CN110008974A - Behavioral data prediction technique, device, electronic equipment and computer storage medium - Google Patents
Behavioral data prediction technique, device, electronic equipment and computer storage medium Download PDFInfo
- Publication number
- CN110008974A CN110008974A CN201811410240.3A CN201811410240A CN110008974A CN 110008974 A CN110008974 A CN 110008974A CN 201811410240 A CN201811410240 A CN 201811410240A CN 110008974 A CN110008974 A CN 110008974A
- Authority
- CN
- China
- Prior art keywords
- historical behavior
- action
- data
- predicted
- feature
- 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.)
- Pending
Links
- 230000003542 behavioural effect Effects 0.000 title claims abstract description 121
- 238000000034 method Methods 0.000 title claims abstract description 57
- 239000013598 vector Substances 0.000 claims abstract description 192
- 230000009471 action Effects 0.000 claims abstract description 116
- 230000006399 behavior Effects 0.000 claims description 280
- 230000010354 integration Effects 0.000 claims description 25
- 238000012545 processing Methods 0.000 claims description 10
- 230000015654 memory Effects 0.000 claims description 8
- 238000012360 testing method Methods 0.000 abstract description 19
- 238000004458 analytical method Methods 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 13
- 230000000694 effects Effects 0.000 description 9
- 230000006870 function Effects 0.000 description 8
- 238000003064 k means clustering Methods 0.000 description 7
- 241000406668 Loxodonta cyclotis Species 0.000 description 6
- 230000006854 communication Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000004590 computer program Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000007717 exclusion Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003319 supportive effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Operations Research (AREA)
- Artificial Intelligence (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the invention discloses a kind of behavioral data prediction technique, device, electronic equipment and computer storage mediums, which comprises obtains the feature vector and its historical behavior data of historical behavior object；The feature vector of object of action to be predicted is obtained, and calculates the similarity between the feature vector of the object of action to be predicted and the feature vector of the historical behavior object；The historical behavior data of the corresponding historical behavior object of the similarity for meeting preset condition are determined as to the predictive behavior data of the object of action to be predicted.The technical solution passes through the analysis for object of action to be predicted and historical behavior object behavior similarity, use and meets the historical behavior data of the corresponding historical behavior object of preset condition similarity as the predictive behavior data of object of action to be predicted, to realize the accurate prediction of object of action behavioral data to be predicted, improve data forecasting accuracy, the case where reducing the unstable bring puzzlement of test data early period, avoiding the occurrence of incorrect decision.
Description
Technical field
The present embodiments relate to technical field of data processing, and in particular to a kind of behavioral data prediction technique, device, electricity
Sub- equipment and computer storage medium.
Background technique
With the development of internet technology, for many e-platforms, whenever thering is new product or new business to need to issue
Or when carrying out, usually all can first selected part user test, if new product performance in test user is good, just meeting
Consider to the open product of full dose user, and before opening to full dose user, it is also necessary to consider open to full dose user's
Decision is carried out after effect again.The related data of user is tested before being typically based in the prior art, it is pre- using time series etc.
Survey method predicts full dose user's using effect data, and is judged accordingly.But since new product or new business are online
Time is shorter, and it is less and data fluctuations are larger to test number of users early period, by traditional prediction technique be difficult for it is open to
Effect when other users is precisely predicted, so that data forecasting accuracy is poor, it is also possible to the case where leading to incorrect decision.
Summary of the invention
The embodiment of the present invention provides a kind of behavioral data prediction technique, device, electronic equipment and computer storage medium.
In a first aspect, providing a kind of behavioral data prediction technique in the embodiment of the present invention.
Specifically, the behavioral data prediction technique, comprising:
Obtain the feature vector and its historical behavior data of historical behavior object；
Obtain the feature vector of object of action to be predicted, and calculate the feature vector of the object of action to be predicted with it is described
Similarity between the feature vector of historical behavior object；
The historical behavior data of the corresponding historical behavior object of the similarity for meeting preset condition are determined as described to pre-
Survey the predictive behavior data of object of action.
With reference to first aspect, the embodiment of the present invention is in the first implementation of first aspect, the acquisition history row
For the feature vector and its historical behavior data of object, comprising:
Determine feature vector element；
The feature vector of the historical behavior object is obtained according to determining feature vector element；
Obtain historical behavior data of the historical behavior object in default historical time section.
With reference to first aspect with the first implementation of first aspect, second in first aspect of the embodiment of the present invention
It is described to obtain the historical behavior object after the historical behavior data in default historical time section in implementation, also wrap
It includes:
Calculate average behavioral data of the historical behavior object in default historical time section；
The historical behavior object is divided into N class based on the difference between the average behavioral data, wherein N is integer；
Calculate averaged historical behavioral data in the integration characteristics vector and class of historical behavior object described in every one kind；
The integration characteristics vector is determined as to the feature vector of the historical behavior object, by averaged historical in the class
Behavioral data is determined as the historical behavior data of the historical behavior object.
With reference to first aspect, second of implementation of the first implementation of first aspect and first aspect, this public affairs
It is opened in the third implementation of first aspect, it is described to calculate the historical behavior object putting down in default historical time section
After equal behavioral data, further includes:
The average behavioral data is normalized.
With reference to first aspect, the first implementation of first aspect, first aspect second of implementation and first
The third implementation of aspect, the disclosure are described to obtain behavior pair to be predicted in the 4th kind of implementation of first aspect
The feature vector of elephant, and calculate the object of action to be predicted feature vector and the historical behavior object feature vector it
Between similarity, comprising:
Obtain the feature vector of the object of action to be predicted；
The feature vector of feature vector and the historical behavior object for the object of action to be predicted carries out pre-
If coded treatment, the coding of the corresponding coding characteristic vector for obtaining the object of action to be predicted and the historical behavior object
Feature vector；
Calculate the coding characteristic vector of the object of action to be predicted and the coding characteristic vector of the historical behavior object
Between similarity.
With reference to first aspect, the first implementation, second of implementation of first aspect, first party of first aspect
The third implementation in face and the 4th kind of implementation of first aspect, five kind implementation of the disclosure in first aspect
In, the historical behavior data by the corresponding historical behavior object of the similarity for meeting preset condition are determined as described to be predicted
The predictive behavior data of object of action, comprising:
The historical behavior data of the corresponding historical behavior object of highest similarity are determined as the object of action to be predicted
Predictive behavior data.
Second aspect provides a kind of behavioral data prediction meanss in the embodiment of the present invention.
Specifically, the behavioral data prediction meanss, comprising:
Module is obtained, is configured as obtaining the feature vector and its historical behavior data of historical behavior object；
Computing module is configured as obtaining the feature vector of object of action to be predicted, and calculates the behavior pair to be predicted
Similarity between the feature vector of elephant and the feature vector of the historical behavior object；
Determining module is configured as that the historical behavior number of the corresponding historical behavior object of similarity of preset condition will be met
According to the predictive behavior data for being determined as the object of action to be predicted.
In conjunction with second aspect, the embodiment of the present invention is in the first implementation of second aspect, the acquisition module packet
It includes:
First determines submodule, is configured to determine that feature vector element；
First acquisition submodule is configured as obtaining the spy of the historical behavior object according to determining feature vector element
Levy vector；
Second acquisition submodule is configured as obtaining history row of the historical behavior object in default historical time section
For data.
In conjunction with the first of second aspect and second aspect implementation, second in second aspect of the embodiment of the present invention
In implementation, the acquisition module further include:
First computational submodule is configured as calculating average row of the historical behavior object in default historical time section
For data；
Classification submodule is configured as dividing the historical behavior object based on the difference between the average behavioral data
For N class, wherein N is integer；
Second computational submodule, be configured as calculating historical behavior object described in every one kind integration characteristics vector and
Averaged historical behavioral data in class；
Second determines submodule, is configured as the integration characteristics vector being determined as the feature of the historical behavior object
Averaged historical behavioral data in the class is determined as the historical behavior data of the historical behavior object by vector.
In conjunction with the first implementation of second aspect, second aspect and second of implementation of second aspect, this public affairs
It is opened in the third implementation of second aspect, the acquisition module further include:
Submodule is normalized, is configured as that the average behavioral data is normalized.
In conjunction with the first implementation of second aspect, second aspect, second of implementation and second of second aspect
The third implementation of aspect, in the 4th kind of implementation of second aspect, the computing module includes: the disclosure
Third acquisition submodule is configured as obtaining the feature vector of the object of action to be predicted；
Encoding submodule is configured as feature vector and the historical behavior pair for the object of action to be predicted
The feature vector of elephant carries out pre-arranged code processing, the corresponding coding characteristic vector for obtaining the object of action to be predicted and described
The coding characteristic vector of historical behavior object；
Third computational submodule is configured as calculating the coding characteristic vector of the object of action to be predicted and the history
Similarity between the coding characteristic vector of object of action.
The first implementation, second of implementation of second aspect, second party in conjunction with second aspect, second aspect
The third implementation in face and the 4th kind of implementation of second aspect, five kind implementation of the disclosure in second aspect
In, the determining module is configured as:
The historical behavior data of the corresponding historical behavior object of highest similarity are determined as the object of action to be predicted
Predictive behavior data.
The third aspect, the embodiment of the invention provides a kind of electronic equipment, including memory and processor, the memories
It is executed in above-mentioned first aspect based on behavioral data prediction technique by storing one or more supportive behavior data prediction meanss
Calculation machine instruction, the processor is configured to for executing the computer instruction stored in the memory.The behavioral data
Prediction meanss can also include communication interface, for behavioral data prediction meanss and other equipment or communication.
Fourth aspect, it is pre- for storing behavioral data the embodiment of the invention provides a kind of computer readable storage medium
Computer instruction used in device is surveyed, it includes be behavioral data for executing behavioral data prediction technique in above-mentioned first aspect
Computer instruction involved in prediction meanss.
Technical solution provided in an embodiment of the present invention can include the following benefits:
Above-mentioned technical proposal by the analysis for behavior similarity between object of action to be predicted and historical behavior object,
Use the historical behavior data of the corresponding historical behavior object of the similarity for meeting preset condition as object of action to be predicted
Predictive behavior data realize the accurate prediction for object of action behavioral data to be predicted with this, and then improve data prediction
The case where accuracy reduces the unstable bring puzzlement of test data early period, avoids the occurrence of incorrect decision.
It should be understood that above general description and following detailed description be only it is exemplary and explanatory, not
The embodiment of the present invention can be limited.
Detailed description of the invention
In conjunction with attached drawing, pass through the detailed description of following non-limiting embodiment, other feature, the mesh of the embodiment of the present invention
And advantage will be apparent.In the accompanying drawings:
Fig. 1 shows the flow chart of behavioral data prediction technique according to an embodiment of the present invention；
Fig. 2 shows the flow charts of the step S101 of the behavioral data prediction technique of embodiment according to Fig. 1；
Fig. 3 shows the flow chart of the step S101 of the behavioral data prediction technique of another embodiment according to the present invention；
Fig. 4 shows the flow chart of the step S102 of the behavioral data prediction technique of embodiment according to Fig. 1；
Fig. 5 shows the structural block diagram of behavioral data prediction meanss according to an embodiment of the present invention；
Fig. 6 shows the structural block diagram of the acquisition module 501 of the behavioral data prediction meanss of embodiment according to Fig.5,；
Fig. 7 shows the structural frames of the acquisition module 501 of the behavioral data prediction meanss of another embodiment according to the present invention
Figure；
Fig. 8 shows the structural block diagram of the computing module 502 of the behavioral data prediction meanss of embodiment according to Fig.5,；
Fig. 9 shows the structural block diagram of electronic equipment according to an embodiment of the present invention；
Figure 10 is adapted for the computer system for realizing behavioral data prediction technique according to an embodiment of the present invention
Structural schematic diagram.
Specific embodiment
Hereinafter, the illustrative embodiments of the embodiment of the present invention will be described in detail with reference to the attached drawings, so that art technology
Them are easily implemented in personnel.In addition, for the sake of clarity, being omitted in the accompanying drawings unrelated with description illustrative embodiments
Part.
In embodiments of the present invention, it should be appreciated that the term of " comprising " or " having " etc. is intended to refer in this specification
The presence of disclosed feature, number, step, behavior, component, part or combinations thereof, and be not intended to exclude it is one or more its
A possibility that his feature, number, step, behavior, component, part or combinations thereof exist or are added.
It also should be noted that in the absence of conflict, the feature in embodiment and embodiment in the present invention
It can be combined with each other.Embodiment that the present invention will be described in detail below with reference to the accompanying drawings and embodiments.
Technical solution provided in an embodiment of the present invention passes through for row between object of action to be predicted and historical behavior object
For the analysis of similarity, use the historical behavior data of the corresponding historical behavior object of the similarity for meeting preset condition as to
The predictive behavior data of predictive behavior object realize the accurate prediction for object of action behavioral data to be predicted with this, in turn
The case where improving the accuracy of data prediction, reducing the unstable bring puzzlement of test data early period, avoid the occurrence of incorrect decision.
Fig. 1 shows the flow chart of behavioral data prediction technique according to an embodiment of the present invention, as shown in Figure 1, described
Behavioral data prediction technique includes the following steps S101-S103:
In step s101, the feature vector and its historical behavior data of historical behavior object are obtained；
In step s 102, the feature vector of object of action to be predicted is obtained, and calculates the object of action to be predicted
Similarity between feature vector and the feature vector of the historical behavior object；
In step s 103, the historical behavior data of the corresponding historical behavior object of the similarity for meeting preset condition are true
It is set to the predictive behavior data of the object of action to be predicted.
It is mentioned above, with the development of internet technology, for many e-platforms, whenever having new product or new business
When needing to issue or carry out, usually all can first selected part user test, if new product shows in test user
Well, can just consider to the open product of full dose user, and before opening to full dose user, it is also necessary to consider open to complete
The effect for measuring user carries out decision again later.The related data that user is tested before being typically based in the prior art, utilizes the time
The prediction techniques such as sequence predict full dose user's using effect data, and are judged accordingly.But due to new product or newly
Business on-line time is shorter, and it is less and data fluctuations are larger to test number of users early period, is difficult pair by traditional prediction technique
Effect when opening to other users is precisely predicted, so that data forecasting accuracy is poor, it is also possible to decision be caused to be lost
Accidentally the case where.
In view of the above problem, in this embodiment, propose a kind of behavioral data prediction technique, this method by for
The analysis of behavior similarity between object of action and historical behavior object to be predicted, it is corresponding using the similarity for meeting preset condition
Historical behavior object predictive behavior data of the historical behavior data as object of action to be predicted, realized with this for pre-
The accurate prediction of object of action behavioral data is surveyed, and then improves the accuracy of data prediction, it is unstable to reduce test data early period
The case where bring perplexs, and avoids the occurrence of incorrect decision.
In an optional implementation of the present embodiment, the object of action refers to implementing a certain specified object
The user of a certain behavior, wherein the specified object can for new product, new demand servicing, new business or new projects, with new product,
New demand servicing, similar product, service, business or the project or other appointed products of new business or new projects, service, business or
Project；The behavioral data refers to that a certain specified object occurs for object of action the number of a certain default behavior.For example, if
The specified object is a new demand servicing, then the object of action can be the user using the new demand servicing, the history row
It can be the user that the new demand servicing was used in default historical time section, the historical behavior of the historical behavior object for object
Data can be the number that the user uses the new demand servicing in the default historical time section, and the behavior pair to be predicted
As being exactly in default future time section there is a possibility that with the user of the new demand servicing, the predictive behavior of the object of action to be predicted
Data be exactly the user that is predicted according to above-mentioned historical data in the default future time section there is a possibility that new with this
The number of service.
Wherein, for the standard uniformly compared, accurately prediction data, the feature vector of the historical behavior object are obtained
It is corresponding identical as the feature vector element of the object of action to be predicted.
In an optional implementation of the present embodiment, the feature vector of the object of action may include in following characteristics
It is one or more: the behavioural characteristic of the attributive character of the object of action and the object of action, wherein the object of action
Attributive character may include one of following element or a variety of: age, gender, occupation, education level, marital status etc.
Deng the behavioural characteristic of the object of action may include one of following element or a variety of: the view rate of specified object, specified pair
Whether the click volume of elephant, the forward rate of specified object were bought and specify object, the amount of placing an order of specified object etc..Certainly, described
Feature vector may also comprise other characteristic elements, those skilled in the art can according to the needs of practical application, the spy of object of action
Point is selected, and the present invention is not especially limited it.
In an optional implementation of the present embodiment, as shown in Fig. 2, the step S101, i.e. acquisition historical behavior
The step of feature vector and its historical behavior data of object, include the following steps S201-S203:
In step s 201, feature vector element is determined；
In step S202, the feature vector of the historical behavior object is obtained according to determining feature vector element；
In step S203, historical behavior data of the historical behavior object in default historical time section are obtained.
In this embodiment, first according to the needs of practical application and the characteristics of object of action, the spy of specified object
Point is to determine feature vector element；Then the feature of the historical behavior object according to determining feature vector element extraction to
Amount；Finally obtain historical behavior data of the historical behavior object in default historical time section.
Wherein, the default historical time section can determine according to the needs of practical application, and the present invention does not make it specifically
It limits.
Assuming that the behavioral data is the new demand servicing order amount of user, when the default historical time section is that new demand servicing is tested
Between section, then historical behavior data of the historical behavior object in default historical time section are exactly that the user tests in new demand servicing
The new demand servicing order amount generated in period.
In an optional implementation of the present embodiment, the step S203 obtains the historical behavior object and exists
It further include that the historical behavior data are handled after the step of presetting the historical behavior data in historical time section
Step, i.e., as shown in figure 3, the step S101 includes the following steps S301-S307:
In step S301, feature vector element is determined；
In step s 302, the feature vector of the historical behavior object is obtained according to determining feature vector element；
In step S303, historical behavior data of the historical behavior object in default historical time section are obtained；
In step s 304, average behavioral data of the historical behavior object in default historical time section is calculated；
In step S305, the historical behavior object is divided by N class based on the difference between the average behavioral data,
Wherein, N is integer；
In step S306, calculates and averagely gone through in the integration characteristics vector and class of historical behavior object described in every one kind
History behavioral data；
In step S307, the integration characteristics vector is determined as to the feature vector of the historical behavior object, by institute
State the historical behavior data that averaged historical behavioral data in class is determined as the historical behavior object.
In order to embody the statistical property of behavioral data, the behavioral data is more reasonably used, in this embodiment, also
Cluster average treatment is carried out for the historical behavior data, specifically, obtains the historical behavior object in default history
Between after historical behavior data in section, calculate average behavior of each historical behavior object in default historical time section first
Data, for example, if the behavioral data is the new demand servicing order amount of user, when the default historical time section is that new demand servicing is tested
Between section, then the average behavioral data is exactly the user average daily new demand servicing order amount in new demand servicing testing time section；So
The historical behavior object is divided by N class based on the difference between the average behavioral data afterwards, wherein N is integer；Then it counts
Calculate averaged historical behavioral data in the integration characteristics vector and class of historical behavior object described in every one kind；It finally will be described whole
The feature vector that feature vector is determined as the historical behavior object is closed, averaged historical behavioral data in the class is determined as institute
The historical behavior data of historical behavior object are stated, it is subsequent that the corresponding data of itself and the object of action to be predicted is subjected to similarity
It calculates.
In an optional implementation of the present embodiment, when the historical behavior object is divided into N class, it can be used poly-
Class method, such as k-means clustering method, naturally it is also possible to use other classification methods.When use k-means clustering method
When, ancon rule can be used to determine categorical measure K, i.e., first obtain the corresponding cost function of different K values, with the increase of K value,
Cluster bring average distortion degree can become smaller, and the number of samples in every one kind can reduce therewith, but with the increase of K value, meeting
There is an inflection point, after inflection point, even if K value increases again, average distortion degree also no longer becomes smaller, and one in the present embodiment can
It selects in implementation, K value corresponding to this inflection point can be taken as the categorical measure K in k-means clustering method.
In an optional implementation of the present embodiment, when carrying out k-means cluster, it can be commented by Euclidean distance
Difference between average behavioral data described in valence, with the other division of implementation of class.
It is special in the integration for calculating historical behavior object described in every one kind in an optional implementation of the present embodiment
It, can be to the corresponding eigenvalue averaged in different historical behavior characteristics of objects vectors, then by each characteristic value when levying vector
Corresponding average value forms the integration characteristics vector of historical behavior object in such.
For example, if the historical behavior object is divided into 3 classes, each historical behavior by above-mentioned k-means clustering method
The feature vector of object includes 5 characteristic values, it is assumed that includes 6 historical behavior objects in the 1st class, then each going through in the 1st class
The feature vector of history object of action can indicate are as follows: object 1 { feature 11, feature 12, feature 13, feature 14, feature 15 }, object
2 { feature 21, feature 22, feature 23, feature 24, features 25 }, object 3 { feature 31, feature 32, feature 33, feature 34, feature
35 }, object 4 { feature 41, feature 42, feature 43, feature 44, feature 45 }, object 5 { feature 51, feature 52, feature 53, feature
54, feature 55 }, object 6 { feature 61, feature 62, feature 63, feature 64, feature 65 }, then after feature integration, in the 1st class
The integration characteristics vector of the historical behavior object can indicate are as follows: integrating vector, { feature 1, feature 2, feature 3, feature 4 are special
Sign 5 }, wherein feature 1=(feature 11+ feature 21+ feature 31+ feature 41+ feature 51+ feature 61)/6, feature 2=(feature 12
+ feature 22+ feature 32+ feature 42+ feature 52+ feature 62)/6, (feature 13+ feature 23+ feature 33+ feature 43+ is special by feature 3=
Levy 53+ feature 63)/6, feature 4=(feature 14+ feature 24+ feature 34+ feature 44+ feature 54+ feature 64)/6, feature 5=
(feature 15+ feature 25+ feature 35+ feature 45+ feature 55+ feature 65)/6.The integration characteristics of historical behavior object in other classes
The calculating of vector can and so on, the present invention repeats no more.
In an optional implementation of the present embodiment, put down in the class for calculating historical behavior object described in every one kind
When equal historical behavior data, the average method of data can also be used, i.e., by the history row of historical behavior objects all in certain one kind
It is averaged for data, averaged historical behavioral data in the class can be obtained.
In an optional implementation of the present embodiment, the calculating of subsequent similarity for convenience, the step S304,
The historical behavior object is calculated after the step of presetting the average behavioral data in historical time section, further includes following step
It is rapid:
The average behavioral data is normalized.
Wherein, those skilled in the art can according to the needs of practical application and the characteristics of average behavioral data selects normalizing
Change mode, the present invention are not especially limited it.
In an optional implementation of the present embodiment, normalizing is carried out for the average behavioral data using following formula
Change:
Y=(col-min (col))/(max (col)-min (col)),
Wherein, y indicates the average behavioral data that a certain user obtains after normalized, col indicate the user without
The average behavioral data of normalized, min (col) indicate it is similar in average behavior number of all users without normalized
Minimum value in, max (col) indicate it is similar in all users without the maximum in the average behavioral data of normalized
Value.
In addition, if the average behavior data packet of a certain user includes two or more subdatas, different user
The subdata of corresponding position is corresponding in average behavioral data, then needing useful for institute when being normalized
The corresponding subdata in family is normalized respectively.For example, the average behavior subdata of each user can be in line, that
The average behavior subdata of multiple users just constitutes a data matrix after being arranged successively, wherein in the data matrix
Every column data is exactly the subdata of corresponding different user, every column data can be returned respectively according to above formula at this time
One changes, wherein col can indicate a certain user without the average behavior subdata on a certain column of normalized, min
(col) minimum value in the column in all average behavior subdatas without normalized is indicated, max (col) indicates the column
In maximum value in all average behavior subdatas without normalized, finally can be obtained by dimension and every in this way
The average behavioral data of a user is identical to normalize average behavioral data.
In an optional implementation of the present embodiment, as shown in figure 4, the step S102, that is, obtain row to be predicted
For the feature vector of object, and calculate the object of action to be predicted feature vector and the historical behavior object feature to
The step of similarity between amount, include the following steps S401-S403:
In step S401, the feature vector of the object of action to be predicted is obtained；
In step S402, the spy of feature vector and the historical behavior object for the object of action to be predicted
It levies vector and carries out pre-arranged code processing, correspondence obtains the coding characteristic vector and the history row of the object of action to be predicted
For the coding characteristic vector of object；
In step S403, the coding characteristic vector and the historical behavior object of the object of action to be predicted are calculated
Similarity between coding characteristic vector.
In order to facilitate the calculating of similarity between feature vector, in this embodiment, also for the behavior to be predicted
The feature vector of object and the feature vector of the historical behavior object carry out pre-arranged code processing, then calculate what correspondence obtained
The coding characteristic vector of the coding characteristic vector of the object of action to be predicted and the historical behavior object.
In an optional implementation of the present embodiment, using one-hot coding (One-Hot Encoding, One-Hot)
To carry out coded treatment to the feature vector of the object of action to be predicted and the feature vector of the historical behavior object.
One-Hot coding is also known as an efficient coding, and method is to be encoded using N bit status register to N number of state, often
A state has its independent register-bit, and when any, and only wherein one effectively, that is to say, that for each
A feature, if it has m probable value, after one-hot coding, the corresponding characteristic value of this feature has reformed into m binary
Characteristic value, also, these feature mutual exclusions, only one each value are active.
In an optional implementation of the present embodiment, in the coding characteristic vector for calculating the object of action to be predicted
When similarity between the coding characteristic vector of the historical behavior object, included angle cosine value similarity evaluation side can be used
Formula, that is, calculate the cosine value for the angle that two feature vectors are formed, and obtained cosine value range of results is described between [- 1,1]
The size of cosine value can be used to evaluate the similarity degree between two feature vectors: angle is smaller, cosine value closer to 1,
With regard to illustrating that the direction of two feature vectors more coincide, then two feature vectors are more similar, conversely, angle is bigger, cosine
Value just illustrates that the direction difference of two feature vectors is bigger, then two feature vectors are more dissimilar closer to -1.
Wherein, the cosine value for the angle that two feature vectors are formed can be represented by the following formula:
Wherein, θ indicates that the angle that two feature vectors are formed, cos (θ) indicate the cosine value of the angle, xiAnd yiTable respectively
Show the ith feature value in two feature vectors, n indicates the quantity of characteristic value in two feature vectors.
In an optional implementation of the present embodiment, the step S103 will meet the similarity of preset condition
The historical behavior data of corresponding historical behavior object are determined as the step of predictive behavior data of the object of action to be predicted,
The following steps are included:
The historical behavior data of the corresponding historical behavior object of highest similarity are determined as the object of action to be predicted
Predictive behavior data.
In this embodiment, it is believed that the behavior with the most similar historical behavior object of object of action to be predicted is practised
Used closest with the behavioural habits of the object of action to be predicted, therefore, the historical behavior data of the historical behavior object can
By as the behavioral data predicted for the object of action to be predicted.
It is subsequent, if it is desired, the predictive behavior data of all object of action to be predicted can be added, can be obtained described pre-
If the predictive behavior total data in future time, which can provide the reference in data for technical staff, to help it to make
Corresponding behaviour decision making.
Following is apparatus of the present invention embodiment, can be used for executing embodiment of the present invention method.
Fig. 5 shows the structural block diagram of behavioral data prediction meanss according to an embodiment of the present invention, which can lead to
Cross being implemented in combination with as some or all of of electronic equipment of software, hardware or both.As shown in figure 5, the behavior number
It is predicted that device includes:
Module 501 is obtained, is configured as obtaining the feature vector and its historical behavior data of historical behavior object；
Computing module 502 is configured as obtaining the feature vector of object of action to be predicted, and calculates the behavior to be predicted
Similarity between the feature vector of object and the feature vector of the historical behavior object；
Determining module 503 is configured as that the history row of the corresponding historical behavior object of similarity of preset condition will be met
It is determined as the predictive behavior data of the object of action to be predicted for data.
It is mentioned above, with the development of internet technology, for many e-platforms, whenever having new product or new business
When needing to issue or carry out, usually all can first selected part user test, if new product shows in test user
Well, can just consider to the open product of full dose user, and before opening to full dose user, it is also necessary to consider open to complete
The effect for measuring user carries out decision again later.The related data that user is tested before being typically based in the prior art, utilizes the time
The prediction techniques such as sequence predict full dose user's using effect data, and are judged accordingly.But due to new product or newly
Business on-line time is shorter, and it is less and data fluctuations are larger to test number of users early period, is difficult pair by traditional prediction technique
Effect when opening to other users is precisely predicted, so that data forecasting accuracy is poor, it is also possible to decision be caused to be lost
Accidentally the case where.
In view of the above problem, in this embodiment, propose a kind of behavioral data prediction meanss, the device by for
The analysis of behavior similarity between object of action and historical behavior object to be predicted, it is corresponding using the similarity for meeting preset condition
Historical behavior object predictive behavior data of the historical behavior data as object of action to be predicted, realized with this for pre-
The accurate prediction of object of action behavioral data is surveyed, and then improves the accuracy of data prediction, it is unstable to reduce test data early period
The case where bring perplexs, and avoids the occurrence of incorrect decision.
In an optional implementation of the present embodiment, the object of action refers to implementing a certain specified object
The user of a certain behavior, wherein the specified object can for new product, new demand servicing, new business or new projects, with new product,
New demand servicing, similar product, service, business or the project or other appointed products of new business or new projects, service, business or
Project；The behavioral data refers to that a certain specified object occurs for object of action the number of a certain default behavior.For example, if
The specified object is a new demand servicing, then the object of action can be the user using the new demand servicing, the history row
It can be the user that the new demand servicing was used in default historical time section, the historical behavior of the historical behavior object for object
Data can be the number that the user uses the new demand servicing in the default historical time section, and the behavior pair to be predicted
As being exactly in default future time section there is a possibility that with the user of the new demand servicing, the predictive behavior of the object of action to be predicted
Data be exactly the user that is predicted according to above-mentioned historical data in the default future time section there is a possibility that new with this
The number of service.
Wherein, for the standard uniformly compared, accurately prediction data, the feature vector of the historical behavior object are obtained
It is corresponding identical as the feature vector element of the object of action to be predicted.
In an optional implementation of the present embodiment, the feature vector of the object of action may include in following characteristics
It is one or more: the behavioural characteristic of the attributive character of the object of action and the object of action, wherein the object of action
Attributive character may include one of following element or a variety of: age, gender, occupation, education level, marital status etc.
Deng the behavioural characteristic of the object of action may include one of following element or a variety of: the view rate of specified object, specified pair
Whether the click volume of elephant, the forward rate of specified object were bought and specify object, the amount of placing an order of specified object etc..Certainly, described
Feature vector may also comprise other characteristic elements, those skilled in the art can according to the needs of practical application, the spy of object of action
Point is selected, and the present invention is not especially limited it.
In an optional implementation of the present embodiment, as shown in fig. 6, the acquisition module 501 includes:
First determines submodule 601, is configured to determine that feature vector element；
First acquisition submodule 602 is configured as obtaining the historical behavior object according to determining feature vector element
Feature vector；
Second acquisition submodule 603 is configured as obtaining the historical behavior object going through in default historical time section
History behavioral data.
In this embodiment, first determine submodule 601 according to the needs of practical application and the characteristics of object of action,
The characteristics of specified object, determines feature vector element；First acquisition submodule 602 is according to determining feature vector element extraction
The feature vector of the historical behavior object；Second acquisition submodule 603 obtains the historical behavior object in default history
Between historical behavior data in section.
Wherein, the default historical time section can determine according to the needs of practical application, and the present invention does not make it specifically
It limits.
Assuming that the behavioral data is the new demand servicing order amount of user, when the default historical time section is that new demand servicing is tested
Between section, then historical behavior data of the historical behavior object in default historical time section are exactly that the user tests in new demand servicing
The new demand servicing order amount generated in period.
In an optional implementation of the present embodiment, the acquisition module 501 further includes for the historical behavior
The part that data are handled, i.e., as shown in fig. 7, the acquisition module 501 includes:
First determines submodule 701, is configured to determine that feature vector element；
First acquisition submodule 702 is configured as obtaining the historical behavior object according to determining feature vector element
Feature vector；
Second acquisition submodule 703 is configured as obtaining the historical behavior object going through in default historical time section
History behavioral data；
First computational submodule 704 is configured as calculating the historical behavior object putting down in default historical time section
Equal behavioral data；
Classify submodule 705, is configured as the historical behavior pair based on the difference between the average behavioral data
As being divided into N class, wherein N is integer；
Second computational submodule 706 is configured as calculating the integration characteristics vector of historical behavior object described in every one kind
And averaged historical behavioral data in class；
Second determines submodule 707, is configured as the integration characteristics vector being determined as the historical behavior object
Averaged historical behavioral data in the class is determined as the historical behavior data of the historical behavior object by feature vector.
In order to embody the statistical property of behavioral data, the behavioral data is more reasonably used, in this embodiment, also
Cluster average treatment is carried out for the historical behavior data, specifically, the second acquisition submodule 703 obtains the historical behavior
For object after the historical behavior data in default historical time section, the first computational submodule 704 calculates each historical behavior pair
As the average behavioral data in default historical time section, for example, if the behavioral data is the new demand servicing order amount of user, institute
Stating default historical time section is new demand servicing testing time section, then the average behavioral data is exactly the user in new demand servicing test
Between average daily new demand servicing order amount in section；Classifying submodule 705 will be described based on the difference between the average behavioral data
Historical behavior object is divided into N class, wherein N is integer；Second computational submodule 706 calculates historical behavior pair described in every one kind
Averaged historical behavioral data in the integration characteristics vector and class of elephant；Second determines submodule 707 by the integration characteristics vector
It is determined as the feature vector of the historical behavior object, averaged historical behavioral data in the class is determined as the historical behavior
The historical behavior data of object, it is subsequent that the corresponding data of itself and the object of action to be predicted is subjected to similarity calculation.
In an optional implementation of the present embodiment, the historical behavior object is being divided into N by classification submodule 705
When class, clustering method, such as k-means clustering method can be used, naturally it is also possible to use other classification methods.When use k-
When means clustering method, ancon rule can be used to determine categorical measure K, i.e., first obtain the corresponding cost function of different K values,
With the increase of K value, clustering bring average distortion degree can become smaller, and the number of samples in every one kind can reduce therewith, but with
The increase of K value, it may appear that an inflection point, after inflection point, even if K value increases again, average distortion degree also no longer becomes smaller, at this
In one optional implementation of embodiment, K value corresponding to this inflection point can be taken as the class in k-means clustering method
Other quantity K.
In an optional implementation of the present embodiment, classifies submodule 705 when carrying out k-means cluster, can borrow
Euclidean distance is helped to evaluate the difference between the average behavioral data, with the other division of implementation of class.
In an optional implementation of the present embodiment, the second computational submodule 706 is gone through described in the every one kind of calculating
When the integration characteristics vector of history object of action, the corresponding eigenvalue in different historical behavior characteristics of objects vectors can be sought average
It is worth, then the corresponding average value of each characteristic value is formed to the integration characteristics vector of historical behavior object in such.
For example, if the historical behavior object is divided into 3 classes, each historical behavior by above-mentioned k-means clustering method
The feature vector of object includes 5 characteristic values, it is assumed that includes 6 historical behavior objects in the 1st class, then each going through in the 1st class
The feature vector of history object of action can indicate are as follows: object 1 { feature 11, feature 12, feature 13, feature 14, feature 15 }, object
2 { feature 21, feature 22, feature 23, feature 24, features 25 }, object 3 { feature 31, feature 32, feature 33, feature 34, feature
35 }, object 4 { feature 41, feature 42, feature 43, feature 44, feature 45 }, object 5 { feature 51, feature 52, feature 53, feature
54, feature 55 }, object 6 { feature 61, feature 62, feature 63, feature 64, feature 65 }, then after feature integration, in the 1st class
The integration characteristics vector of the historical behavior object can indicate are as follows: integrating vector, { feature 1, feature 2, feature 3, feature 4 are special
Sign 5 }, wherein feature 1=(feature 11+ feature 21+ feature 31+ feature 41+ feature 51+ feature 61)/6, feature 2=(feature 12
+ feature 22+ feature 32+ feature 42+ feature 52+ feature 62)/6, (feature 13+ feature 23+ feature 33+ feature 43+ is special by feature 3=
Levy 53+ feature 63)/6, feature 4=(feature 14+ feature 24+ feature 34+ feature 44+ feature 54+ feature 64)/6, feature 5=
(feature 15+ feature 25+ feature 35+ feature 45+ feature 55+ feature 65)/6.The integration characteristics of historical behavior object in other classes
The calculating of vector can and so on, the present invention repeats no more.
In an optional implementation of the present embodiment, the second computational submodule 706 is gone through described in the every one kind of calculating
In the class of history object of action when averaged historical behavioral data, the average method of data can also be used, i.e., gone through all in certain one kind
The historical behavior data of history object of action are averaged, and averaged historical behavioral data in the class can be obtained.
In an optional implementation of the present embodiment, the calculating of subsequent similarity for convenience, the acquisition module
501 further include:
Submodule is normalized, is configured as that the average behavioral data is normalized.
Wherein, those skilled in the art can according to the needs of practical application and the characteristics of average behavioral data selects normalizing
Change mode, the present invention are not especially limited it.
In an optional implementation of the present embodiment, the normalization submodule can be using following formula for described average
Behavioral data is normalized:
Y=(col-min (col))/(max (col)-min (col)),
Wherein, y indicates the average behavioral data that a certain user obtains after normalized, col indicate the user without
The average behavioral data of normalized, min (col) indicate it is similar in average behavior number of all users without normalized
Minimum value in, max (col) indicate it is similar in all users without the maximum in the average behavioral data of normalized
Value.
In addition, if the average behavior data packet of a certain user includes two or more subdatas, different user
The subdata of corresponding position is corresponding in average behavioral data, then needing useful for institute when being normalized
The corresponding subdata in family is normalized respectively.For example, the average behavior subdata of each user can be in line, that
The average behavior subdata of multiple users just constitutes a data matrix after being arranged successively, wherein in the data matrix
Every column data is exactly the subdata of corresponding different user, every column data can be returned respectively according to above formula at this time
One changes, wherein col can indicate a certain user without the average behavior subdata on a certain column of normalized, min
(col) minimum value in the column in all average behavior subdatas without normalized is indicated, max (col) indicates the column
In maximum value in all average behavior subdatas without normalized, finally can be obtained by dimension and every in this way
The average behavioral data of a user is identical to normalize average behavioral data.
In an optional implementation of the present embodiment, as shown in figure 8, the computing module 502 includes:
Third acquisition submodule 801 is configured as obtaining the feature vector of the object of action to be predicted；
Encoding submodule 802 is configured as feature vector and the history row for the object of action to be predicted
Carry out pre-arranged code processing for the feature vector of object, the corresponding coding characteristic vector for obtaining the object of action to be predicted and
The coding characteristic vector of the historical behavior object；
Third computational submodule 803, be configured as calculating the coding characteristic vector of the object of action to be predicted with it is described
Similarity between the coding characteristic vector of historical behavior object.
In order to facilitate the calculating of similarity between feature vector, in this embodiment, also for the behavior to be predicted
The feature vector of object and the feature vector of the historical behavior object carry out pre-arranged code processing, then calculate what correspondence obtained
The coding characteristic vector of the coding characteristic vector of the object of action to be predicted and the historical behavior object.
In an optional implementation of the present embodiment, encoding submodule 802 uses one-hot coding (One-Hot
Encoding, One-Hot) come to the feature of the feature vector of the object of action to be predicted and the historical behavior object to
Amount carries out coded treatment.One-Hot coding is also known as an efficient coding, and method is using N bit status register come to N number of
State is encoded, and each state has its independent register-bit, and when any, and only wherein one is effectively,
That is for each feature, if it has m probable value, after one-hot coding, the corresponding characteristic value of this feature
M binary feature value, also, these feature mutual exclusions are reformed into, only one each value is active.
In an optional implementation of the present embodiment, the behavior to be predicted is calculated in third computational submodule 803
When similarity between the coding characteristic vector of object and the coding characteristic vector of the historical behavior object, it can be used more than angle
String value similarity evaluation mode calculates the cosine value for the angle that two feature vectors are formed, obtained cosine value range of results
Between [- 1,1], the size of the cosine value can be used to evaluate the similarity degree between two feature vectors: angle is smaller,
Cosine value just illustrates that the direction of two feature vectors more coincide closer to 1, then two feature vectors are more similar, conversely,
Angle is bigger, and cosine value just illustrates that the direction difference of two feature vectors is bigger closer to -1, then two feature vectors are just
It is more dissimilar.
Wherein, the cosine value for the angle that two feature vectors are formed can be represented by the following formula:
Wherein, θ indicates that the angle that two feature vectors are formed, cos (θ) indicate the cosine value of the angle, xiAnd yiTable respectively
Show the ith feature value in two feature vectors, n indicates the quantity of characteristic value in two feature vectors.
In an optional implementation of the present embodiment, the determining module 503 is configured as:
The historical behavior data of the corresponding historical behavior object of highest similarity are determined as the object of action to be predicted
Predictive behavior data.
In this embodiment, it is believed that the behavior with the most similar historical behavior object of object of action to be predicted is practised
Used closest with the behavioural habits of the object of action to be predicted, therefore, the historical behavior data of the historical behavior object can
By as the behavioral data predicted for the object of action to be predicted.
It is subsequent, if it is desired, the predictive behavior data of all object of action to be predicted can be added, can be obtained described pre-
If the predictive behavior total data in future time, which can provide the reference in data for technical staff, to help it to make
Corresponding behaviour decision making.
The embodiment of the invention also discloses a kind of electronic equipment, Fig. 9 shows electronics according to an embodiment of the present invention and sets
Standby structural block diagram, as shown in figure 9, the electronic equipment 900 includes memory 901 and processor 902；Wherein,
The memory 901 is for storing one or more computer instruction, wherein one or more computer refers to
It enables and being executed by the processor 902 to realize any of the above-described method and step.
Figure 10 is suitable for being used to realize the knot of the computer system of the behavioral data prediction technique of embodiment according to the present invention
Structure schematic diagram.
As shown in Figure 10, computer system 1000 include central processing unit (CPU) 1001, can according to be stored in only
It reads the program in memory (ROM) 1002 or is loaded into random access storage device (RAM) 1003 from storage section 1008
Program and execute the various processing in above embodiment.In RAM1003, be also stored with system 1000 operate it is required various
Program and data.CPU1001, ROM1002 and RAM1003 are connected with each other by bus 1004.Input/output (I/O) interface
1005 are also connected to bus 1004.
I/O interface 1005 is connected to lower component: the importation 1006 including keyboard, mouse etc.；Including such as cathode
The output par, c 1007 of ray tube (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage section including hard disk etc.
1008；And the communications portion 1009 of the network interface card including LAN card, modem etc..Communications portion 1009 passes through
Communication process is executed by the network of such as internet.Driver 1010 is also connected to I/O interface 1005 as needed.It is detachable to be situated between
Matter 1011, such as disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 1010, so as to
In being mounted into storage section 1008 as needed from the computer program read thereon.
Particularly, embodiment according to the present invention, method as described above may be implemented as computer software programs.
For example, embodiments of the present invention include a kind of computer program product comprising be tangibly embodied in and its readable medium on
Computer program, the computer program includes program code for executing the behavioral data prediction technique.In this way
Embodiment in, which can be downloaded and installed from network by communications portion 1009, and/or from removable
Medium 1011 is unloaded to be mounted.
Flow chart and block diagram in attached drawing illustrate system, method and computer according to the various embodiments of the present invention
The architecture, function and operation in the cards of program product.In this regard, each box in course diagram or block diagram can be with
A part of a module, section or code is represented, a part of the module, section or code includes one or more
Executable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in box
The function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practical
On can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it wants
It is noted that the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart, Ke Yiyong
The dedicated hardware based system of defined functions or operations is executed to realize, or can be referred to specialized hardware and computer
The combination of order is realized.
Being described in unit or module involved in embodiment of the present invention can be realized by way of software, can also
It is realized in a manner of through hardware.Described unit or module also can be set in the processor, these units or module
Title do not constitute the restriction to the unit or module itself under certain conditions.
As on the other hand, the embodiment of the invention also provides a kind of computer readable storage mediums, this is computer-readable
Storage medium can be computer readable storage medium included in device described in above embodiment；It is also possible to individually
In the presence of without the computer readable storage medium in supplying equipment.Computer-readable recording medium storage has one or one
Procedure above, described program are used to execute the method for being described in the embodiment of the present invention by one or more than one processor.
Above description is only presently preferred embodiments of the present invention and the explanation to institute's application technology principle.Those skilled in the art
Member is it should be appreciated that invention scope involved in the embodiment of the present invention, however it is not limited to which the specific combination of above-mentioned technical characteristic forms
Technical solution, while should also cover in the case where not departing from the inventive concept, by above-mentioned technical characteristic or its equivalent spy
Levy the other technical solutions for carrying out any combination and being formed.Such as features described above with it is (but unlimited disclosed in the embodiment of the present invention
In) technical characteristic with similar functions is replaced mutually and the technical solution that is formed.
Claims (14)
1. a kind of behavioral data prediction technique characterized by comprising
Obtain the feature vector and its historical behavior data of historical behavior object；
The feature vector of object of action to be predicted is obtained, and calculates the feature vector and the history of the object of action to be predicted
Similarity between the feature vector of object of action；
The historical behavior data of the corresponding historical behavior object of the similarity for meeting preset condition are determined as the row to be predicted
For the predictive behavior data of object.
2. the method according to claim 1, wherein it is described obtain historical behavior object feature vector and its go through
History behavioral data, comprising:
Determine feature vector element；
The feature vector of the historical behavior object is obtained according to determining feature vector element；
Obtain historical behavior data of the historical behavior object in default historical time section.
3. according to the method described in claim 2, it is characterized in that, described obtain the historical behavior object in default history
Between after historical behavior data in section, further includes:
Calculate average behavioral data of the historical behavior object in default historical time section；
The historical behavior object is divided into N class based on the difference between the average behavioral data, wherein N is integer；
Calculate averaged historical behavioral data in the integration characteristics vector and class of historical behavior object described in every one kind；
The integration characteristics vector is determined as to the feature vector of the historical behavior object, by averaged historical behavior in the class
Data are determined as the historical behavior data of the historical behavior object.
4. according to the method described in claim 3, it is characterized in that, described calculate the historical behavior object in default history
Between after average behavioral data in section, further includes:
The average behavioral data is normalized.
5. method according to claim 1 to 4, which is characterized in that the feature for obtaining object of action to be predicted to
Amount, and calculate similar between the feature vector of the object of action to be predicted and the feature vector of the historical behavior object
Degree, comprising:
Obtain the feature vector of the object of action to be predicted；
The feature vector of feature vector and the historical behavior object for the object of action to be predicted carries out default volume
Code processing, the coding characteristic of the corresponding coding characteristic vector for obtaining the object of action to be predicted and the historical behavior object
Vector；
It calculates between the coding characteristic vector of the object of action to be predicted and the coding characteristic vector of the historical behavior object
Similarity.
6. -5 any method according to claim 1, which is characterized in that the similarity that will meet preset condition is corresponding
The historical behavior data of historical behavior object be determined as the predictive behavior data of the object of action to be predicted, comprising:
The historical behavior data of the corresponding historical behavior object of highest similarity are determined as the pre- of the object of action to be predicted
Survey behavioral data.
7. a kind of behavioral data prediction meanss characterized by comprising
Module is obtained, is configured as obtaining the feature vector and its historical behavior data of historical behavior object；
Computing module is configured as obtaining the feature vector of object of action to be predicted, and calculates the object of action to be predicted
Similarity between feature vector and the feature vector of the historical behavior object；
Determining module, the historical behavior data for being configured as to meet the corresponding historical behavior object of similarity of preset condition are true
It is set to the predictive behavior data of the object of action to be predicted.
8. device according to claim 7, which is characterized in that the acquisition module includes:
First determines submodule, is configured to determine that feature vector element；
First acquisition submodule, be configured as being obtained according to determining feature vector element the feature of the historical behavior object to
Amount；
Second acquisition submodule is configured as obtaining historical behavior number of the historical behavior object in default historical time section
According to.
9. device according to claim 8, which is characterized in that the acquisition module further include:
First computational submodule is configured as calculating average behavior number of the historical behavior object in default historical time section
According to；
Classification submodule is configured as that the historical behavior object is divided into N based on the difference between the average behavioral data
Class, wherein N is integer；
Second computational submodule is configured as calculating in integration characteristics vector and the class of historical behavior object described in every one kind
Averaged historical behavioral data；
Second determines submodule, be configured as the integration characteristics vector being determined as the feature of the historical behavior object to
Averaged historical behavioral data in the class, is determined as the historical behavior data of the historical behavior object by amount.
10. device according to claim 9, which is characterized in that the acquisition module further include:
Submodule is normalized, is configured as that the average behavioral data is normalized.
11. according to any device of claim 7-10, which is characterized in that the computing module includes:
Third acquisition submodule is configured as obtaining the feature vector of the object of action to be predicted；
Encoding submodule is configured as feature vector and the historical behavior object for the object of action to be predicted
Feature vector carries out pre-arranged code processing, the corresponding coding characteristic vector for obtaining the object of action to be predicted and the history
The coding characteristic vector of object of action；
Third computational submodule is configured as calculating the coding characteristic vector of the object of action to be predicted and the historical behavior
Similarity between the coding characteristic vector of object.
12. according to any device of claim 7-11, which is characterized in that the determining module is configured as:
The historical behavior data of the corresponding historical behavior object of highest similarity are determined as the pre- of the object of action to be predicted
Survey behavioral data.
13. a kind of electronic equipment, which is characterized in that including memory and processor；Wherein,
The memory is for storing one or more computer instruction, wherein one or more computer instruction is by institute
Processor is stated to execute to realize method and step described in any one of claims 1-6.
14. a kind of computer readable storage medium, is stored thereon with computer instruction, which is characterized in that the computer instruction quilt
Processor realizes method and step described in any one of claims 1-6 when executing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811410240.3A CN110008974A (en) | 2018-11-23 | 2018-11-23 | Behavioral data prediction technique, device, electronic equipment and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811410240.3A CN110008974A (en) | 2018-11-23 | 2018-11-23 | Behavioral data prediction technique, device, electronic equipment and computer storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110008974A true CN110008974A (en) | 2019-07-12 |
Family
ID=67164985
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811410240.3A Pending CN110008974A (en) | 2018-11-23 | 2018-11-23 | Behavioral data prediction technique, device, electronic equipment and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110008974A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111026956A (en) * | 2019-11-20 | 2020-04-17 | 拉扎斯网络科技（上海）有限公司 | Data list processing method and device, electronic equipment and computer storage medium |
| WO2023050534A1 (en) * | 2021-09-29 | 2023-04-06 | 广州新科佳都科技有限公司 | Method and device for predicting energy consumption of rail transit station apparatus, and apparatus and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103823908A (en) * | 2014-03-21 | 2014-05-28 | 北京飞流九天科技有限公司 | Method and server for content recommendation on basis of user preferences |
| CN106375369A (en) * | 2016-08-18 | 2017-02-01 | 南京邮电大学 | Mobile Web service recommendation method and collaborative recommendation system based on user behavior analysis |
| CN107247753A (en) * | 2017-05-27 | 2017-10-13 | 深圳大学 | A kind of similar users choosing method and device |
| CN104809243B (en) * | 2015-05-15 | 2017-11-17 | 东南大学 | It is a kind of that method is recommended based on the mixing excavated to user behavior composite factor |
| 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 |
-
2018
- 2018-11-23 CN CN201811410240.3A patent/CN110008974A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103823908A (en) * | 2014-03-21 | 2014-05-28 | 北京飞流九天科技有限公司 | Method and server for content recommendation on basis of user preferences |
| CN104809243B (en) * | 2015-05-15 | 2017-11-17 | 东南大学 | It is a kind of that method is recommended based on the mixing excavated to user behavior composite factor |
| CN106375369A (en) * | 2016-08-18 | 2017-02-01 | 南京邮电大学 | Mobile Web service recommendation method and collaborative recommendation system based on user behavior analysis |
| CN107247753A (en) * | 2017-05-27 | 2017-10-13 | 深圳大学 | A kind of similar users choosing method and device |
| 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 |
Non-Patent Citations (3)
| Title |
|---|
| 潘书敏等: "基于用户相似度和特征分化的广告点击率预测研究", 《计算机科学》 * |
| 葛晓琳: "考虑需求响应不确定性的电动汽车负荷Sigmoid云模型预测", 《中国电机工程学报》 * |
| 金倩倩等: "基于相似度分析的电力信息内网用户行为异常预警方法", 《计算机系统应用》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111026956A (en) * | 2019-11-20 | 2020-04-17 | 拉扎斯网络科技（上海）有限公司 | Data list processing method and device, electronic equipment and computer storage medium |
| WO2023050534A1 (en) * | 2021-09-29 | 2023-04-06 | 广州新科佳都科技有限公司 | Method and device for predicting energy consumption of rail transit station apparatus, and apparatus and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11488055B2 (en) | Training corpus refinement and incremental updating | |
| Vorobiev et al. | Towards the theory of fuzzy differential equations | |
| CN108280477B (en) | Method and apparatus for clustering images | |
| US11475235B2 (en) | Clustering techniques for machine learning models | |
| CN110909222B (en) | User portrait establishing method and device based on clustering, medium and electronic equipment | |
| US20230297847A1 (en) | Machine-learning techniques for factor-level monotonic neural networks | |
| CN109408641A (en) | It is a kind of based on have supervision topic model file classification method and system | |
| CN111783039B (en) | Risk determination method, risk determination device, computer system and storage medium | |
| CN110019790A (en) | Text identification, text monitoring, data object identification, data processing method | |
| US20230021338A1 (en) | Conditionally independent data generation for training machine learning systems | |
| CN115809887B (en) | Method and device for determining main business scope of enterprise based on invoice data | |
| Zheng et al. | Identification of grey sheep users by histogram intersection in recommender systems | |
| CN110008974A (en) | Behavioral data prediction technique, device, electronic equipment and computer storage medium | |
| CN111475541A (en) | Data decision method and device, electronic equipment and storage medium | |
| Afuecheta et al. | Flexible models for stock returns based on Student's t distribution | |
| KR20070092727A (en) | Feature reduction method for decision machines | |
| Sun | Linear regression with interval‐valued data | |
| CN111209403A (en) | Data processing method, device, medium and electronic equipment | |
| CN112446777A (en) | Credit evaluation method, device, equipment and storage medium | |
| CN110309278A (en) | Keyword retrieval method, apparatus, medium and electronic equipment | |
| US11983207B2 (en) | Method, electronic device, and computer program product for information processing | |
| WO2022141838A1 (en) | Model confidence analysis method and apparatus, electronic device and computer storage medium | |
| CN114529191A (en) | Method and apparatus for risk identification | |
| CN112115316B (en) | Box separation method and device, electronic equipment and storage medium | |
| CN113554307A (en) | RFM (recursive filter) model-based user grouping method and device and readable medium |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200918 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Innovative advanced technology Co.,Ltd. Address before: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant before: Advanced innovation technology Co.,Ltd. Effective date of registration: 20200918 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Advanced innovation technology Co.,Ltd. Address before: A four-storey 847 mailbox in Grand Cayman Capital Building, British Cayman Islands Applicant before: Alibaba Group Holding Ltd. |
|
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190712 |