementation showing the inventive concept, and the invention is not limited to the above implementation. Some of the processing in the above-described implementations may be omitted or skipped without departing from the spirit and scope of the present invention.
The foregoing methods may be embodied in the form of executable program instructions by various computer devices and recorded in computer-readable recording media. In this case, the computer-readable recording medium may include a program command, a data file, a data structure, or a combination thereof alone. Meanwhile, the program command recorded in the recording medium may be specially designed or configured for the present invention or may be applied as known to those skilled in the art of computer software. The computer-readable recording medium includes a magnetic medium such as a hard disk, a floppy disk, or a magnetic tape, an optical medium such as a compact disc read only memory (CD-ROM) or a Digital Versatile Disc (DVD), a magneto-optical medium such as a magneto-optical floppy disk, and a hardware device such as a ROM, a RAM, a flash memory, etc. which stores and executes a program command. Further, the program command includes a machine language code formed by a compiler and a high-level language executable by a computer by using an interpreter. The foregoing hardware devices may be configured to operate as at least one software module to perform the operations of the present invention, and vice versa.
Although the operations of the methods herein are shown and described in a particular order, the order of the operations of each method may be changed so that the particular operations may be performed in a reverse order or so that the particular operations may be performed at least partially concurrently with other operations. Furthermore, the present invention is not limited to the above-described exemplary embodiments, and it may include one or more other components or operations or omit one or more other components or operations without departing from the spirit and scope of the present disclosure.
The present invention has been shown in connection with the preferred embodiments of the present invention, but it will be understood by those skilled in the art that various modifications, substitutions and changes may be made thereto without departing from the spirit and scope of the present invention. Accordingly, the present invention should not be limited by the above-described embodiments, but should be defined by the appended claims and their equivalents.
Claims (12)
1. A content distribution method, comprising:
determining a content tag and a corresponding confidence for each content to be distributed;
for each user group, determining a user group label and a corresponding confidence level;
determining the preference degree of each user group to each content to be distributed according to the relevance of the user group label of each user group and the content label of each content to be distributed and the corresponding confidence degrees respectively corresponding to the user group label and the content label; and
distributing the contents to be distributed to each user group according to the preference degrees to ensure that the total preference degree reaches the maximum,
wherein distributing the content to be distributed to each user group according to the preference degrees so that the total preference degree reaches the maximum comprises:
for each user group of the user groups, performing the following operations:
acquiring the number of distributable user groups aiming at each content to be distributed;
determining and storing a target allocation scheme for the user group by traversing each allocation scheme for the user group based on the acquired number of allocable user groups, the target allocation scheme supporting maximizing the total preference; and
updating the number of distributable user groups corresponding to the contents to be distributed of the target distribution scheme; and
and distributing the content to be distributed of the corresponding target distribution scheme to each user group so as to maximize the total preference.
2. The method of claim 1, wherein the determining content tags and corresponding confidences comprises: the content attribute is determined as a content tag.
3. The method of claim 1, further comprising: determining a maximum number of assigned user groups allowed for the content.
4. The method of claim 1, wherein the determining user group labels and corresponding confidences comprises: and determining the user group label according to the historical behavior record of the user group.
5. The method of claim 1, wherein the determining the preference of each user group for each content to be distributed according to the relevance of the user group label of each user group to the content label of each content to be distributed and the respective corresponding confidence comprises:
evaluating a quality score of the content;
if one or more content tags of the content to be distributed are associated with one or more user group tags of the user group, calculating, for each associated content tag and user group tag, a product of a respective confidence of the content tag and a respective confidence of the user group tag, respectively; adding products calculated aiming at each associated content label and a user group label to be used as a basic preference score of the user group to the content to be distributed; and is
And calculating the product of the basic preference score and the quality score as the preference degree of the user group to the content to be distributed.
6. The method of claim 5, further comprising:
and if any content label of the content to be distributed is not associated with any user group label of the user group, recording the basic preference score of the user group to the content to be distributed as 0.
7. A content distribution apparatus comprising:
a content tagging module configured to: determining a content tag and a corresponding confidence for each content to be distributed;
a user group tagging module configured to: for each user group, determining a user group label and a corresponding confidence level;
a preference calculation module configured to: determining the preference degree of each user group to each content to be distributed according to the relevance of the user group label of each user group determined by the user group label module and the content label of each content to be distributed determined by the content label module and the corresponding confidence degrees respectively corresponding to the user group label and the content label; and
a content distribution module configured to: distributing the contents to be distributed to each user group according to the preference determined by the preference calculation module so that the total preference reaches the maximum,
wherein the content distribution module is further configured to:
for each user group of the user groups, performing the following operations:
acquiring the number of distributable user groups aiming at each content to be distributed;
determining and storing a target allocation scheme for the user group by traversing each allocation scheme for the user group based on the acquired number of allocable user groups, the target allocation scheme supporting maximizing the total preference; and
updating the number of distributable user groups corresponding to the contents to be distributed of the target distribution scheme; and
and distributing the content to be distributed of the corresponding target distribution scheme to each user group so as to maximize the total preference.
8. The apparatus of claim 7, wherein the preference calculation module is further configured to:
evaluating a quality score of the content;
if one or more content tags of the content to be distributed are associated with one or more user group tags of the user group, calculating, for each associated content tag and user group tag, a product of a respective confidence of the content tag and a respective confidence of the user group tag, respectively; adding products calculated aiming at each associated content label and a user group label to be used as a basic preference score of the user group to the content; and is
And calculating the product of the basic preference score and the quality score as the preference degree of the user group to the content to be distributed.
9. The apparatus of claim 8, wherein the preference calculation module is further configured to:
and if any content label of the content to be distributed is not associated with any user group label of the user group, recording the basic preference score of the user group to the content to be distributed as 0.
10. A content distribution system comprising:
a memory configured to store content tags and respective confidences for content to be distributed and user group tags and respective confidences for each user group; and
a processor connected to the memory via a wired or wireless connection and configured to perform the content distribution method of any one of claims 1-6.
11. A content distribution apparatus comprising:
a memory; and
a processor coupled to the memory, the processor configured to perform the content distribution method of any of claims 1-6 based on instructions stored in the memory.
12. A computer-readable storage medium storing computer instructions which, when executed by a processor, implement the content distribution method of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710063920.1A CN108389066B (en) | 2017-02-03 | 2017-02-03 | Content distribution method, device, system and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710063920.1A CN108389066B (en) | 2017-02-03 | 2017-02-03 | Content distribution method, device, system and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108389066A CN108389066A (en) | 2018-08-10 |
| CN108389066B true CN108389066B (en) | 2022-02-01 |
Family
ID=63076040
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710063920.1A Active CN108389066B (en) | 2017-02-03 | 2017-02-03 | Content distribution method, device, system and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108389066B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109961308B (en) * | 2017-12-25 | 2021-05-25 | 北京京东尚科信息技术有限公司 | Method and apparatus for evaluating tag data |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104504098A (en) * | 2014-12-29 | 2015-04-08 | 北京奇虎科技有限公司 | Information recommending method and device |
| CN104915861A (en) * | 2015-06-15 | 2015-09-16 | 浙江经贸职业技术学院 | An electronic commerce recommendation method for a user group model constructed based on scores and labels |
| CN105740468A (en) * | 2016-03-07 | 2016-07-06 | 达而观信息科技（上海）有限公司 | Individuation recommendation method and system combined with content publisher information |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2589932B2 (en) * | 1992-06-15 | 1997-03-12 | インターナショナル・ビジネス・マシーンズ・コーポレイション | Global optimization method and system for device allocation |
| US7283971B1 (en) * | 2000-09-06 | 2007-10-16 | Masterlink Corporation | System and method for managing mobile workers |
| US20070220103A1 (en) * | 2006-03-14 | 2007-09-20 | Michael Rogers | Targeting personalized content to mobile users |
| US9235853B2 (en) * | 2012-09-11 | 2016-01-12 | Google Inc. | Method for recommending musical entities to a user |
| CN106327090A (en) * | 2016-08-29 | 2017-01-11 | 安徽慧达通信网络科技股份有限公司 | Real task allocation method applied to preference crowd-sourcing system |
-
2017
- 2017-02-03 CN CN201710063920.1A patent/CN108389066B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104504098A (en) * | 2014-12-29 | 2015-04-08 | 北京奇虎科技有限公司 | Information recommending method and device |
| CN104915861A (en) * | 2015-06-15 | 2015-09-16 | 浙江经贸职业技术学院 | An electronic commerce recommendation method for a user group model constructed based on scores and labels |
| CN105740468A (en) * | 2016-03-07 | 2016-07-06 | 达而观信息科技（上海）有限公司 | Individuation recommendation method and system combined with content publisher information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108389066A (en) | 2018-08-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10452612B2 (en) | Efficient data caching management in scalable multi-stage data processing systems | |
| US20220335338A1 (en) | Feature processing tradeoff management | |
| US11809408B2 (en) | Incremental refresh of a materialized view | |
| US10372723B2 (en) | Efficient query processing using histograms in a columnar database | |
| US20200050968A1 (en) | Interactive interfaces for machine learning model evaluations | |
| US11100420B2 (en) | Input processing for machine learning | |
| US20150379425A1 (en) | Consistent filtering of machine learning data | |
| US20080065721A1 (en) | Proximity-based web page content placement mechanism | |
| US9904705B2 (en) | Data table performance optimization | |
| JP5253519B2 (en) | Method, apparatus and storage medium for generating smart text | |
| US8332381B1 (en) | Tab search utility | |
| WO2019085463A1 (en) | Department demand recommendation method, application server, and computer-readable storage medium | |
| US20110179013A1 (en) | Search Log Online Analytic Processing | |
| US8341181B2 (en) | Method for performance tuning a database | |
| US20130238972A1 (en) | Look-alike website scoring | |
| CN111639077A (en) | Data management method and device, electronic equipment and storage medium | |
| CN114328632A (en) | User data analysis method and device based on bitmap and computer equipment | |
| CN108389066B (en) | Content distribution method, device, system and computer readable storage medium | |
| US20030160609A9 (en) | Method and facility for storing and indexing web browsing data | |
| CN110147381B (en) | Information processing method, system and recording medium | |
| Blamey et al. | Adapting the secretary hiring problem for optimal hot-cold tier placement under top-K workloads | |
| US10628452B2 (en) | Providing multidimensional attribute value information | |
| CN101796493B (en) | information search system, information search method, and program | |
| Diener et al. | Modeling memory access behavior for data mapping | |
| KR20230153189A (en) | Apparatus and method for analyzing customers visiting merchants based on card approval list |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |