CN107229557B - Abnormal click detection method and device and click quantity statistical method and device - Google Patents
Abnormal click detection method and device and click quantity statistical method and device Download PDFInfo
- Publication number
- CN107229557B CN107229557B CN201710492409.3A CN201710492409A CN107229557B CN 107229557 B CN107229557 B CN 107229557B CN 201710492409 A CN201710492409 A CN 201710492409A CN 107229557 B CN107229557 B CN 107229557B
- Authority
- CN
- China
- Prior art keywords
- click
- detection period
- detected user
- clicks
- user
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3452—Performance evaluation by statistical analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- 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
- G06Q30/0241—Advertisements
- G06Q30/0248—Avoiding fraud
-
- 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
- G06Q30/0241—Advertisements
- G06Q30/0277—Online advertisement
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- General Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Databases & Information Systems (AREA)
- General Business, Economics & Management (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Bioinformatics & Computational Biology (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Data Mining & Analysis (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to an abnormal click detection method and device, and belongs to the technical field of information processing. The method of the invention is based on the user angle, the click behavior of a specific user within a period of time is compared and analyzed with the click behavior of the similar user group where the specific user is located, so that the user with larger difference with the click behavior of the whole similar user group is found out, and the click behavior sent by the user within the detection period is judged as the abnormal click behavior. The invention also discloses a click rate statistical method and a click rate statistical device. Compared with the prior art, the method has the advantages of more accurate detection result, simpler detection process and better detection effect on abnormal click behaviors from legal users.
Description
Technical Field
The invention relates to an abnormal click detection method and device, and belongs to the technical field of information processing.
Background
Today, when technologies such as computers and the internet are rapidly developed, the click rate of programs or websites has important reference significance for advertisement publishers, content providers and the like. However, the current false click rate, such as the brushing rate of the navy, makes the counted click rate inconsistent with the actual click rate. The counterfeiting of the click behavior causes deviation between the estimation and the practice, and the evaluation of advertisers and investors is watered, which has a serious influence on the whole industry. It is therefore necessary to effectively identify click behavior to provide true and effective click volume statistics.
In the conventional technology for preventing cheating on click rate, from the perspective of a clicked object (for example, a commodity in an e-commerce website, an advertisement in a webpage, an audio/video provided by a multimedia on-demand system, etc.), whether click behavior of a certain object is abnormal or not is determined by comparing the clicked amount of the object with historical click data of the object. The existing detection mode usually aims at a certain specific webpage or commodity, the behavior and the discrimination standard of the existing detection mode cannot be directly copied to other click objects, so that different parameter standards need to be formulated for each specific webpage or at least each specific type of webpage; on one hand, the algorithm of cheating detection is complex, the requirements on software and hardware are too high, and on the other hand, the detection accuracy is also low. It is therefore necessary to explore new click cheating detection techniques from other perspectives.
Disclosure of Invention
The technical problem to be solved by the invention is to overcome the defects of the prior art and provide the abnormal click detection method and the abnormal click detection device, so that the abnormal click is detected from the perspective of the user and based on the click behavior characteristics of the similar user group, the detection result is more accurate, and the detection process is simpler.
The abnormal click detection method comprises the following steps:
step 1, extracting click behavior characteristics of a detected user in a detection period from click behavior statistical data of the detected user in the detection period; extracting click behavior characteristics of the similar user group in the detection period from click behavior statistical data of the similar user group in which the detected user is located in the detection period;
step 2, calculating the difference between the click behavior characteristics of the detected user in the detection period and the first standard click behavior characteristics, and the difference between the click behavior characteristics of the similar user group in the detection period and the second standard click behavior characteristics;
step 3, judging whether the difference value between the two difference degrees exceeds a preset range, if so, judging that the click behaviors of the detected user in the detection period are abnormal clicks; otherwise, judging that the click behaviors of the detected user in the detection period are normal clicks.
Preferably, the click behavior characteristics of the detected user in the detection period are the time distribution of the number of clicks of the detected user in the detection period; and the click behavior characteristics of the similar user group in the detection period are the time distribution mean value of the number of clicks of each user in the similar user group in the detection period.
Or the click behavior characteristics of the detected user in the detection period are the time distribution of the number of clicks of the detected user in the detection period; and the click behavior characteristics of the similar user group in the detection period are the time distribution mean value of the number of clicks of all users except the detected user in the similar user group in the detection period.
Preferably, only click behaviors that can determine the user information are counted in the click behavior statistical data.
Preferably, the first standard click behavior characteristic is equal to the second standard click behavior characteristic.
Preferably, the first standard click behavior feature and/or the second standard click behavior feature is/are the average value of click behavior features of a similar user group where the detected user is located in a plurality of previous detection periods.
Preferably, the degree of difference is the inverse of the degree of similarity between the two features.
In order to further improve the detection accuracy, if the click behaviors of the detected user in the detection period are all judged to be abnormal clicks in step 3, the click behaviors are further judged according to the following method:
step 4, for each object clicked by the detected user in the detection period, respectively obtaining the number of times of clicking the object by the similar user group in which the detected user is located in the detection period, taking the maximum value of the number of times of clicking the object by the person, multiplying the maximum value by a preset coefficient which is more than or equal to 1, and taking the obtained product as a standard number of clicks;
and 5, judging whether the number of clicks of the detected user on the object in the detection period is smaller than the standard number of clicks or not for each object clicked by the detected user in the detection period, and if so, rejecting all clicks of the detected user on the object in the detection period from the abnormal clicks judged in the step 3.
The value range of the coefficient is preferably (1, 2).
The abnormal click detection device of the present invention includes:
the characteristic extraction module is used for extracting the click behavior characteristics of the detected user in the detection period from the click behavior statistical data of the detected user in the detection period; extracting click behavior characteristics of the similar user group in the detection period from click behavior statistical data of the similar user group in which the detected user is located in the detection period;
the difference degree calculation module is used for calculating the difference degree between the click behavior characteristic of the detected user in the detection period and a first standard click behavior characteristic, and the difference degree between the click behavior characteristic of the similar user group in the detection period and a second standard click behavior characteristic;
the judging module is used for judging whether the difference value between the two difference degrees exceeds a preset range, and if so, judging that the clicking behaviors of the detected user in the detection period are abnormal clicks; otherwise, judging that the click behaviors of the detected user in the detection period are normal clicks.
Preferably, the click behavior characteristics of the detected user in the detection period are the time distribution of the number of clicks of the detected user in the detection period; and the click behavior characteristics of the similar user group in the detection period are the time distribution mean value of the number of clicks of each user in the similar user group in the detection period.
Or the click behavior characteristics of the detected user in the detection period are the time distribution of the number of clicks of the detected user in the detection period; and the click behavior characteristics of the similar user group in the detection period are the time distribution mean value of the number of clicks of all users except the detected user in the similar user group in the detection period.
Preferably, only click behaviors that can determine the user information are counted in the click behavior statistical data.
Preferably, the first standard click behavior characteristic is equal to the second standard click behavior characteristic.
Preferably, the first standard click behavior feature and/or the second standard click behavior feature is/are the average value of click behavior features of a similar user group where the detected user is located in a plurality of previous detection periods.
Preferably, the degree of difference is the inverse of the degree of similarity between the two features.
In order to further improve the detection accuracy, the device further comprises:
the detection result correction module is used for further judging the abnormal click output by the judgment module and comprises a standard click number calculation submodule and a correction submodule; the standard click number calculation submodule is used for respectively acquiring the number of times of clicking on each object clicked by the detected user in the detection period by the similar user group where the detected user is located in the detection period, taking the maximum value of the number of times of clicking on the object by the person, multiplying the maximum value by a preset coefficient which is more than or equal to 1, and taking the obtained product as the standard click number; the correction submodule is used for judging whether the number of clicks of the detected user on the object in the detection period is smaller than the standard number of clicks or not for each object clicked by the detected user in the detection period, if so, all the clicks of the detected user on the object in the detection period are removed from the abnormal clicks output by the judgment module.
Preferably, the value range of the coefficient is (1, 2).
The following technical scheme can be obtained according to the same invention concept:
a click rate statistical method comprises the steps of firstly, recording all click behaviors; then, carrying out abnormal click detection by using the method of any one of the technical schemes; and finally, removing the detected abnormal click from all the recorded click behaviors and counting the click quantity of the rest click behaviors.
A click volume statistics apparatus, comprising:
the recording unit is used for recording all click behaviors;
the abnormal click detection device according to any one of the above technical solutions is configured to perform abnormal click detection;
and the click rate counting unit is used for removing the abnormal click detected by the abnormal click detection device from all the click behaviors recorded by the recording unit and counting the click rate of the rest click behaviors.
Compared with the prior art, the invention has the following beneficial effects:
the method and the device provided by the invention are used for detecting the abnormal click from the perspective of the user and based on the click behavior characteristics of the similar user group, so that the detection result is more accurate, and the detection process is simpler.
Detailed Description
Aiming at the defects of the prior art, the method and the device have the idea that from the perspective of a user, the clicking behavior of a specific user within a period of time is compared and analyzed with the clicking behavior of a similar user group where the specific user is located, so that the user with the larger difference with the whole clicking behavior of the similar user group is found out, and the clicking behavior sent by the user within a detection period is judged to be the abnormal clicking behavior. The method has the advantages of more accurate detection result, simpler detection process and better detection effect on abnormal click behaviors (such as the bill brushing or flow brushing behaviors of registered users) from legal users.
The abnormal click detection device of the present invention includes:
the characteristic extraction module is used for extracting the click behavior characteristics of the detected user in the detection period from the click behavior statistical data of the detected user in the detection period; extracting click behavior characteristics of the similar user group in the detection period from click behavior statistical data of the similar user group in which the detected user is located in the detection period;
the difference degree calculation module is used for calculating the difference degree between the click behavior characteristic of the detected user in the detection period and a first standard click behavior characteristic, and the difference degree between the click behavior characteristic of the similar user group in the detection period and a second standard click behavior characteristic;
the judging module is used for judging whether the difference value between the two difference degrees exceeds a preset range, and if so, judging that the clicking behaviors of the detected user in the detection period are abnormal clicks; otherwise, judging that the click behaviors of the detected user in the detection period are normal clicks.
To facilitate understanding of the public, the technical solution of the present invention will be described in further detail below.
The abnormal click detection method comprises the following steps:
step 1, extracting click behavior characteristics of a detected user in a detection period from click behavior statistical data of the detected user in the detection period; and extracting the click behavior characteristics of the similar user group in the detection period from the click behavior statistical data of the similar user group in which the detected user is positioned in the detection period.
The method of the invention is used for detecting each user. The statistical data of the click behaviors of each user in any time period can be counted according to the click behaviors recorded by the system. The specific detection period can be set according to actual needs, such as a week, a month, a quarter, and the like. In order to improve the detection accuracy, the click behavior statistical data only counts the click behaviors of the determinable user information, and the click behaviors of the user information cannot be determined and are not counted into the click behavior statistical data.
The similar user group is that users with similar preferences or similar behavior patterns are classified into the same user group according to a certain preset standard, and users with larger preferences or behavior differences are classified into different user groups. The grouping of similar users has great significance for enhancing the pertinence of services and improving the user experience, so that the method is widely applied to various aspects such as electronic commerce, multimedia online on-demand and the like. The grouping method of similar users is usually implemented by a clustering algorithm based on some user similarity measure (such as cosine similarity, pearson coefficient, adjusted cosine similarity, euclidean distance, etc. which are most commonly used at present). Which is a well-established technology, and is not described herein for brevity.
The users in the similar user group can show consistency to the click behaviors of objects such as movie works, commodities, advertisements and the like. To achieve a comparison of click behavior, click behavior is first abstracted into comparable behavior features. The specific click behavior characteristics may be distribution of click behaviors in the dimension of the object, distribution of click behaviors in the time dimension, or a combination of the two. For example, for the multimedia online on-demand service, the click number distribution of the user on different types of videos (terrorism, action, thrillery, love, ethics and the like) in the detection period can be used as the click behavior feature of the user in the detection period, the click number distribution of the user on each period in the detection period can be used as the click behavior feature of the user in the detection period, and the click number distribution of the user on different types of videos (terrorism, action, thrillery, love, ethics and the like) in each period in the detection period can be used as the click behavior feature of the user in the detection period. The concrete expression of the click behavior characteristics can be in the forms of curves, graphs, matrixes, vectors and the like.
Considering the accuracy and complexity of the algorithm comprehensively, the invention preferably uses the time domain distribution of the click behavior as the click behavior feature, specifically: the click behavior characteristics of the detected user in the detection period are the time distribution of the number of clicks of the detected user in the detection period; the click behavior characteristics of the similar user group in the detection period are the time distribution mean value of the number of clicks of each user in the similar user group in the detection period, or the time distribution mean value of the number of clicks of each user except the detected user in the similar user group in the detection period.
And 2, calculating the difference between the click behavior characteristic of the detected user in the detection period and a first standard click behavior characteristic, and the difference between the click behavior characteristic of the similar user group in the detection period and a second standard click behavior characteristic.
After the click behavior characteristics of the detected user in the detection period and the click behavior characteristics of the similar user groups in the detection period are obtained, the difference between the click behavior characteristics of the detected user in the detection period and the click behavior characteristics of the similar user groups in the detection period can be used for judging that the click behavior of the detected user in the detection period deviates from the normal behaviors of most users. However, considering the uncertainty of the time period, such as the specific holiday, the promotion activity, the important collective activity, etc., which may cause the situation that the difference occurs in the period repetition, the present invention does not directly compare the two, but calculates the difference between the click behavior feature of the detected user in the detection period and the first standard click behavior feature, and the difference between the click behavior feature of the similar user group in the detection period and the second standard click behavior feature.
As two preset comparison criteria, the first standard click behavior feature and the second standard click behavior feature may be the same or different. For example, the first standard click behavior feature may be an average of click behavior features of the detected user in a plurality of previous detection periods in which the abnormal click behavior is not detected, that is, the first standard click behavior feature is generated by using the historical data of the user; similarly, the second standard click behavior feature may be an average of click behavior features of a group of similar users in which the detected user is located within a previous number of detection cycles. The preferred scheme of the invention is as follows: the first standard click behavior characteristic and/or the second standard click behavior characteristic are/is the average value of click behavior characteristics of a similar user group where the detected user is located in a plurality of previous detection periods. At least one of the two uses the historical data of the similar user group as the comparison basis.
There are many specific measures of the degree of difference, most conveniently in the form of the inverse of the degree of similarity, and other measures that vary inversely with the degree of similarity may be used. The similarity measure between two feature vectors has various ways, such as distance (e.g., euclidean distance, chebyshev distance, manhattan distance, etc.), pearson correlation coefficient, mutual information entropy, and the like.
Step 3, judging whether the difference value between the two difference degrees exceeds a preset range, if so, judging that the click behaviors of the detected user in the detection period are abnormal clicks; otherwise, judging that the click behaviors of the detected user in the detection period are normal clicks.
Assuming that the calculated difference degree between the click behavior feature of the detected user in the detection period and the first standard click behavior feature is X1, the difference degree between the click behavior feature of the similar user group in which the detected user is located in the detection period and the second standard click behavior feature is X2. If the difference value between the X1 and the X2 exceeds a preset range, the difference between the click behavior of the detected user in the detection period and the overall click behavior of the similar user group in which the detected user is located in the detection period is over large, and the click behaviors of the detected user in the detection period can be judged to be abnormal clicks; otherwise, the click is determined to be a normal click.
The click behaviors of the detected user in the detection period can be preliminarily detected through the three steps, and the click behaviors of all the users in the detection period are detected in the same mode, so that the detection results of all the click behaviors in the detection period are obtained.
According to the scheme, all clicking behaviors of the detected user with the difference degree exceeding the preset range in the detection period are judged to be abnormal clicking in the final judgment, although the method is simple, the user is considered to possibly click a plurality of different objects in the detection period, the possibility that the user clicks part of the objects normally, and only clicks a small number of specific objects are abnormal exists, and thus partial misjudgment is inevitable. Therefore, in order to solve the above problems and further improve the detection accuracy, the present invention provides a dimension of an object, and performs a second detection on the click behavior determined to be abnormal, specifically:
and 4, respectively acquiring the number of times of clicking the object by the similar user group in which the detected user is located in the detection period for each object clicked by the detected user in the detection period, taking the maximum value of the number of times of clicking the object by the person, multiplying the maximum value by a preset coefficient which is more than or equal to 1, and taking the obtained product as the standard number of clicks.
Assuming that the detected user clicks the N objects for M times in the detection period, the distribution of the M clicks on the N objects can be obtained according to the click record. For each of the N specific objects, acquiring the number of clicks of the similar user group where the detected user is located on the object in the detection period, namely dividing the total number of clicks of the similar user group on the object in the detection period by the number of the similar user group, or dividing the total number of clicks of all users except the detected user in the similar user group on the object in the detection period by the total number of clicks of the similar user group except the detected user; then, the maximum value is found out from the number of clicks of the N persons, the maximum value is multiplied by a preset coefficient (the value range is preferably (1, 2)) which is larger than or equal to 1, and the product obtained finally serves as the standard number of clicks and serves as a judgment reference for further judgment.
And 5, judging whether the number of clicks of the detected user on the object in the detection period is smaller than the standard number of clicks or not for each object clicked by the detected user in the detection period, and if so, rejecting all clicks of the detected user on the object in the detection period from the abnormal clicks judged in the step 3.
Comparing the number of clicks of each object by the detected user in the detection period with the standard number of clicks, if the number of clicks is smaller than the standard number of clicks, considering that the clicking behavior of the detected user on the object in the detection period belongs to normal clicking behavior, and removing the normal clicking behavior from the detected abnormal clicking; and if the number of clicks is equal to or larger than the standard number of clicks, the click behavior of the detected user on the object in the detection period is considered to be abnormal, and the abnormal click behavior is output as the final abnormal click.
Correspondingly, the abnormal click detection device of the present invention may further include:
the detection result correction module is used for further judging the abnormal click output by the judgment module and comprises a standard click number calculation submodule and a correction submodule; the standard click number calculation submodule is used for respectively acquiring the number of times of clicking on each object clicked by the detected user in the detection period by the similar user group where the detected user is located in the detection period, taking the maximum value of the number of times of clicking on the object by the person, multiplying the maximum value by a preset coefficient which is more than or equal to 1, and taking the obtained product as the standard click number; the correction submodule is used for judging whether the number of clicks of the detected user on the object in the detection period is smaller than the standard number of clicks or not for each object clicked by the detected user in the detection period, if so, all the clicks of the detected user on the object in the detection period are removed from the abnormal clicks output by the judgment module.
The abnormal click detection scheme can be applied to click quantity statistics in the aspects of electronic commerce, multimedia on-line on-demand, webpage advertisements and the like, so that the accuracy of click quantity data is effectively improved, and the click quantity statistics method specifically comprises the following steps: firstly, recording all click behaviors; then, carrying out abnormal click detection by using the method; and finally, removing the detected abnormal click from all the recorded click behaviors and counting the click quantity of the rest click behaviors.
Similarly, the click rate statistic device of the invention comprises:
the recording unit is used for recording all click behaviors;
the abnormal click detection device is used for detecting abnormal clicks;
and the click rate counting unit is used for removing the abnormal click detected by the abnormal click detection device from all the click behaviors recorded by the recording unit and counting the click rate of the rest click behaviors.
In addition, when the final click rate is counted, the number of detected abnormal clicks can be replaced or smoothly corrected by using the normal click history data of the detected user or the click history data of the similar user group where the detected user is located or the current click data.
Claims (20)
1. An abnormal click detection method is characterized by comprising the following steps:
step 1, extracting click behavior characteristics of a detected user in a detection period from click behavior statistical data of the detected user in the detection period; extracting click behavior characteristics of the similar user group in the detection period from click behavior statistical data of the similar user group in which the detected user is located in the detection period;
step 2, calculating a first difference between the click behavior characteristics of the detected user in the detection period and a first standard click behavior characteristic, and a second difference between the click behavior characteristics of the similar user group in the detection period and a second standard click behavior characteristic;
step 3, judging whether the difference value between the first difference degree and the second difference degree exceeds a preset range, if so, judging that the click behaviors of the detected user in the detection period are abnormal clicks; otherwise, judging that the click behaviors of the detected user in the detection period are normal clicks.
2. The method of claim 1, wherein the click behavior of the detected user in the detection period is characterized by a time distribution of the number of clicks of the detected user in the detection period; and the click behavior characteristics of the similar user group in the detection period are the time distribution mean value of the number of clicks of each user in the similar user group in the detection period.
3. The method of claim 1, wherein the click behavior of the detected user in the detection period is characterized by a time distribution of the number of clicks of the detected user in the detection period; and the click behavior characteristics of the similar user group in the detection period are the time distribution mean value of the number of clicks of all users except the detected user in the similar user group in the detection period.
4. The method of claim 1, wherein the click behavior statistics only count click behaviors for which the user information is determinable.
5. The method of claim 1, wherein the first standard click behavior characteristic is equal to the second standard click behavior characteristic.
6. The method of claim 1, wherein the first standard click behavior feature and/or the second standard click behavior feature is an average of click behavior features of a group of similar users in which the detected user is located within a previous number of detection cycles.
7. The method of claim 1, wherein the degree of difference is an inverse of a degree of similarity between two features.
8. The method according to any one of claims 1 to 7, wherein if the click behaviors of the detected user in the detection period are all judged to be abnormal clicks in step 3, the click behaviors are further judged according to the following method:
step 4, for each object clicked by the detected user in the detection period, respectively obtaining the number of times of clicking the object by the similar user group in which the detected user is located in the detection period, taking the maximum value of the number of times of clicking the object by the person, multiplying the maximum value by a preset coefficient which is more than or equal to 1, and taking the obtained product as a standard number of clicks;
and 5, judging whether the number of clicks of the detected user on the object in the detection period is smaller than the standard number of clicks or not for each object clicked by the detected user in the detection period, and if so, rejecting all clicks of the detected user on the object in the detection period from the abnormal clicks judged in the step 3.
9. The method of claim 8, wherein the coefficient has a value in the range of (1, 2).
10. An abnormal click detection apparatus, comprising:
the characteristic extraction module is used for extracting the click behavior characteristics of the detected user in the detection period from the click behavior statistical data of the detected user in the detection period; extracting click behavior characteristics of the similar user group in the detection period from click behavior statistical data of the similar user group in which the detected user is located in the detection period;
the difference degree calculation module is used for calculating a first difference degree between the click behavior feature of the detected user in the detection period and a first standard click behavior feature and a second difference degree between the click behavior feature of the similar user group in the detection period and a second standard click behavior feature;
the judging module is used for judging whether the difference value between the first difference degree and the second difference degree exceeds a preset range, and if so, judging that the click behaviors of the detected user in the detection period are abnormal clicks; otherwise, judging that the click behaviors of the detected user in the detection period are normal clicks.
11. The apparatus of claim 10, wherein the click behavior of the detected user in the detection period is characterized by a time distribution of the number of clicks of the detected user in the detection period; and the click behavior characteristics of the similar user group in the detection period are the time distribution mean value of the number of clicks of each user in the similar user group in the detection period.
12. The apparatus of claim 10, wherein the click behavior of the detected user in the detection period is characterized by a time distribution of the number of clicks of the detected user in the detection period; and the click behavior characteristics of the similar user group in the detection period are the time distribution mean value of the number of clicks of all users except the detected user in the similar user group in the detection period.
13. The apparatus of claim 10, wherein the click behavior statistics only count click behaviors for which the user information is determinable.
14. The apparatus of claim 10, wherein the first standard click behavior characteristic is equal to the second standard click behavior characteristic.
15. The apparatus of claim 10, wherein the first standard click behavior feature and/or the second standard click behavior feature is an average of click behavior features of a group of similar users in which the detected user is located within a previous number of detection cycles.
16. The apparatus of claim 10, wherein the degree of difference is an inverse of a degree of similarity between two features.
17. The apparatus of any one of claims 10 to 16, further comprising:
the detection result correction module is used for further judging the abnormal click output by the judgment module and comprises a standard click number calculation submodule and a correction submodule; the standard click number calculation submodule is used for respectively acquiring the number of times of clicking on each object clicked by the detected user in the detection period by the similar user group where the detected user is located in the detection period, taking the maximum value of the number of times of clicking on the object by the person, multiplying the maximum value by a preset coefficient which is more than or equal to 1, and taking the obtained product as the standard click number; the correction submodule is used for judging whether the number of clicks of the detected user on the object in the detection period is smaller than the standard number of clicks or not for each object clicked by the detected user in the detection period, if so, all the clicks of the detected user on the object in the detection period are removed from the abnormal clicks output by the judgment module.
18. The apparatus of claim 17, wherein the coefficient has a value in a range of (1, 2).
19. A click rate statistical method is characterized in that firstly, all click behaviors are recorded; then carrying out abnormal click detection by using the method according to any one of claims 1 to 9; and finally, removing the detected abnormal click from all the recorded click behaviors and counting the click quantity of the rest click behaviors.
20. A click rate statistic device, comprising:
the recording unit is used for recording all click behaviors;
the abnormal click detection device according to any one of claims 10 to 18, configured to perform abnormal click detection;
and the click rate counting unit is used for removing the abnormal click detected by the abnormal click detection device from all the click behaviors recorded by the recording unit and counting the click rate of the rest click behaviors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710492409.3A CN107229557B (en) | 2017-06-26 | 2017-06-26 | Abnormal click detection method and device and click quantity statistical method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710492409.3A CN107229557B (en) | 2017-06-26 | 2017-06-26 | Abnormal click detection method and device and click quantity statistical method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107229557A CN107229557A (en) | 2017-10-03 |
| CN107229557B true CN107229557B (en) | 2020-10-20 |
Family
ID=59935722
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710492409.3A Active CN107229557B (en) | 2017-06-26 | 2017-06-26 | Abnormal click detection method and device and click quantity statistical method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107229557B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107748998A (en) * | 2017-10-13 | 2018-03-02 | 北京奇艺世纪科技有限公司 | The method and system that advertisement is recommended |
| CN108156141B (en) * | 2017-12-14 | 2021-08-27 | 北京奇艺世纪科技有限公司 | Real-time data identification method and device and electronic equipment |
| CN109241994A (en) * | 2018-07-31 | 2019-01-18 | 顺丰科技有限公司 | A kind of user's anomaly detection method, device, equipment and storage medium |
| CN109388548B (en) * | 2018-09-29 | 2020-12-22 | 京东数字科技控股有限公司 | Method and apparatus for generating information |
| CN109447701A (en) * | 2018-10-24 | 2019-03-08 | 麒麟合盛网络技术股份有限公司 | The anti-cheat method of application program, device and server-side |
| CN111292139A (en) * | 2020-03-12 | 2020-06-16 | 上海数川数据科技有限公司 | Anti-cheating method based on time-frequency analysis |
| CN112633766A (en) * | 2020-12-30 | 2021-04-09 | 苏州三六零智能安全科技有限公司 | Channel brushing amount judgment method, device, equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103853839A (en) * | 2014-03-18 | 2014-06-11 | 北京博雅立方科技有限公司 | Method and device for evaluating advertisement page malicious click suspected degree |
| CN104050178A (en) * | 2013-03-13 | 2014-09-17 | 北京思博途信息技术有限公司 | Internet monitoring anti-spamming method and device |
| CN104765874A (en) * | 2015-04-24 | 2015-07-08 | 百度在线网络技术（北京）有限公司 | Method and device for detecting click-cheating |
| CN106355431A (en) * | 2016-08-18 | 2017-01-25 | 晶赞广告（上海）有限公司 | Detection method, device and terminal for cheating traffic |
| CN106651423A (en) * | 2016-09-28 | 2017-05-10 | 华南理工大学 | Quality scoring method for mobile application advertisement putting |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8135615B2 (en) * | 2007-12-18 | 2012-03-13 | Amdocs Software Systems Limited | Systems and methods for detecting click fraud |
-
2017
- 2017-06-26 CN CN201710492409.3A patent/CN107229557B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104050178A (en) * | 2013-03-13 | 2014-09-17 | 北京思博途信息技术有限公司 | Internet monitoring anti-spamming method and device |
| CN103853839A (en) * | 2014-03-18 | 2014-06-11 | 北京博雅立方科技有限公司 | Method and device for evaluating advertisement page malicious click suspected degree |
| CN104765874A (en) * | 2015-04-24 | 2015-07-08 | 百度在线网络技术（北京）有限公司 | Method and device for detecting click-cheating |
| CN106355431A (en) * | 2016-08-18 | 2017-01-25 | 晶赞广告（上海）有限公司 | Detection method, device and terminal for cheating traffic |
| CN106651423A (en) * | 2016-09-28 | 2017-05-10 | 华南理工大学 | Quality scoring method for mobile application advertisement putting |
Non-Patent Citations (1)
| Title |
|---|
| 网络广告中的点击欺诈检测方法研究;乔晨蕊等;《万方数据库》;20170523;第1-62页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107229557A (en) | 2017-10-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107229557B (en) | Abnormal click detection method and device and click quantity statistical method and device | |
| CN106097000B (en) | Information processing method and server | |
| CN106355431B (en) | Cheating flow detection method and device and terminal | |
| CA3013711C (en) | Systems, methods, and articles of manufacture to measure online audiences | |
| AU2012231667B2 (en) | Methods and apparatus to determine media impressions | |
| TWI391867B (en) | Method for scoring user click and click traffic scoring system thereof | |
| CN110009372B (en) | User risk identification method and device | |
| CN105808639B (en) | Network access behavior identification method and device | |
| CN108780479A (en) | For to the abnormal system and method for being detected and scoring | |
| CN110290400B (en) | Suspicious brushing amount video identification method, real playing amount estimation method and device | |
| CN107493277A (en) | The online method for detecting abnormality of big data platform based on maximum information coefficient | |
| CN103327016B (en) | A kind of computing network Streaming Media exception playback volume the method and system to its correction | |
| US20140324522A1 (en) | Detecting Fraud In Internet-Based Lead Generation Utilizing Neural Networks | |
| US10348844B2 (en) | Method and device for monitoring push effect of push information | |
| CN109726737B (en) | Track-based abnormal behavior detection method and device | |
| CN108768743B (en) | User identification method and device and server | |
| CN114626024A (en) | Internet infringement video low-consumption detection method and system based on block chain | |
| JP6792694B1 (en) | Equipment, methods and programs for evaluating the effects of TVCM | |
| CN111191720A (en) | Service scene identification method and device and electronic equipment | |
| CN111612531B (en) | Click fraud detection method and system | |
| CN112819476A (en) | Risk identification method and device, nonvolatile storage medium and processor | |
| KR102172532B1 (en) | Method and system for calculating price influence based on disclosure and media | |
| CN114493784A (en) | Commodity recommendation method and recommendation system based on big data | |
| CN113220741A (en) | Internet advertisement false flow identification method, system, equipment and storage medium | |
| CN110113228B (en) | Network connection detection method and device |
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 |