CN109858977B - User behavior analysis method and system based on self-coding - Google Patents
User behavior analysis method and system based on self-coding Download PDFInfo
- Publication number
- CN109858977B CN109858977B CN201910142937.5A CN201910142937A CN109858977B CN 109858977 B CN109858977 B CN 109858977B CN 201910142937 A CN201910142937 A CN 201910142937A CN 109858977 B CN109858977 B CN 109858977B
- Authority
- CN
- China
- Prior art keywords
- self
- application
- behavior
- user
- coding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention discloses a user behavior analysis method and a system based on self-coding, which relate to the technical field of information coding and comprise the following steps: acquiring user behavior data based on a self-coding mode to obtain self-coding sequences corresponding to different user behaviors; obtaining all possible user behavior data according to a self-coding mode, wherein the possible user behavior data comprise preset standard sequences corresponding to all possible behaviors; and comparing the self-coding sequence with a preset standard sequence, and performing defect detection and behavior detection of the user. The self-coding sequence is compared with a preset standard sequence, so that behavior detection of the user can be performed, data is not required to be obtained through buried points, the accuracy of user behavior acquisition is improved in a mode that the self-coding sequence is compared with the preset standard sequence, and therefore the accuracy of user behavior analysis is improved, and the true intention of the user can be accurately restored.
Description
Technical Field
The invention relates to the technical field of information coding, in particular to a user behavior analysis method and system based on self-coding.
Background
User behavior analysis is in a narrow sense analysis of behavior data of a user, but in a broad sense the term encompasses both result analysis of user behavior and behavior analysis of a user. The results of the user's behavior are different from the user's behavior analysis, one being the result and one being the process. Products on the domestic market for user behavior analysis are now classified into user behavior analysis based on foreground data and user behavior analysis based on background data. The user behavior analysis based on the foreground technique focuses on the behavior analysis of the user, while the user behavior analysis based on the background technique focuses on the result analysis of the user behavior. Both types of products can be said to have some one-sidedness, and only a part of the user behavior analysis is completed.
At present, the conventional user behavior collection manner is to use buried points, as shown in fig. 1, where a product may be provided with a plurality of buried points, such as a user registration buried point, a user login buried point, a user exit buried point, a user logout buried point, a first page single machine buried point, a channel page single machine buried point, a detailed page single machine buried point, a pay page single machine buried point, a collection buried point, a comment buried point, a praise buried point and a payment initiation buried point.
The above approach has several significant drawbacks: firstly, the buried points are scattered, and particularly when some buried points are missing, for example, the buried points are not considered in the design process, so that the real intention of a user is difficult to restore; secondly, the dispersion of the behavior data is high, direct connection between the data is difficult, and a certain obstacle is caused to subsequent analysis.
Disclosure of Invention
Aiming at the defects existing in the prior art, the first aim of the invention is to provide a user behavior analysis method based on self-coding, which can improve the accuracy of user behavior analysis and accurately restore the real intention of a user.
In order to achieve the above purpose, the present invention provides the following technical solutions:
a self-encoding based user behavior analysis method, comprising:
acquiring user behavior data based on a self-coding mode to obtain self-coding sequences corresponding to different user behaviors;
obtaining all possible user behavior data according to a self-coding mode, wherein the possible user behavior data comprise preset standard sequences corresponding to all possible behaviors;
and comparing the self-coding sequence with a preset standard sequence, and performing defect detection and behavior detection of the user.
By adopting the technical scheme, the self-coding mode is adopted, the user behavior is correspondingly coded, the self-coding sequence in operation is obtained, the self-coding sequence is compared with the preset standard sequence, the behavior detection of the user can be carried out, the data is not required to be obtained through buried points, the accuracy of the user behavior acquisition is improved in a mode that the self-coding sequence in operation is compared with the preset standard sequence, the accuracy of the user behavior analysis is improved, and the real intention of the user can be accurately restored.
Further, the self-encoding method includes the steps of:
s101: performing tree-type coding on the application structure according to the substructures in the application, and generating application codes corresponding to the application structures;
s102: performing the same operation as S101 on all possible behavior structures of the user, and generating behavior codes corresponding to the behavior structures;
s103: after triggering a certain behavior in the application, the user starts to perform coding synthesis to obtain self-coding sequences corresponding to different user behaviors; or alternatively
S103': and (3) performing coding synthesis according to all possible user behaviors to obtain a preset standard sequence.
By adopting the technical scheme, the application structure and the behavior structure are separately encoded, and finally a finished behavior sequence segment is synthesized, and a final difference sequence is obtained through comparison among the behavior sequence segments during comparison.
Further, step S101 or step S102 includes the steps of:
s1010: encoding the application structure and the behavior structure based on the accessed weight;
s1011: the two nodes with the smallest weight begin to construct a Huffman tree, and the weight of a newly generated father node is calculated;
s1012: reselecting two nodes with the minimum weight value to synthesize a new subtree;
s1013: step S1011 and step S1012 are repeated until there are no other nodes, generating corresponding application codes and behavior codes.
By adopting the technical scheme, the Huffman coding can obtain variable length coding, and the coding applicability is higher.
Further, according to steps S103 and S103', the self-coding sequence and the preset standard sequence are synthesized in order of application coding, behavior coding, and trigger time in turn; or the self-coding sequence and the preset standard sequence are synthesized in the arrangement mode of the rest sequences of the application codes, the behavior codes and the triggering time.
By adopting the technical scheme, the complete behavior sequence segment can be formed by applying any combination of codes, behavior codes and triggering time, and the selectivity of the codes is higher.
Further, the self-coding sequence and the preset standard sequence are subjected to difference comparison to obtain a difference sequence, the difference sequence is input into a threshold function to calculate a threshold, the calculated threshold is compared with a set threshold, and whether the user behavior is established is judged.
Further, the application structure at least comprises one of the following: an application home page, an application purchase page, an application column page-1, an application column page-2, an application column page-3, an application campaign page, an application detail page-1, an application detail page-2, an application detail page-3, or an application payment page.
Further, the behavior structure at least comprises one of the following: up slide, down slide, back and forth slide, single click, double click, or zoom in.
Further, the preset standard sequence also includes all self-coded sequences generated based on user behavior of other users.
By adopting the technical scheme, the preset standard sequence comprises all self-coding sequences generated by the user behaviors of other users, and the behaviors of different users can be compared and analyzed, so that a client template is established, the user behaviors of different users can be compared and analyzed, and the auxiliary effect is realized on the storage of data and the behavior intention analysis of different types of clients.
The second object of the present invention is to provide a user behavior analysis system based on self-coding, which can improve the accuracy of user behavior analysis and can accurately restore the true intention of the user.
A self-encoding based user behavior analysis system, comprising:
acquiring user behavior data based on a self-coding mode, and obtaining self-encoders of self-coding sequences corresponding to different user behaviors;
acquiring all possible user behavior data according to a self-coding mode, wherein the possible user behavior data comprise preset standard sequences corresponding to all possible behaviors or rule sequence libraries of all self-coding sequences generated based on user behaviors of other users;
the analyzer is used for comparing the difference value of the sequences of the self-encoder and the rule sequence library by using the comparator and obtaining a difference value sequence;
a threshold value function device for inputting the difference value sequence to obtain a calculated threshold value;
and a result output device for comparing the calculated threshold value with the set threshold value and outputting whether the user behavior is established.
By adopting the technical scheme, the self-encoder acquires the user behavior data, encodes the user behavior to obtain the self-encoding sequence, and can continuously obtain the self-encoding sequence; then inputting the self-coding sequence and the information in the rule sequence library into an analyzer for analysis; the analyzer compares the difference value between the self-coding sequence and the sequence of the rule sequence library through the comparator to obtain a difference value sequence, inputs the difference value sequence into the threshold value function device to calculate a threshold value, compares the difference value sequence with a set threshold value, and outputs whether the user behavior is true or not through the result output device.
Further, the self-coding mode includes huffman coding.
Compared with the prior art, the invention has the advantages that: the self-coding sequence is compared with a preset standard sequence, so that the behavior detection of the user can be performed, the data is not required to be obtained through buried points, the accuracy of the acquisition of the user behavior is improved in a mode that the self-coding sequence is compared with the preset standard sequence, the accuracy of the analysis of the user behavior is improved, and the real intention of the user can be accurately restored;
in addition, because the coding form of the self-coding sequence has specific coding for the user behavior, after the self-coding sequence is compared with the preset standard sequence, if the application structure and the behavior structure coding of the user cannot be matched with the corresponding application structure and behavior structure coding in the preset standard sequence, the defect in the self-coding sequence is indicated, the defect problem of the product can be detected by adopting the user behavior data, and the detection of the performance of the product is improved.
Drawings
FIG. 1 is a simplified schematic diagram of a prior art buried point mode of collecting data;
FIG. 2 is a schematic diagram of the data construction of the self-encoder;
FIG. 3 is a simplified schematic diagram of the present invention with respect to tree Huffman coding;
FIG. 4 is a schematic diagram of sequence alignment.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and examples.
Based on the phenomenon that buried points are scattered and missing in the buried point mode, the obtained behavior data are high in dispersion, direct connection between the data is difficult, and a certain obstacle is caused to subsequent analysis. The invention provides a self-coding user behavior analysis method and system, aiming at improving the accuracy of user behavior acquisition according to a self-coding sequence corresponding to acquired user behavior in a self-coding mode; and the self-coding sequence is compared with a preset standard sequence, so that the accuracy of user behavior analysis is improved, and the real intention of the user can be accurately restored.
In the present invention, as shown in fig. 2, the application structure set in the product includes at least one of the following: an application home page, an application purchase page, an application column page-1, an application column page-2, an application column page-3, an application campaign page, an application detail page-1, an application detail page-2, an application detail page-3, or an application payment page. The behavior structure of the user comprises at least one of the following: up slide, down slide, back and forth slide, single click, double click, or zoom in. The user behavior described in the present application includes a combination of an application structure and a behavior structure, such as a stand-alone application purchase page or a double-click application payment page. The content of the application structure has priority, for example, in the application home page, the application purchase page cannot be skipped and the application payment page cannot be directly accessed.
The invention provides a user behavior analysis method based on self-coding, which comprises the following steps:
a: acquiring user behavior data based on a self-coding mode to obtain self-coding sequences corresponding to different user behaviors;
b: obtaining all possible user behavior data according to a self-coding mode, wherein the possible user behavior data comprise preset standard sequences corresponding to all possible behaviors;
c: and comparing the self-coding sequence with a preset standard sequence, and performing defect detection and behavior detection of the user.
Wherein, for step A, a self-coding mode is realized by a self-coder, and the self-coding mode comprises the following steps:
s101: performing tree-type coding on the application structure according to the substructures in the application, and generating application codes corresponding to the application structures;
s102: performing the same operation as S101 on all possible behavior structures of the user, and generating behavior codes corresponding to the behavior structures;
s103: after triggering a certain action in the application, the user starts to perform coding synthesis to obtain self-coding sequences corresponding to different user actions.
Based on step S101, the coding length is selected, and the application is tree-shaped coded according to the substructures in the application, which may be fixed-length coded or variable-length coded. In the application, huffman coding of tree type can be selected. The whole encoding process is as shown in fig. 2:
s1010: coding an application structure and a behavior structure based on the accessed weight, wherein if the application home page is 10, the application column page-1 is 7, the application column page-2 is 5, and the application activity page is 2;
s1011: the two nodes with the smallest weight begin to construct a Huffman tree, and calculate the weight of the newly generated father node, such as applying column page-2 weight 5, applying active page weight 2, and the father node weight is 7
S1012: reselecting two nodes with the minimum weight value to synthesize a new subtree;
s1013: step S1011 and step S1012 are repeated until there are no other nodes, generating corresponding application codes and behavior codes.
The user code generated in the above steps is first page 0, application column page-1 is 10, application column page-2 is 110, and application activity page is 111.
Based on step S102, the above steps of generating the application code are identical. Such as the generated behavior code being user click 0, user swipe up 10, user swipe down 110, etc. In addition, in order to distinguish from the application code, a prefix identification code 00 may be added to the behavior code. The final result is user click 000, user swipe up 0010, user swipe down 00110.
Based on step S103, after a user triggers a certain action in an application, starting to perform coding synthesis, and sequentially using a synthesized self-coding sequence and a preset standard sequence of application coding, action coding and trigger time; or the self-coding sequence and the preset standard sequence are synthesized in the arrangement mode of the rest sequences of the application codes, the behavior codes and the triggering time. In this embodiment, the sequence selected for applying the code, behavior code, and trigger time is self-coded. Such as 50 single clicks on the application home page by the user at 2018, 12, 08, and 14. The time information may generate binary data of a fixed length as part of a complete behavior sequence, e.g., 1450 converted to binary at 8 bits as 0000 1110 0011 0010. The result of the behavior of the first page of the 50-point single click application of the user 14 after the code synthesis with the behavior operation of the user is 0000 0000 1110 0011 0010. The user continues to operate to generate the complete self-coded sequence for subsequent analysis. In the present invention, a user's daily behavior code sequence is constructed in units of "daily".
User behavior can be used as a gene, and each application behavior has a specific code, so that later behavior detection and defect detection are facilitated.
Aiming at the step B, code synthesis is carried out according to all possible user behaviors of the user, and a preset standard sequence is obtained, wherein the self-code mode is consistent with the step S103. Meanwhile, in order to expand the breadth of user behavior analysis, the preset standard sequence also comprises all self-coding sequences generated based on the user behaviors of other users. After the user behavior sequence is generated, the user behavior sequence is stored in a rule sequence library such as a data storage, and after the new user generates the corresponding behavior sequence, the behavior sequence can be compared with the behavior sequences of other users stored before so as to judge the behavior tendency of the user group. Thus, the preset criteria sequence includes a set based on all behavior sequences for itself, as well as a set of behavior sequences for other users. For example, the preset standard sequence may include, but is not limited to, the following sequences:
the login error sequence is noted as S1:0101 0101 1101 1011 0000 0010 0111 … 1000, 1000;
the willingness-to-pay sequence is noted as S2:0111 1100 0001 0110 1100 1010 0001 … 1111;
the repeated click sequence is noted as S3:0110 1101 1111 1001 1010 0001 1111 … 0101;
the refund intent sequence is noted as S4:1010 1001 0101 0001 1001 1001 0010 … 0101.
Aiming at the step C, the analyzer compares the difference value of the sequences of the self-encoder and the rule sequence library by using a comparator to obtain a difference value sequence, and inputs the difference value sequence into a threshold value function device to obtain a calculated threshold value; comparing the calculated threshold value with the set threshold value, and outputting whether the user behavior is established or not through a result output device so as to achieve the purpose of user behavior detection.
If user a has a complete behavior sequence S of one day: 0000 0000 1110 0011 0010 0101 0000 … 0111, the behavioral tendencies of the user can be extracted by comparison according to a preset standard sequence. A single S is a behavioral sequence of one day, then an alignment can be started from the beginning of the S sequence, as shown in fig. 4 (arrow direction is the sliding direction), sliding one complete behavioral operation (application code + behavioral code + time) after each alignment. The comparison formula is as follows:
S∧S1, S∧S2, S∧S3, …，S∧SN。
according to the above formula, the difference sequences D can be obtained, respectively. The result output can be determined by inputting the difference sequence D into a certain threshold function, wherein the threshold function is selected from an S-shaped curve (i.e. a logic stark function):
it should be noted that, before the threshold function is input, the difference sequence D is converted into 10.
Similarly, if the user B has the behavior sequence segment S':1101 1110 0001 0000 1010 1101 0011 … 1010, which compares the degree of difference between user a and user B behavior, can be used to build a client template in this manner. The difference between the whole sequences of S and S' is compared to obtain the behavior difference between the user A and the user B by logical AND operation, and the following formula is adopted: s ∈S' is calculated to obtain a difference sequence D. And calculates the threshold value by a threshold function.
And comparing the calculated threshold with a set threshold, and then judging the user behavior or analyzing the user behavior. This behaviour is considered valid if we set a threshold value for the willingness-to-pay behaviour greater than 0.650. The output of the detected difference sequence D is f (D) =0.682, and since 0.682 > 0.650, there is a willingness to pay of the user.
In addition, because the coding form of the self-coding sequence has specific coding for the user behavior, after the self-coding sequence is compared with the preset standard sequence, if the application structure and the behavior structure coding of the user cannot be matched with the corresponding application structure and behavior structure coding in the preset standard sequence, the defect in the self-coding sequence is indicated, the defect problem of the product can be detected by adopting the user behavior data, and the detection of the performance of the product is improved.
The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above examples, and all technical solutions belonging to the concept of the present invention belong to the protection scope of the present invention. It should be noted that modifications and adaptations to the present invention may occur to one skilled in the art without departing from the principles of the present invention and are intended to be within the scope of the present invention.
Claims (8)
1. A self-encoding based user behavior analysis method, comprising:
acquiring user behavior data based on a self-coding mode to obtain self-coding sequences corresponding to different user behaviors, wherein the self-coding mode comprises the following steps:
s101: performing tree-type coding on the application structure according to the substructures in the application, and generating application codes corresponding to the application structures;
s102: performing the same operation as S101 on all possible behavior structures of the user, and generating behavior codes corresponding to the behavior structures;
s103: after triggering a certain behavior in the application, the user starts to perform coding synthesis to obtain self-coding sequences corresponding to different user behaviors; or alternatively
S103': performing coding synthesis according to all possible user behaviors to obtain a preset standard sequence;
wherein, step S101 or step S102 includes the following steps:
s1010: encoding an application structure or a behavior structure based on the accessed weight;
s1011: the two nodes with the smallest weight begin to construct a Huffman tree, and the weight of a newly generated father node is calculated;
s1012: reselecting two nodes with the minimum weight value to synthesize a new subtree;
s1013: repeating step S1011 and step S1012 until there are no other nodes, generating corresponding application codes or behavior codes;
obtaining all possible user behavior data according to a self-coding mode, wherein the possible user behavior data comprise preset standard sequences corresponding to all possible behaviors;
and comparing the self-coding sequence with a preset standard sequence, and performing defect detection and behavior detection of the user.
2. The method according to claim 1, wherein the self-code sequence and the preset standard sequence are synthesized in order of application code, behavior code, and trigger time in sequence according to steps S103 and S103'; or the self-coding sequence and the preset standard sequence are synthesized in the arrangement mode of the rest sequences of the application codes, the behavior codes and the triggering time.
3. The method according to claim 1, wherein the self-code sequence and the preset standard sequence are subjected to difference comparison to obtain a difference sequence, the difference sequence is input into a threshold function to calculate a threshold, and the calculated threshold is compared with a set threshold to determine whether the user behavior is established.
4. The method of claim 1, wherein the application structure comprises at least one of: an application home page, an application purchase page, an application column page-1, an application column page-2, an application column page-3, an application campaign page, an application detail page-1, an application detail page-2, an application detail page-3, or an application payment page.
5. The method of claim 1, wherein the behavioral structure comprises at least one of: up slide, down slide, back and forth slide, single click, double click, or zoom in.
6. The method of claim 1, wherein the pre-set criteria sequence further comprises all self-coded sequences generated based on user behavior of other users.
7. A self-encoding based user behavior analysis system, comprising:
acquiring user behavior data based on a self-coding mode, and obtaining self-encoders of self-coding sequences corresponding to different user behaviors;
the self-coding mode comprises the following steps:
s101: performing tree-type coding on the application structure according to the substructures in the application, and generating application codes corresponding to the application structures;
s102: performing the same operation as S101 on all possible behavior structures of the user, and generating behavior codes corresponding to the behavior structures;
s103: after triggering a certain behavior in the application, the user starts to perform coding synthesis to obtain self-coding sequences corresponding to different user behaviors; or alternatively
S103': performing coding synthesis according to all possible user behaviors to obtain a preset standard sequence;
wherein, step S101 or step S102 includes the following steps:
s1010: encoding an application structure or a behavior structure based on the accessed weight;
s1011: the two nodes with the smallest weight begin to construct a Huffman tree, and the weight of a newly generated father node is calculated;
s1012: reselecting two nodes with the minimum weight value to synthesize a new subtree;
s1013: repeating step S1011 and step S1012 until there are no other nodes, generating corresponding application codes or behavior codes;
obtaining all possible user behavior data according to a self-coding mode, wherein the possible user behavior data comprise preset standard sequences corresponding to all possible behaviors, and the preset standard sequences comprise all self-coding sequences generated based on user behaviors of other users;
the analyzer is used for comparing the difference value between the self-encoder and a preset standard sequence by using a comparator and obtaining a difference value sequence;
a threshold value function device for inputting the difference value sequence to obtain a calculated threshold value;
and a result output device for comparing the calculated threshold value with the set threshold value and outputting whether the user behavior is established.
8. The system of claim 7, wherein the self-encoding means comprises huffman coding.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910142937.5A CN109858977B (en) | 2019-02-26 | 2019-02-26 | User behavior analysis method and system based on self-coding |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910142937.5A CN109858977B (en) | 2019-02-26 | 2019-02-26 | User behavior analysis method and system based on self-coding |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109858977A CN109858977A (en) | 2019-06-07 |
| CN109858977B true CN109858977B (en) | 2023-06-16 |
Family
ID=66899030
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910142937.5A Active CN109858977B (en) | 2019-02-26 | 2019-02-26 | User behavior analysis method and system based on self-coding |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109858977B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111615149B (en) * | 2020-05-13 | 2021-05-11 | 和智信(山东)大数据科技有限公司 | Signaling track data compression method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107578294A (en) * | 2017-09-28 | 2018-01-12 | 北京小度信息科技有限公司 | User's behavior prediction method, apparatus and electronic equipment |
| CN108153870A (en) * | 2017-12-25 | 2018-06-12 | 四川长虹电器股份有限公司 | A kind of user access path Forecasting Methodology |
| CN108829572A (en) * | 2018-05-30 | 2018-11-16 | 北京奇虎科技有限公司 | The analysis method and device of user's login behavior |
| CN109255640A (en) * | 2017-07-13 | 2019-01-22 | 阿里健康信息技术有限公司 | A kind of method, apparatus and system of determining user grouping |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8108406B2 (en) * | 2008-12-30 | 2012-01-31 | Expanse Networks, Inc. | Pangenetic web user behavior prediction system |
| US8874485B2 (en) * | 2011-12-16 | 2014-10-28 | Palo Alto Research Center Incorporated | Privacy-preserving behavior targeting for digital coupons |
-
2019
- 2019-02-26 CN CN201910142937.5A patent/CN109858977B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109255640A (en) * | 2017-07-13 | 2019-01-22 | 阿里健康信息技术有限公司 | A kind of method, apparatus and system of determining user grouping |
| CN107578294A (en) * | 2017-09-28 | 2018-01-12 | 北京小度信息科技有限公司 | User's behavior prediction method, apparatus and electronic equipment |
| CN108153870A (en) * | 2017-12-25 | 2018-06-12 | 四川长虹电器股份有限公司 | A kind of user access path Forecasting Methodology |
| CN108829572A (en) * | 2018-05-30 | 2018-11-16 | 北京奇虎科技有限公司 | The analysis method and device of user's login behavior |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109858977A (en) | 2019-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111178523B (en) | Behavior detection method and device, electronic equipment and storage medium | |
| CN111026915B (en) | Video classification method, video classification device, storage medium and electronic equipment | |
| CN103548076B (en) | Utilize sound signal to identify the apparatus and method of content | |
| CN109598334B (en) | Sample generation method and device | |
| US20100085481A1 (en) | Frame based video matching | |
| CN112235327A (en) | Abnormal log detection method, device, equipment and computer readable storage medium | |
| KR102042168B1 (en) | Methods and apparatuses for generating text to video based on time series adversarial neural network | |
| KR102359090B1 (en) | Method and System for Real-time Abnormal Insider Event Detection on Enterprise Resource Planning System | |
| CN109858977B (en) | User behavior analysis method and system based on self-coding | |
| Sajadmanesh et al. | Continuous-time relationship prediction in dynamic heterogeneous information networks | |
| CN116383096B (en) | Micro-service system anomaly detection method and device based on multi-index time sequence prediction | |
| CN112231582A (en) | Website recommendation method and equipment based on variational self-coding data fusion | |
| Wei et al. | Sequence-to-segment networks for segment detection | |
| Matos et al. | A comparison of data-driven approaches for mobile marketing user conversion prediction | |
| CN113988156A (en) | Time series clustering method, system, equipment and medium | |
| CN114185766A (en) | Code detection method and device, electronic equipment and storage medium | |
| CN117149986A (en) | Real-time big data processing method and system based on multi-stage data channel | |
| Lillack et al. | Improved prediction of non-functional properties in software product lines with domain context | |
| CN109376138B (en) | Abnormal combination detection method and device for multi-dimensional data | |
| CN111010595A (en) | New program recommendation method and device | |
| CN116415063A (en) | Cloud service recommendation method and device | |
| JPWO2011016281A1 (en) | Information processing apparatus and program for Bayesian network structure learning | |
| CN112509644A (en) | Molecular optimization method, system, terminal equipment and readable storage medium | |
| CN111583991A (en) | Method, system, equipment and medium for gene regulation and control network reconstruction | |
| CN110991508A (en) | Anomaly detector recommendation method, device and equipment |
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 |