CN111737305B - Method and device for determining characteristics of people group in information push and readable storage medium - Google Patents
Method and device for determining characteristics of people group in information push and readable storage medium Download PDFInfo
- Publication number
- CN111737305B CN111737305B CN202010461899.2A CN202010461899A CN111737305B CN 111737305 B CN111737305 B CN 111737305B CN 202010461899 A CN202010461899 A CN 202010461899A CN 111737305 B CN111737305 B CN 111737305B
- Authority
- CN
- China
- Prior art keywords
- information
- determining
- crowd
- characteristic
- 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
- 238000000034 method Methods 0.000 title claims abstract description 30
- 230000011218 segmentation Effects 0.000 claims abstract description 41
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000012216 screening Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000001052 transient effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 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
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2457—Query processing with adaptation to user needs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/219—Managing data history or versioning
Abstract
A crowd characteristic determining method and device in information push and a readable storage medium, wherein the method comprises the following steps: acquiring a history push data set of history push information which is the same as the category of the information to be pushed; dividing the history pushing data set on each dimension by adopting a preset assessment index to obtain a dividing result, wherein the dividing result comprises the following steps of: push data corresponding to each label in each dimension respectively; determining a characteristic label set according to the segmentation result; according to the feature tag set, the historical push data set is segmented, the historical push data corresponding to the feature tag set is screened out, the assessment index is continuously adopted, and the historical push data corresponding to the feature tag set is segmented on each dimension respectively until a final tag set is obtained; and determining the crowd characteristics of the crowd corresponding to the information to be pushed according to the final tag set. By means of the scheme, operators can scientifically and effectively plan target audiences.
Description
Technical Field
The embodiment of the invention relates to the field of information pushing, in particular to a method and a device for determining characteristics of people in information pushing and a readable storage medium.
Background
Before an information pushing activity is started, an advertiser or an operator needs to plan a crowd of the information pushing, such as planning crowd portraits and crowd scales of the crowd to reasonably set a budget or determine a pushing range.
Currently, the business personnel generally delimit the population according to past pushing experience and subjective knowledge. However, in the above-mentioned crowd-scope defining manner, the scientificity and effectiveness of the target audience determination are lower, resulting in lower crowd positioning accuracy, and lower crowd positioning accuracy often results in poor information pushing initial effect, wastes a large amount of budget, and results in higher cost.
Disclosure of Invention
The technical problem solved by the embodiment of the invention is how to improve the scientificity and effectiveness of the crowd characteristic determination of the target audience so as to improve the crowd positioning precision of the information to be pushed.
In order to solve the above technical problems, an embodiment of the present invention provides a method for determining crowd characteristics in information push, including: acquiring a history push data set of history push information which is the same as the category of the information to be pushed; and respectively segmenting the history pushing data set on each dimension by adopting a preset assessment index to obtain a segmentation result, wherein the segmentation result comprises the following steps of: push data corresponding to each label in each dimension respectively; determining a feature tag set according to the segmentation result, wherein the feature tag set comprises one or more feature tags; dividing the history push data set according to the characteristic tag set, screening out the history push data corresponding to the characteristic tag set, continuing to adopt the assessment index, and dividing the history push data corresponding to the characteristic tag set on each dimension respectively until a final tag set is obtained, wherein the final tag set is related to the characteristic tag set; and determining crowd characteristics of the crowd corresponding to the information to be pushed according to the final tag set.
Optionally, the step of dividing the history push data set in each dimension by using a preset assessment index to obtain a division result includes: traversing each label in the history push data in each dimension, determining push data corresponding to each label in each dimension, and calculating the information gain of each label; or traversing each label in the historical push data in each dimension, determining the push data corresponding to each label in each dimension, and calculating the coefficient of the foundation corresponding to each label.
Optionally, the determining the feature tag set according to the segmentation result includes: determining a recommendation label according to the segmentation result; and determining the characteristic label set according to the recommended labels.
Optionally, the determining the recommendation label according to the segmentation result includes: and arranging the information gain of each label in order from big to small, and taking the label with the information gain of N before ranking as the recommended label, wherein N is more than or equal to 1 and is an integer.
Optionally, the determining the recommendation label according to the segmentation result includes: and arranging the coefficient of the base of each label in order from small to large, and taking the label with the coefficient of the base ranked M as the recommended label, wherein M is more than or equal to 1 and is an integer.
Optionally, the determining the feature tag set according to the recommended tag includes: selecting one or more of the recommended labels as characteristic labels and forming the characteristic label set; or selecting one or more tags from other tags except the recommended tag as characteristic tags and forming the characteristic tag set; or selecting one or more tags from the recommended tags and other tags except the recommended tags as the characteristic tags and forming the characteristic tag set.
Optionally, the determining, according to the final tag set, the crowd characteristics of the crowd corresponding to the information to be pushed includes: and determining crowd characteristics of the crowd corresponding to the information to be pushed according to the scale of the historical push data corresponding to each tag in the final tag set, wherein the crowd characteristics are part or all of the tags in the final tag set.
Optionally, after determining the crowd characteristics of the crowd corresponding to the information to be pushed, the method further includes: and pushing the information to be pushed according to the crowd characteristics.
Optionally, the assessment index includes: click rate or conversion rate.
The embodiment of the invention also provides a crowd characteristic determining device in information pushing, which comprises the following steps: the acquisition unit acquires a history push data set of history push information which is the same as the category of the information to be pushed; the segmentation unit is used for respectively segmenting the history pushing data set in each dimension by adopting a preset assessment index to obtain a segmentation result, wherein the segmentation result comprises the following steps of: push data corresponding to each label in each dimension respectively, wherein the assessment index comprises: click rate or conversion rate; the first determining unit is used for determining a characteristic label set according to the segmentation result, wherein the characteristic label set comprises one or more characteristic labels; the second determining unit is suitable for cutting the history push data set according to the characteristic tag set, screening out the history push data corresponding to the characteristic tag set, continuing to adopt the assessment index, and respectively cutting the history push data corresponding to the characteristic tag set in each dimension until a final tag set is obtained, wherein the final tag set is related to the characteristic tag set; and the third determining unit is suitable for determining crowd characteristics of the crowd corresponding to the information to be pushed according to the final tag set.
The embodiment of the invention also provides a crowd characteristic determining device in information pushing, which comprises a memory and a processor, wherein the memory stores computer instructions capable of running on the processor, and the processor executes the steps of any crowd characteristic determining method in information pushing when running the computer instructions.
The embodiment of the invention also provides a computer readable storage medium, which is a nonvolatile storage medium or a non-transient storage medium, and is provided with computer instructions stored thereon, wherein the computer instructions execute the steps of any crowd characteristic determining method in information pushing when running.
Compared with the prior art, the technical scheme of the embodiment of the invention has the following beneficial effects:
the crowd characteristics of the crowd corresponding to the information to be pushed are obtained by adopting a mode that based on a history pushing data set of the history pushing information which is the same as the class of the information to be pushed, preset assessment indexes are adopted to segment the history pushing data in multiple dimensions, a characteristic tag set is determined according to segmentation results, the history pushing data set is further segmented according to the characteristic tag set, the history pushing data corresponding to the characteristic tag set is screened out, the assessment indexes are continuously adopted to segment the history pushing data corresponding to the characteristic tag set until a final tag set is obtained, and the crowd characteristics are obtained according to the final tag set. In addition, it is expected to improve the crowd positioning accuracy of the information to be pushed.
Drawings
Fig. 1 is a flowchart of a crowd feature determining method in information push in an embodiment of the invention;
fig. 2 is a schematic structural diagram of a device for determining characteristics of people in information push according to an embodiment of the present invention.
Detailed Description
As mentioned above, the population is currently generally demarcated by business personnel based on past push experience and subjective knowledge. However, in the above-mentioned crowd-scale demarcation method, because crowd positioning accuracy is low, the initial information pushing effect is often poor, a large amount of budget is wasted, and the cost is high.
In the embodiment of the invention, the crowd characteristics of the crowd corresponding to the information to be pushed are obtained by adopting a mode that the historical push data set of the historical push information which is the same as the class of the information to be pushed is based, the historical push data is segmented in a plurality of dimensions by adopting a preset assessment index, the characteristic tag set is determined according to the segmentation result, the historical push data set is segmented according to the characteristic tag set, the historical push data corresponding to the characteristic tag set is screened out, the assessment index is continuously adopted to segment the historical push data corresponding to the characteristic tag set until a final tag set is obtained, and the crowd characteristics are obtained according to the final tag set.
In addition, through the scientific and effective planning target audience, the crowd positioning accuracy of the information to be pushed can be improved, and then the information pushing effect is improved.
In order to make the above objects, features and advantages of the embodiments of the present invention more comprehensible, the following detailed description of the embodiments of the present invention refers to the accompanying drawings.
Referring to fig. 1, a flowchart of a crowd feature determining method in information push in an embodiment of the present invention is provided, which may include the following steps:
and 11, acquiring a history push data set of the history push information which is the same as the category of the information to be pushed.
In particular implementations, the information to be pushed may include multiple types, such as advertisements, news information, pictures, and the like. For example, the information to be pushed is a real-time bidding (Real Time Bidding, RTB) advertisement, in an information pushing scene related to the RTB advertisement, in order to accurately push the pushing information related to the RTB advertisement to a corresponding crowd, the crowd characteristics of the crowd corresponding to the information to be pushed need to be confirmed, the crowd characteristics of the crowd corresponding to the pushing information of different categories are different, wherein the crowd characteristics can be represented by a label mode.
For example, the categories of push information related to the advertising field may include mobile phone numbers, home appliances, computer offices, mother and infant products, clothing, etc., and it is understood that other classification manners and type names may also exist, which are not exemplified herein. The names of the specific categories may be different depending on the classification method, and the classification method or the category name of the category of the push information is not limited.
According to the category of the information to be pushed, a history pushing data set corresponding to the same history pushing information as the category of the information to be pushed can be obtained from the history pushing information, wherein the history pushing information can comprise a plurality of types, and correspondingly, the history pushing data set consists of a plurality of history pushing data.
For example, the category of the information to be pushed is mother and infant articles, and a history pushing data set corresponding to the category of the history pushing information of the mother and infant articles is obtained from the history pushing information.
For another example, the category of the information to be pushed is a household appliance, and the history pushing data set corresponding to the category of the history pushing information of the household appliance is obtained from the history pushing information.
In a specific implementation, the historical push data set corresponding to the same class of the historical push information can be from the same owner, can be from different owners, can be from data provided by the same information push platform, can be from data provided by a third party information push platform, and can have other paths, so that the sources of the historical push data set corresponding to the historical push information are not illustrated one by one.
And step 12, respectively segmenting the history pushing data set in each dimension by adopting a preset assessment index to obtain a segmentation result.
After the history push data set corresponding to the history push information is obtained, the history push data set can be segmented, that is, the history push data set is classified.
In the embodiment of the invention, the historical push data set is segmented on each dimension by adopting a preset assessment index to obtain a segmentation result, wherein the segmentation result can comprise push data corresponding to each label on each dimension.
The evaluation index may be a click Rate (Click Through Rate, CTR) or a Conversion Rate (CVR), and it is understood that other quantifiable business indexes may be used as the evaluation index, which is not illustrated here.
In particular implementations, the dimensions may include: regional, urban, gender, age group, occupation, hobbies, etc. Wherein the gender dimension may include tags for both males and females. Tags in the regional dimension may include south and north, tags in the urban dimension may include Beijing, shanghai, tianjin, shenzhen, and the like. The labels corresponding to the age groups can comprise young people, middle-aged people or old people, or set age groups and the like.
For example, the history push data set is segmented in the gender dimension by using the CTR as an assessment index, so as to obtain the history push data clicked by the user and having a male tag, or the history push data clicked by the user and having a female tag.
For another example, the CVR is used as an assessment index, and the history push data set is segmented in a regional dimension to obtain history push data which is successfully converted and has a south label and history push data which is successfully converted and has a north label.
In implementations, the history push data set may be sliced using two classifications.
In one embodiment of the invention, the history push data set is sliced using information gain.
Specifically, in each dimension, each tag in the history push data is traversed, and the information gain of each tag in each dimension is calculated.
In the embodiment of the invention, the information gain of each label on each dimension can be calculated by adopting the following formula (1), the information entropy of the positive sample can be calculated by adopting the following formula (2), and the conditional entropy of the label A can be calculated by adopting the formula (4):
g(D,A)＝H(D)-H(D|A)； (1)
P i ＝P(X＝i)； (3)
wherein, D is a history push data set, the history push data set comprises a plurality of history push data, and each history push data can be used as a sample; pi is the probability that the ith historical push data is a positive sample; h (D) is the information entropy of the sample; n is the total number of history push data included in the history push data set; h (D|A) is the conditional entropy of tag A; g (D, A) is the information gain of tag A; i A i The I is the total number of the positive samples of the history push data corresponding to the label A in the dimension; the A| is the total number of tags A in the historical push data set.
In a specific implementation, taking CTR as an example, the history push data clicked by the user is a positive sample, and the history push data not clicked by the user is a negative sample. Taking CVR as an example, the history push data of successful conversion is a positive sample, and the history push data of failed conversion is a negative sample.
In another embodiment of the present invention, the history push data set is sliced using a coefficient of kunning.
Specifically, each tag in the history push data is traversed in each dimension, and the coefficient of the kunit corresponding to each tag in each dimension is calculated.
In an embodiment of the present invention, the following equation (5) may be used to calculate the coefficient of the kuntze of each tag in each dimension:
wherein Gini (P) is the coefficient of the base of tag i, P i The probability that the selected history push data belongs to the label i is provided; 1-P i The probability of the data being misclassified for the selected history is pushed.
And step 13, determining a characteristic label set according to the segmentation result.
In a specific implementation, a feature tag set may be determined according to push data corresponding to each tag in each dimension in the segmentation result, where the feature tag set may include one or more feature tags.
In the embodiment of the invention, the recommended label can be determined according to the segmentation result, and then the characteristic label set is determined according to the recommended label.
In a specific implementation, according to different adopted segmentation modes, the determination modes of the recommended labels are correspondingly different.
In an embodiment of the present invention, the information gain of each tag is arranged in order from large to small, and the tag with the top N of the information gain rank is taken as the recommended tag, where N is greater than or equal to 1 and is an integer.
In another embodiment of the invention, the coefficient of the base of each label is arranged in order from small to large, the label with the top M of the coefficient of the base is taken as the recommended label, M is more than or equal to 1, and the M is an integer.
For example, the segmentation results include: the information gain of the male tag in the sex dimension is 0.67, and the information gain of the female tag is 0.22; the information gain of the north tag in the regional dimension is 0.72, and the information gain of the south tag is 0.25. And determining the recommended label as male or north.
In particular implementations, there may be a variety of ways to determine the feature tag set based on the recommended tags, as exemplified below.
In an embodiment of the invention, one or more of the recommended labels are selected as feature labels and the feature label set is formed.
In another embodiment of the present invention, one or more tags are selected from the other tags than the recommended tag as feature tags and the feature tag set is formed.
In yet another embodiment of the present invention, one or more of the recommended tags and other tags other than the recommended tags are selected as the feature tags and the feature tag set is formed.
In specific implementation, the recommendation tag can be determined according to factors such as input cost, pushing quantity, pushing time, types of feature groups and the like of information to be pushed.
And step 14, judging whether the characteristic tag set is a final tag set.
When the judgment result is yes, executing step 16; when the determination result is no, step 15 is performed.
In specific implementation, the feature tag set obtained after the iteration number reaches the preset number is used as a final tag set. Or when the obtained feature tag set meets the requirement of the user, taking the feature tag set meeting the requirement of the user as a final tag set. It will be appreciated that other manners of determination may exist and are not described in detail herein.
And 15, cutting the history push data set according to the characteristic tag set, and screening out the history push data corresponding to the characteristic tag set.
In specific implementation, when the feature tag set is not the final tag set, the history push data is segmented according to the feature tag set, and the history push data corresponding to the feature tag set is screened out. And taking the history push data corresponding to the feature tag set as the history push data in the step 12, and continuing to execute the step 12, namely continuing to adopt the assessment index, respectively segmenting the history push data corresponding to the feature tag set in each dimension, wherein each segmentation corresponds to a segmentation result until a final tag set is obtained according to the segmentation result. Wherein the final set of tags is associated with the feature set of tags.
For example, if the tag included in the feature tag set is female and Shanghai, the historical push data set is screened out that the tag is female and Shanghai corresponding historical push data. And taking the history push data with the labels corresponding to women and Shanghai as the history push data in the step 12, continuing to divide the history push data corresponding to women and Shanghai in each dimension by adopting the assessment index, and obtaining a corresponding characteristic label set in the next iteration process according to the dividing result until a final label set is obtained according to the dividing result.
And step 16, determining crowd characteristics of the crowd corresponding to the information to be pushed according to the final tag set.
In a specific implementation, after the final tag set is obtained, the crowd characteristics of the crowd corresponding to the information to be pushed can be determined according to the final tag set.
In the embodiment of the invention, the crowd characteristics can be determined according to the scale of the historical push data corresponding to each tag in the final tag set.
The crowd characteristics are related to the final tag set, and in one embodiment of the invention, part of tags in the final tag set are used as crowd characteristics of the crowd corresponding to the information to be pushed. In another embodiment of the present invention, all tags in the final tag set are used as crowd characteristics of the crowd corresponding to the information to be pushed.
For example, the final set of labels includes labels for women, shanghai, young. The scale of the history push data corresponding to the female tag is 70%, the scale of the history push data corresponding to the upper sea tag is 75%, and the scale of the history push data corresponding to the young tag is 92%. The young label can be used as the crowd characteristic of the crowd corresponding to the information to be pushed, the young label and the female label can be used as the crowd characteristic of the crowd corresponding to the information to be pushed, and the young label, the female label and the sea label can be used as the crowd characteristic of the crowd corresponding to the information to be pushed.
In a specific implementation, the pushing of the information to be pushed may be performed according to crowd characteristics.
The determination of crowd characteristics is related to factors such as pushing number and cost of information to be pushed.
In an embodiment of the present invention, the input cost of the information to be pushed is sufficient, the number of information pushing is more, when all the tags in the final tag set are used as crowd features, if there is a larger surplus in the number of information pushing, the information pushing can be performed according to the size of the history pushing data scale corresponding to each tag in the final tag set, the tag corresponding to the tag with the largest rule of the history pushing data is used as crowd features, if all the crowd with the tag with the largest rule of the history pushing data are pushed, there is still a surplus in number, and the pushing can be continued to the crowd with the tag with the second rank of the history pushing data scale, and so on until the information pushing with the set number is completed. And partial tags in the final tag set can be selected as crowd features to push information.
In another embodiment of the invention, the input cost of the information to be pushed is limited, the information pushing number is smaller, and all tags in the final tag set can be selected as crowd features to push the information, so that the crowd positioning accuracy of the limited number of information pushing is higher.
From the above, the crowd characteristics of the crowd corresponding to the information to be pushed are obtained by adopting a mode that based on the history push data set of the history push information which is the same as the class of the information to be pushed, the history push data is segmented in a plurality of dimensions by adopting a preset assessment index, the characteristic tag set is determined according to the segmentation result, the history push data corresponding to the characteristic tag set is further segmented according to the characteristic tag set, the history push data corresponding to the characteristic tag set is screened out, the assessment index is continuously adopted to segment the history push data corresponding to the characteristic tag set until a final tag set is obtained, and the crowd characteristics are obtained according to the final tag set. In addition, the crowd positioning accuracy of the information to be pushed and the information pushing effect can be expected.
To facilitate a better understanding and implementation of embodiments of the invention by those skilled in the art. The embodiment of the invention also provides a crowd characteristic determining device in information pushing.
Referring to fig. 2, a schematic structural diagram of a device for determining characteristics of people in information push in the implementation of the present invention is provided. The crowd characteristic determining device 20 in information push may include:
an acquisition unit 21 that acquires a history push data set of history push information identical to the category of information to be pushed;
the segmentation unit 22 is configured to segment the history push data set on each dimension by using a preset assessment index, so as to obtain a segmentation result, where the segmentation result includes: push data corresponding to each label in each dimension respectively, wherein the assessment index comprises: click rate or conversion rate;
a first determining unit 23 that determines a feature tag set including one or more feature tags according to the result of the segmentation;
the second determining unit 24 is adapted to segment the history push data set according to the feature tag set, screen out the history push data corresponding to the feature tag set, and continue to adopt the assessment index, and segment the history push data corresponding to the feature tag set in each dimension until a final tag set is obtained, where the final tag set is related to the feature tag set;
and the third determining unit 25 is adapted to determine, according to the final tag set, crowd characteristics of the crowd corresponding to the information to be pushed.
In a specific implementation, the working principle and flow of the in-information-pushing crowd feature determining device 20 may refer to the description of the in-information-pushing crowd feature determining method provided in any of the above embodiments of the present invention, which is not repeated herein.
The embodiment of the invention also provides a crowd characteristic determining device in information pushing, which comprises a memory and a processor, wherein the memory stores computer instructions capable of running on the processor, and the processor executes the steps of any crowd characteristic determining method in information pushing when running the computer instructions.
The embodiment of the invention also provides a computer readable storage medium, which is a nonvolatile storage medium or a non-transient storage medium, and is provided with computer instructions stored thereon, wherein the computer instructions execute the steps of any crowd characteristic determining method in information pushing when running.
Those of ordinary skill in the art will appreciate that all or a portion of the steps in the various methods of the above embodiments may be implemented by a program to instruct related hardware, the program may be stored in any computer readable storage medium, and the storage medium may include: ROM, RAM, magnetic or optical disks, etc.
Although the present invention is disclosed above, the present invention is not limited thereto. Various changes and modifications may be made by one skilled in the art without departing from the spirit and scope of the invention, and the scope of the invention should be assessed accordingly to that of the appended claims.
Claims (12)
1. The crowd characteristic determining method in information pushing is characterized by comprising the following steps:
acquiring a history push data set of history push information which is the same as the category of the information to be pushed;
and respectively segmenting the history pushing data set on each dimension by adopting a preset assessment index to obtain a segmentation result, wherein the segmentation result comprises the following steps of: push data corresponding to each label in each dimension respectively;
determining a feature tag set according to the segmentation result, wherein the feature tag set comprises one or more feature tags;
dividing the history push data set according to the characteristic tag set, screening out the history push data corresponding to the characteristic tag set, continuing to adopt the assessment index, and dividing the history push data corresponding to the characteristic tag set on each dimension respectively until a final tag set is obtained, wherein the final tag set is related to the characteristic tag set;
and determining crowd characteristics of the crowd corresponding to the information to be pushed according to the final tag set.
2. The method for determining the characteristics of the crowd in the information push of claim 1, wherein the step of dividing the history push data set in each dimension by using a preset assessment index to obtain a division result comprises the following steps:
traversing each label in the history push data in each dimension, determining push data corresponding to each label in each dimension, and calculating the information gain of each label; or,
and traversing each label in the historical push data in each dimension, determining the push data corresponding to each label in each dimension, and calculating the coefficient of the foundation corresponding to each label.
3. The method for determining the characteristics of the people in the information push according to claim 2, wherein the determining the characteristic tag set according to the segmentation result comprises:
determining a recommendation label according to the segmentation result;
and determining the characteristic label set according to the recommended labels.
4. The method for determining a group of people feature in information push according to claim 3, wherein determining a recommendation tag according to the segmentation result comprises:
and arranging the information gain of each label in order from big to small, and taking the label with the information gain of N before ranking as the recommended label, wherein N is more than or equal to 1 and is an integer.
5. The method for determining a group of people feature in information push according to claim 3, wherein determining a recommendation tag according to the segmentation result comprises:
and arranging the coefficient of the base of each label in order from small to large, and taking the label with the coefficient of the base ranked M as the recommended label, wherein M is more than or equal to 1 and is an integer.
6. The method for determining characteristics of people in information push according to claim 3, wherein determining a characteristic tag set according to the recommended tag includes:
selecting one or more of the recommended labels as characteristic labels and forming the characteristic label set;
or,
selecting one or more tags from other tags except the recommended tag as characteristic tags and forming the characteristic tag set; or,
and selecting one or more tags from the recommended tags and other tags except the recommended tags as the characteristic tags and forming the characteristic tag set.
7. The method for determining the crowd characteristics in information push according to claim 1, wherein the determining the crowd characteristics of the crowd corresponding to the information to be pushed according to the final tag set includes:
and determining crowd characteristics of the crowd corresponding to the information to be pushed according to the scale of the historical push data corresponding to each tag in the final tag set, wherein the crowd characteristics are part or all of the tags in the final tag set.
8. The method for determining the crowd characteristics in information push of claim 1, further comprising, after determining the crowd characteristics of the crowd corresponding to the information to be pushed:
and pushing the information to be pushed according to the crowd characteristics.
9. The method for determining the characteristics of people in information push as set forth in claim 1, wherein the assessment index comprises: click rate or conversion rate.
10. A crowd characteristic determining device in information push, comprising:
the acquisition unit acquires a history push data set of history push information which is the same as the category of the information to be pushed;
the segmentation unit is used for respectively segmenting the history pushing data set in each dimension by adopting a preset assessment index to obtain a segmentation result, wherein the segmentation result comprises the following steps of: push data corresponding to each label in each dimension respectively, wherein the assessment index comprises: click rate or conversion rate;
the first determining unit is used for determining a characteristic label set according to the segmentation result, wherein the characteristic label set comprises one or more characteristic labels;
the second determining unit is suitable for cutting the history push data set according to the characteristic tag set, screening out the history push data corresponding to the characteristic tag set, continuing to adopt the assessment index, and respectively cutting the history push data corresponding to the characteristic tag set in each dimension until a final tag set is obtained, wherein the final tag set is related to the characteristic tag set;
and the third determining unit is suitable for determining crowd characteristics of the crowd corresponding to the information to be pushed according to the final tag set.
11. An in-flight profile feature determination apparatus comprising a memory and a processor, the memory having stored thereon computer instructions executable on the processor, wherein the processor, when executing the computer instructions, performs the steps of the in-flight profile feature determination method of any one of claims 1 to 9.
12. A computer readable storage medium, the computer readable storage medium being a non-volatile storage medium or a non-transitory storage medium, having stored thereon computer instructions, wherein the computer instructions, when executed, perform the steps of the method for determining characteristics of a group of people in information push according to any of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010461899.2A CN111737305B (en) | 2020-05-27 | 2020-05-27 | Method and device for determining characteristics of people group in information push and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010461899.2A CN111737305B (en) | 2020-05-27 | 2020-05-27 | Method and device for determining characteristics of people group in information push and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111737305A CN111737305A (en) | 2020-10-02 |
| CN111737305B true CN111737305B (en) | 2024-03-15 |
Family
ID=72647885
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010461899.2A Active CN111737305B (en) | 2020-05-27 | 2020-05-27 | Method and device for determining characteristics of people group in information push and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111737305B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105989004A (en) * | 2015-01-27 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Information releasing pretreatment method and device |
| CN108629045A (en) * | 2018-05-14 | 2018-10-09 | 平安科技（深圳）有限公司 | A kind of information push method, storage medium and server |
| CN109615487A (en) * | 2019-01-04 | 2019-04-12 | 平安科技（深圳）有限公司 | Products Show method, apparatus, equipment and storage medium based on user behavior |
| CN110263242A (en) * | 2019-01-04 | 2019-09-20 | 腾讯科技（深圳）有限公司 | Content recommendation method, device, computer readable storage medium and computer equipment |
| CN111061979A (en) * | 2018-10-17 | 2020-04-24 | 北京字节跳动网络技术有限公司 | User label pushing method and device, electronic equipment and medium |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5087721B2 (en) * | 2010-05-31 | 2012-12-05 | 楽天株式会社 | Database management apparatus, database management method, database management program, and computer-readable recording medium storing the program |
| JP6516849B2 (en) * | 2015-09-07 | 2019-05-22 | 富士フイルム株式会社 | Search apparatus, search system and search method |
-
2020
- 2020-05-27 CN CN202010461899.2A patent/CN111737305B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105989004A (en) * | 2015-01-27 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Information releasing pretreatment method and device |
| CN108629045A (en) * | 2018-05-14 | 2018-10-09 | 平安科技（深圳）有限公司 | A kind of information push method, storage medium and server |
| CN111061979A (en) * | 2018-10-17 | 2020-04-24 | 北京字节跳动网络技术有限公司 | User label pushing method and device, electronic equipment and medium |
| CN109615487A (en) * | 2019-01-04 | 2019-04-12 | 平安科技（深圳）有限公司 | Products Show method, apparatus, equipment and storage medium based on user behavior |
| CN110263242A (en) * | 2019-01-04 | 2019-09-20 | 腾讯科技（深圳）有限公司 | Content recommendation method, device, computer readable storage medium and computer equipment |
Non-Patent Citations (3)
| Title |
|---|
| The MovieLens Datasets: History and Context;F. Maxwell Harper等;《ACM Transactions on Interactive Intelligent Systems》;第5卷(第4期);1-19 * |
| 基于商品属性的电子商务推荐系统研究;胡新明;《中国博士学位论文全文数据库 (经济与管理科学辑)》(第07期);J157-8 * |
| 基于类别的推荐——一种解决协同推荐中冷启动问题的方法;罗喜军等;《第二十四届中国数据库学术会议论文集（研究报告篇）》;290-295 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111737305A (en) | 2020-10-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10324960B1 (en) | Determining a number of unique viewers of a content item | |
| US10580035B2 (en) | Promotion selection for online customers using Bayesian bandits | |
| WO2016161976A1 (en) | Method and device for selecting data content to be pushed to terminals | |
| CN110489644A (en) | Information-pushing method, device, computer readable storage medium and computer equipment | |
| CN104834641B (en) | The processing method and related system of network media information | |
| CN105930425A (en) | Personalized video recommendation method and apparatus | |
| CN111382361B (en) | Information pushing method, device, storage medium and computer equipment | |
| CN108777701B (en) | Method and device for determining information audience | |
| US20080228537A1 (en) | Systems and methods for targeting advertisements to users of social-networking and other web 2.0 websites and applications | |
| CN103338223A (en) | Recommendation method, client and server of mobile application | |
| US11593860B2 (en) | Method, medium, and system for utilizing item-level importance sampling models for digital content selection policies | |
| KR20210006505A (en) | Methods and apparatus to estimate demographics of users employing social media | |
| CN104915734A (en) | Commodity popularity prediction method based on time sequence and system thereof | |
| CN103295145A (en) | Mobile phone advertising method based on user consumption feature vector | |
| CN110753920A (en) | System and method for optimizing and simulating web page ranking and traffic | |
| CN111639988B (en) | Broker recommendation method, device, electronic equipment and storage medium | |
| CN110880127B (en) | Consumption level prediction method and device, electronic equipment and storage medium | |
| CN105488163B (en) | Information-pushing method and device | |
| CN109949089B (en) | Method, device and terminal for determining display rate | |
| EP3537365A1 (en) | Method, device, and system for increasing users | |
| CN105824818A (en) | Informationized management method, platform and system | |
| CN109598544A (en) | A kind of media object matching process and system | |
| CN106713651A (en) | Information display method and mobile terminal | |
| EP3818488A2 (en) | Availability ranking system and method | |
| CN104735117A (en) | Information processing method, device and system |
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 |