CN109241133A - Data monitoring method, calculates equipment and storage medium at device - Google Patents
Data monitoring method, calculates equipment and storage medium at device Download PDFInfo
- Publication number
- CN109241133A CN109241133A CN201810923329.3A CN201810923329A CN109241133A CN 109241133 A CN109241133 A CN 109241133A CN 201810923329 A CN201810923329 A CN 201810923329A CN 109241133 A CN109241133 A CN 109241133A
- Authority
- CN
- China
- Prior art keywords
- curve
- user activity
- user
- category
- data
- 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
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
This specification provides a kind of data monitoring method, device, calculates equipment and storage medium, which comprises obtains the liveness data relevant to the first setting date and liveness data relevant with the second setting date of target object;The first curve category of the first user activity curve and the second curve category of second user liveness curve are determined according to the first user activity data and second user liveness data;State transition probability between first curve category and second curve category is calculated according to state transition model;According to the state transition probability between first curve category and second curve category, preset threshold value between the state transition probability and first curve category and second curve category is compared, in the case where the state transition probability is lower than threshold value, alert notification is sent.
Description
Technical field
This specification is related to data monitoring technical field, in particular to a kind of data monitoring method, device, calculate equipment and
Storage medium.
Background technique
With the development of big data, each company is more and more deep to the understanding of data, we can generate largely daily
Data are in analysis and decision, in this course, manpower to be limited always.Therefore for the need of the system of garbled data
Asking just will appear.It is for way common in terms of data screening at present, to some key values of data configuration as threshold value, if
In the case that the state transition probability is lower than threshold value, alert notification is sent.
Summary of the invention
In view of this, this specification embodiment provides a kind of data monitoring method, device, calculates equipment and storage is situated between
Matter, to solve technological deficiency existing in the prior art.
According to this specification embodiment in a first aspect, providing a kind of data monitoring method, comprising:
It obtains the first user activity data relevant to the first setting date of target object and sets day with second
Phase relevant second user liveness data;
The first user activity curve is obtained according to the first user activity data and first user is active
Write music the first change to attributes feature of line, according to the second user liveness data obtain second user liveness curve and
Second change to attributes feature of the second user liveness curve;
It is living that first user is obtained based on the first change to attributes characteristic use user activity curve classification model
First curve category of jerk diagram, based on user activity curve classification model described in the second change to attributes characteristic use
Obtain the second curve category of the second user liveness curve;
It is general that state transfer between first curve category and second curve category is calculated according to state transition model
Rate;
According to the state transition probability between first curve category and second curve category, the state is turned
It moves probability to be compared with the preset threshold value between first curve category and second curve category, described
In the case that state transition probability is lower than threshold value, alert notification is sent.
Optionally, the first user activity data relevant to the first setting date for obtaining target object include:
From the user activity number extracted in journal file in the preset time section that target object sets the date first
According to.
Optionally, the second user liveness data relevant to the second setting date for obtaining target object include:
From the user activity number extracted in journal file in the preset time section that target object sets the date second
According to.
Optionally, first curve category, second curve category respectively include: grow steadily type, type of uprushing, and puts down
Slow type, counter-rotative type, preiodic type, rapidly ascending-type and dramatic decrease type.
Optionally, the first change to attributes feature, the second change to attributes feature respectively include: current value is default
The ranking in sample value in time interval, increment of the current value than proxima luce (prox. luc) numerical value, current value ratio maximum in addition to current value
Three numerical value average value increment, the growth of current value average value of the smallest three numerical value than in addition to current value
Amount, increment of the current value than average value, the number for the curve medium wave peak that user activity data are formed, user activity data
The number of trough in the curve of formation.
Optionally, the user enliven curve classification model as follows training obtain:
Obtain the training sample data collection that user enlivens curve classification model, wherein the training sample data concentrate packet
Include the corresponding user activity curve of user activity data and each preset time section in multiple preset time sections;
Change to attributes feature is constructed for the user activity data in each preset time section;
Preset model is trained by the user activity curve and the change to attributes feature, obtains described point
Class model, it is associated with the change to attributes feature that the disaggregated model makes the user enliven curve.
Optionally, the preset model is the Xgboost model using R language.
Optionally, the state transition model is Markov model.
Optionally, preparatory between the state transition probability and first curve category and second curve category
The threshold value of setting is compared, wherein the threshold value between every two curve categories is all different.
According to the second aspect of this specification embodiment, a kind of data monitoring device is provided, comprising:
Module is obtained, is configured as obtaining the first user activity data relevant to the first setting date of target object
And second user liveness data relevant to the second setting date;
Processing module, be configured as being obtained according to the first user activity data the first user activity curve and
First change to attributes feature of the first user activity curve, it is active according to the second user
Degree evidence obtains the second change to attributes of second user liveness curve and the second user liveness curve
Feature;
Categorization module is configured as obtaining based on the first change to attributes characteristic use user activity curve classification model
It is living based on user described in the second change to attributes characteristic use to the first curve category of the first user activity curve
Jerk diagram disaggregated model obtains the second curve category of the second user liveness curve;
Computing module is configured as calculating first curve category and second class of a curve according to state transition model
State transition probability between not.
Notification module is configured as according to the state transfer between first curve category and second curve category
Probability, by the preset threshold between the state transition probability and first curve category and second curve category
Value is compared, and in the case where the state transition probability is lower than threshold value, sends alert notification.
According to the third aspect of this specification embodiment, a kind of calculating equipment is provided, including memory, processor and deposit
The computer instruction that can be run on a memory and on a processor is stored up, the processor realizes above-mentioned side when executing described instruction
The step of method.
According to the fourth aspect of this specification embodiment, a kind of computer readable storage medium is provided, computer is contained
The step of instruction, which realizes above-mentioned method when being executed by processor.
This application provides a kind of data monitoring methods, obtain target object the first user relevant to the first setting date
Liveness data and second set the active data of date relevant second user, according to the first user activity data and second
It is bent that user activity data respectively obtain the first user activity curve, second user liveness curve, the first user activity
The corresponding first change to attributes feature of line and the corresponding second change to attributes feature of second user liveness curve, according to described
One change to attributes characteristic use disaggregated model obtains the first curve category, and the second change to attributes characteristic use disaggregated model obtains
To the second curve category, classification between the first curve category and the second curve category is calculated by state transition model and is shifted
Probability is monitored the lesser probability being calculated, and the application utilizes machine learning algorithm, by between different curves
Classification transition probability set different threshold values, realize the monitoring to Various types of data development trend.
Detailed description of the invention
Fig. 1 is a kind of structural schematic diagram for data monitoring system that one embodiment of this specification provides;
Fig. 2 is the flow chart of one of one embodiment of this specification data monitoring method;
Fig. 3 is one of one embodiment of this specification user activity curve graph;
Fig. 4 is one of one embodiment of this specification user activity curve graph;
Fig. 5 is one of one embodiment of this specification data monitoring device module map.
Specific embodiment
Many details are explained in the following description in order to fully understand this specification.But this specification energy
Enough to be implemented with being much different from other way described herein, those skilled in the art can be without prejudice to this specification intension
In the case where do similar popularization, therefore this specification is not limited by following public specific implementation.
Below by specific embodiment, the present invention is described in detail.
With reference to Fig. 1, Fig. 1 is a kind of structural schematic diagram for data monitoring system that one embodiment of this specification provides, and is being situated between
Before the technical solution for the application that continues, the framework of data monitoring system involved in the application is illustrated first.
Fig. 1 is to show the structural schematic diagram of the data monitoring system of one embodiment of this specification.Including server-side 110,
Network 130 and terminal 120.
The component of the server-side 110 and the terminal 120 includes but is not limited to memory and processor.Processor with deposit
Reservoir is connected by bus, and database is for saving data.
Server-side 110 and the terminal 120 further include access device, and access device makes server-side 110 and the terminal
120 can communicate via one or more networks 130.The example of these networks includes public switched telephone network (PSTN), local
Net (LAN), wide area network (WAN), personal area network (PAN) or such as internet communication network combination.
Access device may include wired or wireless any kind of network interface (for example, network interface card (NIC))
One or more of, such as IEEE802.11 WLAN (WLAN) wireless interface, worldwide interoperability for microwave accesses (Wi-
MAX) interface, Ethernet interface, universal serial bus (USB) interface, cellular network interface, blue tooth interface, near-field communication (NFC)
Interface, etc..
Fig. 2 shows the flow charts of one of one embodiment of this specification data monitoring method, are applied to server-side, such as
Shown in Fig. 2, including step 202 is to step 208.
Step 202: obtain target object to the first setting date relevant first user activity data and with the
Two setting dates relevant second user liveness data.
In a kind of embodiment of this specification, the target object can be obtained by the log of target object, described
Obtain target object includes: to the first setting date relevant first user activity data
From the user activity number extracted in journal file in the preset time section that target object sets the date first
According to.
For the network equipment, system and service routine etc., the logout for being log can be all generated in running, it should
Logout is the journal file for assessing object.Assess object journal file in every a line log all recite the date,
The description of the relevant operations such as time, user and movement.
In the embodiment of the present application, target object can be application program or Webpage, and the user of target object is active
Degree evidence can be pageview, downloading viewing amount or thumb up comment amount.
In this application, target object can be to be a variety of, such as the pageview in Taobao shop, wechat public platform article is clear
Downloading viewing amount, the pageview of microblogging small video and the amount of thumbing up etc. of file in the amount of looking at, Baidu library;The application will be multiple
Sample as a preset time section, according to the first setting date match set with first set the date match it is pre-
If time interval, after determining preset time section, the user activity of target object is extracted.
Described first set date of the preset time section on date to be monitored be contained in it is pre-set
Time range in preset time section, such as setting time section be 31 days, first 30 days for setting the date are exactly currently to set
The time interval fixed the date.
In a kind of embodiment of this specification, the second user relevant to the second setting date for obtaining target object is living
Jerk data include:
From the user activity number extracted in journal file in the preset time section that target object sets the date second
According to.
Described first set date of the preset time section on date to be monitored be contained in it is pre-set
Time range in preset time section.
Step 204: obtaining the first user activity curve and described first according to the first user activity data
First change to attributes feature of user activity curve obtains second user liveness according to the second user liveness data
Second change to attributes feature of curve and the second user liveness curve.
In a kind of embodiment of this specification, first curve category, second curve category respectively include: on steadily
The type of liter, type of uprushing, high order smooth pattern, counter-rotative type, preiodic type, rapidly ascending-type and dramatic decrease type;The application utilizes Xgboost pairs
User activity data are classified, and user activity data are divided into above 7 kinds of classifications, 7 kinds of above-mentioned classifications can cover
Most of form of data.
Xgboost is a kind of algorithm classified and returned by certain indexs of data, and Xgboost passes through decision tree
To the restriction in data carry out condition, the division by the index tree of restriction achievees the purpose that classification or recurrence, classification
It is obtaining the result is that discrete, the numerical value returned be it is continuous, return and classification essence be all feature to result or mark
Mapping between label, the sample output of regression tree are the forms of numerical value, and the sample output of classification tree is the form of class.Xgboost
The advantages of be that speed is fast, effect is good, be capable of handling large-scale data, support multilingual and support customized loss function etc.
Deng.The classic place Xgboost is exactly to support parallelization data processing, and direct effect is exactly that handle data speed fast.
In a kind of embodiment of this specification, the first change to attributes feature, the second change to attributes feature are wrapped respectively
It includes: ranking of the current value in the sample value in preset time section, increment of the current value than proxima luce (prox. luc) numerical value, current value ratio
The increment of the average value of maximum three numerical value in addition to current value, current value the smallest three numerical value than in addition to current value
The increment of average value, increment of the current value than average value, the number for the curve medium wave peak that user activity data are formed, use
The number of trough in the curve that family liveness data are formed;By data prediction part in the application, by user activity number
According to identifiable characteristic value is converted into, first curve category, second curve category include the first change described above
7 change to attributes features in change attributive character, the second attributive character.
Step 206: obtaining described based on the first change to attributes characteristic use user activity curve classification model
First curve category of one user activity curve, based on user activity curve described in the second change to attributes characteristic use
Disaggregated model obtains the second curve category of the second user liveness curve.
In a kind of embodiment of this specification, the user enlivens curve classification model, and training is obtained as follows:
Obtain the training sample data collection that user enlivens curve classification model, wherein the training sample data concentrate packet
Include the corresponding user activity curve of user activity data and each preset time section in multiple preset time sections;
Change to attributes feature is constructed for the user activity data in each preset time section;
Preset model is trained by the user activity curve and the change to attributes feature, obtains described point
Class model, it is associated with the change to attributes feature that the disaggregated model makes the user enliven curve.
In a kind of embodiment of this specification, the preset model is the Xgboost model using R language.
In this application, it is substituted into using the Xgboost model of the R language comprising the user in multiple preset time sections
The training sample data collection of user activity curve corresponding to liveness data and each preset time section, according to multiple pre-
If any active ues degree evidence in time interval, the user activity data in each preset time section are constructed into variation and are belonged to
Property feature, is trained Xgboost model by the change to attributes feature and the user activity curve, is trained
The data classification model of end.
Step 208: shape between first curve category and second curve category is calculated according to state transition model
State transition probability.
In a kind of embodiment of this specification, the state transition model is Markov model.
In this application, first curve category and second curve category have a curve different in 7 respectively, and every kind
State transition probability is different between curve category, the probability of state transfer between available 7*7 classification, can be to lesser
State transition probability is monitored between user activity curve category.
Step 210: according to the state transition probability between first curve category and second curve category, by institute
The preset threshold value stated between state transition probability and first curve category and second curve category is compared
Compared with, in the case where the state transition probability is lower than threshold value, transmission alert notification.
In a kind of embodiment of this specification, the state transition probability and first curve category and second curve
Preset threshold value between classification is compared, wherein the threshold value between every two curve categories is all different.
This application provides a kind of data monitoring methods, obtain target object the first user relevant to the first setting date
Liveness data and second set the active data of date relevant second user, according to the first user activity data and second
It is bent that user activity data respectively obtain the first user activity curve, second user liveness curve, the first user activity
The corresponding first change to attributes feature of line and the corresponding second change to attributes feature of second user liveness curve, according to described
One change to attributes characteristic use disaggregated model obtains the first curve category, and the second change to attributes characteristic use disaggregated model obtains
To the second curve category, classification between the first curve category and the second curve category is calculated by state transition model and is shifted
Probability is monitored the lesser probability being calculated, and the application utilizes machine learning algorithm, by between different curves
Classification transition probability set different threshold values, realize the monitoring to Various types of data development trend.
Fig. 3 shows the user activity curve graph in the application one embodiment, as shown in figure 3, the curve graph is certain
Product abscissa in May 1 to the user activity curve graph on May 31, the curve graph is the date, and ordinate is the product
Newly-increased enrollment, May 31 was obtained into for the first setting date in preset time as the first setting date in the present embodiment
The first user activity data in section, wherein preset time section is 31 days, active according to default 31 days daily users
Degree evidence obtains the user activity curve graph in May 1 shown in Fig. 3 to May 31.
Table 1 shows user's change to attributes mark sheet in the application one embodiment.
Table 1
According to table 1, the sample value on May 31 arranges 12 in preset 31 days sample intervals, is proxima luce (prox. luc) sample
0.9703 times of this numerical value is 0.5218 times of maximum sample values average value on the three, is minimum three days sample values
2.2967 times, it is 1.0741 times of whole 31 days sample values average value, there are 4 waves in the curve in May 1 to May 31
Peak, is indicated with A, B, C and D in Fig. 3 respectively, is had 3 troughs in curve, is indicated respectively with E, F and G in Fig. 3, according to above-mentioned
The first change to attributes characteristic use user activity curve classification model of the first user activity curve obtain the first user
First curve category of liveness, in the present embodiment, the first curve category are preiodic type curve.
Fig. 4 shows the user activity curve graph in the application one embodiment, as shown in figure 4, the curve graph is certain
Product abscissa in May 2 to the user activity curve graph on June 1, the curve graph is the date, and ordinate is the product
Newly-increased enrollment, June 1 was obtained into for the second setting date in preset time as the second setting date in the present embodiment
Second user liveness data in section, wherein preset time section is 31 days, active according to default 31 days daily users
Degree evidence obtains the user activity curve graph in May 2 shown in Fig. 4 to June 1.
Table 2 shows user's change to attributes mark sheet in the application one embodiment.
Table 2
According to table 2, the sample value on June 1 arranges 19 in preset 31 days sample intervals, is proxima luce (prox. luc) sample
0.812 times of numerical value is 0.4237 times of maximum sample values average value on the three, is 1.865 times of minimum sample values on the three,
It is 0.8696 times of whole 31 days sample values average value, there are 4 wave crests in the curve in May 2 to June 1, respectively with figure
H, I, J and K in 4 indicate there are 3 troughs in curve, indicated respectively with L, M and N in Fig. 4, according to above-mentioned second user
Second change to attributes characteristic use user activity curve classification model of liveness curve obtains the of second user liveness
Two curve categories, in the present embodiment, the second curve category are preiodic type curve.
By calculating a large amount of data, passed through according to the curve category of the curve category on May 31 and June 1
Markov model carries out the calculating of state transition probability, and the state that available preiodic type curve is transferred to preiodic type curve turns
Moving probability is 91%, and 91% and pre-set threshold value are compared, will be reminded lower than threshold value, on May 31 to 6
State transition probability is 91% between the curve on the moon 1, belongs to the event not less than threshold value, therefore thinks May 31 to June 1
State conversion does not need to cause to pay close attention to.
The embodiment of the present application provides a kind of data monitoring method, general by shifting between the state user activity curve
Rate is calculated, by the state transition probability of first curve category and the second curve category and pre-set threshold value into
Row compares, and is considered as normal condition more than pre-set threshold value, is considered as abnormal condition lower than pre-set threshold value, right
In the case that the state transition probability is lower than threshold value, alert notification is sent, and pay close attention to the case where being lower than threshold value, reached
The purpose of data monitoring.
Fig. 5 shows one of the embodiment of the present application data monitoring device module map, as shown in figure 5, the data monitoring
Device 500 includes obtaining module 502, processing module 504, categorization module 506, computing module 508 and notification module 510:
It obtains module 502: being configured as obtaining the first user activity relevant to the first setting date of target object
Data and second user liveness data relevant to the second setting date;
Processing module 504: it is configured as obtaining the first user activity curve according to the first user activity data
And the first change to attributes feature of the first user activity curve, is obtained according to the second user liveness data
Second change to attributes feature of two user activity curves and the second user liveness curve;
Categorization module 506: it is configured as based on the first change to attributes characteristic use user activity curve classification mould
Type obtains the first curve category of the first user activity curve, based on using described in the second change to attributes characteristic use
Family liveness curve classification model obtains the second curve category of the second user liveness curve;
Computing module 508: it is configured as calculating first curve category and second song according to state transition model
State transition probability between line classification.
Notification module 510: it is configured as according to the state between first curve category and second curve category
Transition probability, by presetting between the state transition probability and first curve category and second curve category
Threshold value be compared, the state transition probability be lower than threshold value in the case where, send alert notification.
In an optional embodiment, the acquisition module is configured to: target pair is extracted from journal file
As the user activity data in the preset time section on the first setting date.
In an optional embodiment, the acquisition module is configured to: target pair is extracted from journal file
As the user activity data in the preset time section on the second setting date.
In an optional embodiment, first curve category, second curve category respectively include: grow steadily
Type, type of uprushing, high order smooth pattern, counter-rotative type, preiodic type, rapidly ascending-type and dramatic decrease type.
In an optional embodiment, the first change to attributes feature, the second change to attributes feature respectively include:
Ranking of the current value in the sample value in preset time section, than the increment of proxima luce (prox. luc) numerical value, current value ratio removes current value
The increment of the average value of maximum three numerical value outside current value, current value the smallest three numerical value than in addition to current value are put down
The increment of mean value, increment of the current value than average value, the number for the curve medium wave peak that user activity data are formed, user
The number of trough in the curve that liveness data are formed.
In an optional embodiment, the curve classification module includes:
Acquisition submodule is configured as obtaining the training sample data collection that user enlivens curve classification model, wherein described
Training sample data concentration includes user activity data and each preset time section pair in multiple preset time sections
The user activity curve answered;
Building submodule: it is configured as constructing change to attributes spy for the user activity data in each preset time section
Sign;
Training submodule: it is configured as through the user activity curve and the change to attributes feature to preset model
It is trained, obtains the disaggregated model, the disaggregated model makes the user enliven curve and the change to attributes feature
It is associated.
In an optional embodiment, the preset model can use the Xgboost model of R language.
In an optional embodiment, the state transition model can be Markov model.
In an optional embodiment, the state transition probability and first curve category and second class of a curve
Preset threshold value between not is compared, wherein the threshold value between every two curve categories is all different.
The function of each unit and the realization process of effect are specifically detailed in the above method and correspond to step in above-mentioned apparatus
Realization process, details are not described herein.
For device embodiment, since it corresponds essentially to embodiment of the method, so related place is referring to method reality
Apply the part explanation of example.The apparatus embodiments described above are merely exemplary, wherein described be used as separation unit
The unit of explanation may or may not be physically separated, and component shown as a unit can be or can also be with
It is not physical unit, it can it is in one place, or may be distributed over multiple network units.It can be according to actual
The purpose for needing to select some or all of the modules therein to realize this specification scheme.Those of ordinary skill in the art are not
In the case where making the creative labor, it can understand and implement.
One embodiment of this specification also provides a kind of calculating equipment, including memory, processor and storage are on a memory
And the computer instruction that can be run on a processor, the processor realize the data monitoring method when executing described instruction
The step of.
One embodiment of this specification also provides a kind of computer readable storage medium, is stored with computer instruction, this refers to
Enable the step of data monitoring method is realized when being executed by processor.
A kind of exemplary scheme of above-mentioned computer readable storage medium for the present embodiment.It should be noted that this is deposited
The technical solution of storage media and the technical solution of above-mentioned automated testing method belong to same design, the technical side of storage medium
The detail content that case is not described in detail may refer to the description of the technical solution of above-mentioned automated testing method.
The computer instruction includes computer program code, the computer program code can for source code form,
Object identification code form, executable file or certain intermediate forms etc..The computer-readable medium may include: that can carry institute
State any entity or device, recording medium, USB flash disk, mobile hard disk, magnetic disk, CD, the computer storage of computer program code
Device, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory),
Electric carrier signal, telecommunication signal and software distribution medium etc..It should be noted that the computer-readable medium include it is interior
Increase and decrease appropriate can be carried out according to the requirement made laws in jurisdiction with patent practice by holding, such as in certain jurisdictions of courts
Area does not include electric carrier signal and telecommunication signal according to legislation and patent practice, computer-readable medium.
It should be noted that for the various method embodiments described above, describing for simplicity, therefore, it is stated as a series of
Combination of actions, but those skilled in the art should understand that, this specification is not limited by the described action sequence, because
For according to this specification, certain steps can use other sequences or carry out simultaneously.Secondly, those skilled in the art also should
Know, the embodiments described in the specification are all preferred embodiments, and related actions and modules might not all be this
Necessary to specification.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment
Point, it may refer to the associated description of other embodiments.
This specification preferred embodiment disclosed above is only intended to help to illustrate this specification.Alternative embodiment is not
All details of detailed descriptionthe, do not limit the invention to the specific embodiments described.Obviously, according in this specification
Hold, can make many modifications and variations.These embodiments are chosen and specifically described to this specification, is in order to preferably explain this theory
The principle and practical application of bright book, so that skilled artisan be enable to better understand and utilize this specification.This
Specification is limited only by the claims and their full scope and equivalents.
Claims (18)
1. a kind of data monitoring method characterized by comprising
It obtains the first user activity data relevant to the first setting date of target object and sets date phase with second
The second user liveness data of pass;
The first user activity curve is obtained according to the first user activity data and first user activity is bent
First change to attributes feature of line obtains second user liveness curve and described according to the second user liveness data
Second change to attributes feature of second user liveness curve;
First user activity is obtained based on the first change to attributes characteristic use user activity curve classification model
First curve category of curve is obtained based on user activity curve classification model described in the second change to attributes characteristic use
Second curve category of the second user liveness curve;
State transition probability between first curve category and second curve category is calculated according to state transition model;
According to the state transition probability between first curve category and second curve category, the state is shifted general
Preset threshold value between rate and first curve category and second curve category is compared, in the state
In the case that transition probability is lower than threshold value, alert notification is sent.
2. the method according to claim 1, wherein described obtain the related to the first setting date of target object
The first user activity data include:
From the user activity data extracted in journal file in the preset time section that target object sets the date first.
3. the method according to claim 1, wherein described obtain the related to the second setting date of target object
Second user liveness data include:
From the user activity data extracted in journal file in the preset time section that target object sets the date second.
4. the method according to claim 1, wherein first curve category, second curve category point
It does not include: the type that grows steadily, type of uprushing, high order smooth pattern, counter-rotative type, preiodic type, rapidly ascending-type and dramatic decrease type;
The first change to attributes feature, the second change to attributes feature respectively include: current value is in preset time section
Sample value in ranking, increment of the current value than proxima luce (prox. luc) numerical value, current value ratio maximum three numerical value in addition to current value
Average value increment, the increment of current value average value of the smallest three numerical value than in addition to current value, current value ratio
The increment of average value, the number for the curve medium wave peak that user activity data are formed, the curve that user activity data are formed
The number of middle trough.
5. the method according to claim 1, wherein the user enlivens curve classification model as follows
Training obtains:
Obtain the training sample data collection that user enlivens curve classification model, wherein it includes more that the training sample data, which are concentrated,
The corresponding user activity curve of user activity data and each preset time section in a preset time section;
Change to attributes feature is constructed for the user activity data in each preset time section;
Preset model is trained by the user activity curve and the change to attributes feature, it is living to obtain the user
Jump curve classification model, and the user enlivens curve classification model and the user is made to enliven curve and the change to attributes feature
It is associated.
6. according to the method described in claim 5, it is characterized in that, the preset model is the Xgboost mould using R language
Type.
7. the method according to claim 1, wherein the state transition model is Markov model.
8. the method according to claim 1, wherein the state transition probability and first curve category and
Preset threshold value between second curve category is compared, wherein the threshold value between every two curve categories is all
Different.
9. a kind of data monitoring device characterized by comprising
Obtain module, be configured as obtaining target object to the first setting date relevant first user activity data and
Second user liveness data relevant to the second setting date;
Processing module is configured as obtaining the first user activity curve and described according to the first user activity data
It is living to obtain second user according to the second user liveness data for first change to attributes feature of the first user activity curve
Second change to attributes feature of jerk diagram and the second user liveness curve;
Categorization module is configured as obtaining institute based on the first change to attributes characteristic use user activity curve classification model
The first curve category of the first user activity curve is stated, based on user activity described in the second change to attributes characteristic use
Curve classification model obtains the second curve category of the second user liveness curve;
Computing module, be configured as being calculated according to state transition model first curve category and second curve category it
Between state transition probability.
Notification module is configured as general according to the state transfer between first curve category and second curve category
Rate, by the preset threshold value between the state transition probability and first curve category and second curve category
It is compared, in the case where the state transition probability is lower than threshold value, sends alert notification.
10. device according to claim 9, which is characterized in that the acquisition module is configured to: from log text
User activity data of the target object in the preset time section on the first setting date are extracted in part.
11. device according to claim 9, which is characterized in that the acquisition module is configured to: from log text
User activity data of the target object in the preset time section on the second setting date are extracted in part.
12. device according to claim 9, which is characterized in that first curve category, second curve category point
It does not include: the type that grows steadily, type of uprushing, high order smooth pattern, counter-rotative type, preiodic type, rapidly ascending-type and dramatic decrease type;
The first change to attributes feature, the second change to attributes feature respectively include: current value is in preset time section
Sample value in ranking, increment of the current value than proxima luce (prox. luc) numerical value, current value ratio maximum three numerical value in addition to current value
Average value increment, the increment of current value average value of the smallest three numerical value than in addition to current value, current value ratio
The increment of average value, the number for the curve medium wave peak that user activity data are formed, the curve that user activity data are formed
The number of middle trough.
13. device according to claim 9, which is characterized in that the curve classification module includes:
Acquisition submodule is configured as obtaining the training sample data collection that user enlivens curve classification model, wherein the training
Sample data concentrate include multiple preset time sections in user activity data and each preset time section it is corresponding
User activity curve;
Building submodule: it is configured as constructing change to attributes feature for the user activity data in each preset time section;
Training submodule: it is configured as carrying out preset model by the user activity curve and the change to attributes feature
Training, obtains the disaggregated model, it is related to the change to attributes feature that the disaggregated model makes the user enliven curve
Connection.
14. device according to claim 13, which is characterized in that the preset model is the Xgboost mould using R language
Type.
15. device according to claim 9, which is characterized in that the state transition model is Markov model.
16. device according to claim 9, which is characterized in that the state transition probability and first curve category
Preset threshold value between second curve category is compared, wherein setting between every two curve categories in advance
Fixed threshold value is all different.
17. a kind of calculating equipment including memory, processor and stores the calculating that can be run on a memory and on a processor
The step of machine instruction, the processor realizes claim 1-8 described in any item methods when executing described instruction.
18. a kind of computer readable storage medium, is stored with computer instruction, which realizes right when being executed by processor
It is required that the step of 1-8 described in any item methods.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810923329.3A CN109241133A (en) | 2018-08-14 | 2018-08-14 | Data monitoring method, calculates equipment and storage medium at device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810923329.3A CN109241133A (en) | 2018-08-14 | 2018-08-14 | Data monitoring method, calculates equipment and storage medium at device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109241133A true CN109241133A (en) | 2019-01-18 |
Family
ID=65070246
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810923329.3A Pending CN109241133A (en) | 2018-08-14 | 2018-08-14 | Data monitoring method, calculates equipment and storage medium at device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109241133A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110032670A (en) * | 2019-04-17 | 2019-07-19 | 腾讯科技(深圳)有限公司 | Method for detecting abnormality, device, equipment and the storage medium of time series data |
CN112633573A (en) * | 2020-12-21 | 2021-04-09 | 北京达佳互联信息技术有限公司 | Prediction method of active state and determination method of activity threshold |
CN113590925A (en) * | 2020-04-30 | 2021-11-02 | 中国移动通信集团北京有限公司 | User determination method, device, equipment and computer storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102749589A (en) * | 2012-07-13 | 2012-10-24 | 哈尔滨工业大学深圳研究生院 | Recession-mode predicting method of power battery of electric automobile |
CN107451832A (en) * | 2016-05-30 | 2017-12-08 | 北京京东尚科信息技术有限公司 | The method and apparatus of pushed information |
-
2018
- 2018-08-14 CN CN201810923329.3A patent/CN109241133A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102749589A (en) * | 2012-07-13 | 2012-10-24 | 哈尔滨工业大学深圳研究生院 | Recession-mode predicting method of power battery of electric automobile |
CN107451832A (en) * | 2016-05-30 | 2017-12-08 | 北京京东尚科信息技术有限公司 | The method and apparatus of pushed information |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110032670A (en) * | 2019-04-17 | 2019-07-19 | 腾讯科技(深圳)有限公司 | Method for detecting abnormality, device, equipment and the storage medium of time series data |
CN110032670B (en) * | 2019-04-17 | 2022-11-29 | 腾讯科技(深圳)有限公司 | Method, device and equipment for detecting abnormity of time sequence data and storage medium |
CN113590925A (en) * | 2020-04-30 | 2021-11-02 | 中国移动通信集团北京有限公司 | User determination method, device, equipment and computer storage medium |
CN112633573A (en) * | 2020-12-21 | 2021-04-09 | 北京达佳互联信息技术有限公司 | Prediction method of active state and determination method of activity threshold |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110046952B (en) | Recommendation model training method and device, and recommendation method and device | |
CN109767255B (en) | Method for realizing intelligent operation and accurate marketing through big data modeling | |
CN107766929B (en) | Model analysis method and device | |
CN110008399A (en) | A kind of training method and device, a kind of recommended method and device of recommended models | |
CN106874522A (en) | Information recommendation method, device, storage medium and processor | |
CN108363821A (en) | A kind of information-pushing method, device, terminal device and storage medium | |
CN106779087A (en) | A kind of general-purpose machinery learning data analysis platform | |
CN109241133A (en) | Data monitoring method, calculates equipment and storage medium at device | |
CN106951925A (en) | Data processing method, device, server and system | |
CN110033342A (en) | A kind of training method and device, a kind of recommended method and device of recommended models | |
CN112153426B (en) | Content account management method and device, computer equipment and storage medium | |
CN104133817A (en) | Online community interaction method and device and online community platform | |
CN109165799A (en) | The class's of walking education course arrangement system based on genetic algorithm | |
TW202133005A (en) | Method and system for online data collection | |
Alshmrany | Adaptive learning style prediction in e-learning environment using levy flight distribution based CNN model | |
CN106846082A (en) | Tourism cold start-up consumer products commending system and method based on hardware information | |
CN110147389A (en) | Account number treating method and apparatus, storage medium and electronic device | |
CN109918574A (en) | Item recommendation method, device, equipment and storage medium | |
Gu et al. | [Retracted] Application of Fuzzy Decision Tree Algorithm Based on Mobile Computing in Sports Fitness Member Management | |
CN111368131B (en) | User relationship identification method and device, electronic equipment and storage medium | |
Maharani et al. | Weighted product method for selection of superior seeds catfish in the Clarias gariepinus types (Sangkuriang) | |
CN111368060B (en) | Self-learning method, device and system for conversation robot, electronic equipment and medium | |
CN108205727B (en) | Digital workshop process decision method based on decision tree and expert system | |
CN107844536A (en) | The methods, devices and systems of application program selection | |
CN112231299A (en) | Method and device for dynamically adjusting feature library |
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 | ||
CB02 | Change of applicant information |
Address after: Units F01-03 and 05-10 on the 6th floor of No.1 Building, No.8 Courtyard, Guangshun South Street, Chaoyang District, Beijing Applicant after: Beijing Ape Power Future Technology Co., Ltd. Address before: Units F01-03 and 05-10 on the 6th floor of No.1 Building, No.8 Courtyard, Guangshun South Street, Chaoyang District, Beijing Applicant before: Beijing chalk Future Technology Co., Ltd. |
|
CB02 | Change of applicant information |