CN109583949A - A kind of user changes planes prediction technique and system - Google Patents
A kind of user changes planes prediction technique and system Download PDFInfo
- Publication number
- CN109583949A CN109583949A CN201811400416.7A CN201811400416A CN109583949A CN 109583949 A CN109583949 A CN 109583949A CN 201811400416 A CN201811400416 A CN 201811400416A CN 109583949 A CN109583949 A CN 109583949A
- Authority
- CN
- China
- Prior art keywords
- user
- particle
- information
- planes
- quanta particle
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 239000002245 particle Substances 0.000 claims abstract description 190
- 238000003066 decision tree Methods 0.000 claims abstract description 38
- 238000012549 training Methods 0.000 claims abstract description 33
- 238000005457 optimization Methods 0.000 claims abstract description 23
- 230000006870 function Effects 0.000 claims description 30
- 230000006399 behavior Effects 0.000 claims description 8
- 235000012054 meals Nutrition 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 5
- 235000013399 edible fruits Nutrition 0.000 claims 2
- 230000010365 information processing Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 7
- 239000002096 quantum dot Substances 0.000 description 6
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- 230000035772 mutation Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000009153 huxin Substances 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/40—Business processes related to the transportation industry
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Game Theory and Decision Science (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Tourism & Hospitality (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
It changes planes prediction technique, system, device and a kind of computer readable storage medium the invention discloses a kind of user, is related to technical field of information processing.The present invention encodes the user information of acquisition, optimization is iterated to user information coding using complex parameter quantum particle swarm optimization, determine input of the user characteristics label as decision tree training pattern, according to the decision tree training pattern training user characteristics label, determine optimal objective function value, output user changes planes prediction result.It not only can handle high-dimensional data, user can also be improved and changed planes the accuracy of prediction result.
Description
Technical field
The present invention relates to technical field of information processing more particularly to a kind of based on complex parameter quantum particle swarm decision tree
User changes planes method, system, device and a kind of computer readable storage medium of prediction.
Background technique
Currently, intelligent terminal has become social, consumption in for people's lives, amusement and that obtain information etc. indispensable is auxiliary
Assistant engineer's tool, with the fast development of mobile communication technology and intelligent terminal, usage scenario, brand, the model of mobile terminal etc. are more
New to regenerate rapidly, operator and equipment vendor etc. generally require to carry out the relevant marketing activity of terminal, and use to improve telecommunications
Family is changed planes the accuracy rate of prediction, is needed as far as possible adequately with all kinds of teledata sources.
User in the prior art changes planes, and prediction technique is mainly based upon terminal shipment, user's questionnaire survey, terminal are opened
Identification information that machine is sent, change planes with existing decision Tree algorithms the modes such as prediction, and the side based on terminal shipment
Formula, analysis result lay particular emphasis on the shipment amount and city's accounting of mobile device manufacturer, and it is new mobile can not accurately to analyze the practical enabling of user
The case where equipment and can not knowing, accurately changes planes location information;Based on the mode of user's questionnaire survey, since the period is longer,
The timeliness of analysis is poor, and the application value for analyzing result is also restricted;Identification information based on terminal booting transmission
Mode can not then analyze the strategy that user changes planes in long period scale, and the aforesaid way of the prior art obtains
Analysis result accuracy it is lower;And change planes with existing decision Tree algorithms and predict to be easy over-fitting, cause practical pre-
The effect of survey is not high, and is not suitable for processing high dimensional data, and when number of attributes is excessive, decision tree effect is poor, in addition,
Existing decision Tree algorithms are too sensitive to exceptional value, are easy to cause the structure of tree to generate variation, and generalization ability is too poor, for
The user characteristics label not occurred can not be handled.
Summary of the invention
The embodiment of the present invention provides a kind of user and changes planes prediction technique, can mention while being suitable for high dimensional data
High user changes planes prediction result accuracy.
In order to achieve the above objectives, the embodiment of the present invention provides a kind of user and changes planes prediction technique, changes planes applied to user
Prediction meanss, which comprises
User information is obtained, the user information includes: terminal parameter information, terminal behavior information, user information, position
Information, consumption information, user preference information and user browse information;
The user information is encoded;
Optimize the coding according to complex parameter quantum particle swarm optimization, obtains user characteristics label;
According to the decision tree training pattern training user characteristics label, obtains user and change planes prediction result.
Accordingly, it changes planes forecasting system the embodiment of the invention also provides a kind of user, the system comprises acquisition of information
Unit, coding unit and analytical unit;
The information acquisition unit, for obtaining user information, the user information includes: terminal parameter information, terminal
Behavioural information, user information, location information, consumption information, user preference information and user browse information;
The coding unit, for being encoded to the user information;
The analytical unit obtains user for optimizing the coding according to complex parameter quantum particle swarm optimization
Feature tag;And
According to the decision tree training pattern training user characteristics label, obtains user and change planes prediction result.
The present invention is iterated optimization to user information coding by using complex parameter quantum particle swarm optimization, obtains
The input of decision tree training pattern, i.e. user characteristics label are taken, the user characteristics label is trained according to decision tree training pattern,
Determine optimal objective function value, output user changes planes prediction result.It not only can handle high-dimensional data, use can also be improved
Family is changed planes the accuracy of prediction result.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing.
Fig. 1 is that the user of the embodiment of the present invention changes planes the flow chart of prediction technique;
Fig. 2 is the flow chart of the decision-tree model algorithm of the embodiment of the present invention;
Fig. 3 is the flow chart of the complex parameter quantum particle swarm optimization of the embodiment of the present invention;
Fig. 4 is that the user of the embodiment of the present invention changes planes the structural schematic diagram of forecasting system;
Fig. 5 is that the user of the embodiment of the present invention changes planes the structural schematic diagrams of prediction meanss.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The embodiment of the present invention provides a kind of user and changes planes prediction technique, as shown in Figure 1, changing planes prediction for user of the invention
Method flow diagram, the method are applied to user and change planes prediction meanss, and the method specifically includes:
S101, user information is obtained, the user information includes: terminal parameter information, terminal behavior information, Yong Huxin
Breath, location information, consumption information, user preference information and user browse information.
Wherein it is possible to obtain user information from user, the user information can also be obtained from network side, it can also be from
Third party's information platform obtains the user information, in this regard, the embodiment of the present invention is not specifically limited.
The terminal parameter information can include but is not limited to terminal brand, terminal models, terminal price, Time To Market and
Operating system;Whether it is contract machine that the terminal behavior information can include but is not limited to terminal using duration and terminal;It is described
User information can include but is not limited to age of user, user's gender and user gradation;The location information may include but not
It is limited to user's permanent residence on daytime and user's night permanent residence;The consumption information can include but is not limited to user's monthly average and disappear
Take, user's monthly average uses flow, user's monthly average duration of call and user's set meal type;The user preference information can wrap
Include but be not limited to the tracking area list TA List that user often uses the daily active time of application program, user, user to be passed through daily
Quantity and user pay close attention to classification;The user, which browses information, can include but is not limited to the number that user browses mobile phone sales platform
And duration.User information classification can be adjusted according to specific circumstances.In this regard, the embodiment of the present invention is without limiting.
S102, the user information is encoded.
Specifically, carrying out one-hot coding using the method for enumerated value to discrete type user information;To continuous type user information
Using etc. frequency divisions case method carry out interval division, each section is encoded according to the interval division.
Wherein, the discrete type user information includes: terminal brand, terminal models, operating system, user's gender, user
Grade, user's permanent residence on daytime, user's night permanent residence, user's set meal type, user often use application program, user active daily
Time and user pay close attention to classification;The continuous type user information includes: that terminal price, Time To Market, age of user, user month are flat
It consumes, the tracking area list TA that user's monthly average is passed through daily using flow, user's monthly average duration of call, user
List quantity, user browse the number and duration of mobile phone sales platform.
Such as: user's gender is encoded, male can be encoded to 0, women is encoded to 1;Terminal brand is carried out
Coding, can be by the common brand of TOP20 according to 1~20 coding, other brands are encoded to 0;To user daytime/night permanent residence into
Row coding can be encoded for the different zones in different cities;Terminal price is encoded, it can be by terminal price
Be divided into section: (0,2000], (2000,4000], (4000,6000] and (6000 ,+∞), by above-mentioned section be separately encoded for
1,2,3,4.The embodiment of the present invention is to specific coding mode and method without limiting.
S103, the coding is optimized according to complex parameter quantum particle swarm optimization, obtains user characteristics label.
S104, according to the decision tree training pattern training user characteristics label, obtain user and change planes prediction result.
The process is specially as shown in Fig. 2, being the decision-tree model algorithm flow chart of the embodiment of the present invention, comprising:
S1041, the input that the decision tree is determined according to the user characteristics label;
Specifically, according toDetermine user characteristics label;Wherein,It is i-th of quanta particle
D-th of bit, the t times iteration position, r is that each iteration generates equally distributed random number between [0,1] at random,It is the speed of d-th of bit, the t times iteration of i-th of quanta particle;Wherein,
Wherein,It is i-th of quanta particle
D-th of bit, the t times iteration rotation angle;It is of d-th of bit, the t-1 times iteration of i-th of quanta particle
Body optimal location,It is the global optimum position of d-th of bit, the t-1 times iteration of i-th of quanta particle;Wherein
Wherein, a1And a2It is constant, for identifying the individual of quanta particle
Optimal location, global optimum position and influence journey of the position to the speed of quanta particle in current iteration for initializing quanta particle
Degree.
Specifically, if the position of d-th of bit, the t times iteration of i-th of quanta particleIt is 1, then it is its is corresponding
Input of the user characteristics label as decision tree, if the position of d-th of bit, the t times iteration of i-th of quanta particleFor
0, then give up the user characteristics label.
S1042, the input that the decision tree is trained according to decision tree training pattern, if the number of iterations is greater than pre-determined threshold,
Determine optimal objective function value, output user changes planes prediction result.
Further, as shown in figure 3, the process of the complex parameter quantum particle swarm optimization for the embodiment of the present invention
Figure, i.e. step S103 may include:
S1031, according to complex parameter quantum particle swarm optimization, initialize position, speed and the individual of quanta particle
Optimal location;Wherein, the position x of the initialization quanta particleiIt is 0 or 1, initializes the speed v of quanta particlei=(αi1 αi2
… αil)=(vi1 vi2 … vil), initialize the personal best particle p of quanta particlei=xi;Wherein, i refers to i-th of quantum grain
Son, l refer to the dimension of quanta particle, and the dimension is for identifying the user information.
Wherein, xi=(xi1,xi2,…,xil), 1≤i≤P;pi=(pi1 pi2 … pil)=(xi1,xi2,…,xil),
In, P indicates the scale of quantum particle swarm, i.e. the quanta particle number of quantum particle swarm.
Specifically, for conventional discrete particle swarm algorithm, since the bit of quanta particle can be expressed as a pair of of plural number
(α, β), wherein α and β is respectively indicated to after user information coding, obtains the probability amplitude of different coding, wherein 0≤α≤1, | α
|2+|β|2=1, therefore the speed of quanta particle can indicate are as follows:
vi=(αi1 αi2 … αil)=(vi1 vi2 … vil)。
Invention introduces the concepts of Quantum rotating gate, are updated with speed of the quantum rotation angle θ to quanta particle, rotation
The thought turned are as follows: in an iterative process, particle is moved by the rotation angle θ, the direction bigger to objective function, it is assumed that quanta particle
Position takes " 1 " that objective function can be made to obtain greater value, then | α |2It should reduce, | β |2It should increase, if | αi+1|2<|αi|2, | βi+1|2>|
βi|2, then Quantum rotating gate indicates are as follows:
Wherein, j-th of quantum bit of j expression particle, j-th of i-th of particle
The speed vij of quantum bit can be indicated are as follows:
Wherein, t indicates the t times iteration;
It can also be expressed as:
Wherein,
Wherein, a1And a2It is constant, for identifying the individual of quanta particle
Optimal location, global optimum position and influence journey of the position to the speed of quanta particle in current iteration for initializing quanta particle
Degree.
Further, locally optimal solution is fallen into order to prevent, introduces mutation probability c,
Wherein, μ is decay factor, meets 0 < μ < 1, and T is maximum number of iterations, and t indicates the t times iteration.The variation is general
Rate c reduces with the increase of the number of iterations and levels off to 0.
At this point, if the current location of some bit of quanta particle, personal best particle and global optimum position are identical,
And meet random quantity r < c, the speed v of j-th of quantum bit of i-th of particleijIt can indicate are as follows:
Wherein, it is equally distributed random between [0,1] to be that each iteration generates at random by r
Number, N is quantum non-gate.The formula can simplify are as follows:
Therefore, the speed of d-th of bit, the t times iteration of i-th of quanta particle can indicate are as follows:
Wherein,It is the rotation angle of d-th of bit, the t times iteration of i-th of quanta particle;It is i-th of quantum grain
The personal best particle of d-th of bit, the t-1 times iteration of son,It is d-th of bit t-1 of i-th of quanta particle
The global optimum position of secondary iteration.
S1032, the objective function for calculating the quanta particle, obtain the global optimum position p of populationg。
Wherein, the global optimum position p of the populationgIndicate entire population in hyperspace motion process, with
The calculating to the target function value of each quanta particle, and iteration each time obtain in all populations of memory optimal
The particle of target function value optimal location achieved.
S1033, the position and speed for updating the quanta particle.
Specifically, according to formulaUpdate the amount
The speed of seed, whereinIt is the rotation angle of d-th of bit, the t times iteration of i-th of quanta particle;It is i-th
The personal best particle of d-th of bit, the t-1 times iteration of quanta particle,It is d-th of bit of i-th of quanta particle
The global optimum position of the t-1 times iteration in position.
According to formulaUpdate the position of the quanta particle, whereinIt is i-th of quantum
The position of d-th of bit, the t times iteration of particle, it is equally distributed random between [0,1] that r is that each iteration generates at random
Number,It is the speed of d-th of bit, the t times iteration of i-th of quanta particle.
S1034, the objective function that the quanta particle is calculated according to the position and speed of the updated quanta particle,
Obtain personal best particle, global optimum position and the mean place of the quanta particle.
Wherein, each particle is in hyperspace motion process, can to make its objective function obtain the position of maximum value into
Row memory, the personal best particle of the quanta particle is for characterizing particle optimal location obtained itself.
The global optimum position p of the populationgEntire population is indicated in hyperspace motion process, with to every
The calculating of the target function value of a quanta particle, and iteration each time obtain optimal objective letter in all populations of memory
The particle of numerical value optimal location achieved.
Mean place z=[the z1,z2,...,zM], whereinJ=1,2 ..., M.
S1035, user spy is determined according to the personal best particle of the quanta particle, global optimum position and mean place
Levy label.
Wherein, the objective function F (x) of the quanta particle=(TP+TN)/(TP+TN+FN+FP);Wherein, TP indicates pre-
It surveys result to change planes for user, and user is practical changes planes;FP indicates that prediction result is changed planes for user, and actual user does not change planes;TN table
Show that prediction user is to change planes, actual user does not change planes;FN indicates that prediction user does not change planes, and actual user changes planes.
If the number of iterations is not up to preset maximum number of iterations, repeatedly step S1033, S1034;If institute
It states the number of iterations and has reached preset maximum number of iterations, then algorithm terminates.
The present invention is iterated optimization to user information coding by using complex parameter quantum particle swarm optimization, obtains
The input of decision tree training pattern, i.e. user characteristics label are taken, the user characteristics label is trained according to decision tree training pattern,
Determine optimal objective function value, output user changes planes prediction result.It not only can handle high-dimensional data, use can also be improved
Family is changed planes the accuracy of prediction result.
It changes planes forecasting system the embodiment of the invention also provides a kind of user, is the use of the embodiment of the present invention as described in Figure 4
Family is changed planes the structural schematic diagram of forecasting system, and user forecasting system 10 of changing planes includes information acquisition unit 110, coding unit
120 and analytical unit 130.Shown in Fig. 4 is only schematic diagram, is not changed planes other modules of forecasting system to the user, and
The structural relation of modules, which is constituted, to be limited.Wherein,
Information acquisition unit 110, for obtaining user information, the user information includes: terminal parameter information, terminal row
Information is browsed for information, user information, location information, consumption information, user preference information and user.
Wherein it is possible to obtain user information from user, the user information can also be obtained from network side, it can also be from
Third party's information platform obtains the user information, in this regard, the embodiment of the present invention is not specifically limited.
The terminal parameter information can include but is not limited to terminal brand, terminal models, terminal price, Time To Market and
Operating system;Whether it is contract machine that the terminal behavior information can include but is not limited to terminal using duration and terminal;It is described
User information can include but is not limited to age of user, user's gender and user gradation;The location information may include but not
It is limited to user's permanent residence on daytime and user's night permanent residence;The consumption information can include but is not limited to user's monthly average and disappear
Take, user's monthly average uses flow, user's monthly average duration of call and user's set meal type;The user preference information can wrap
Include but be not limited to the tracking area list TAList that user often uses the daily active time of application program, user, user to be passed through daily
Quantity and user pay close attention to classification;The user, which browses information, can include but is not limited to the number that user browses mobile phone sales platform
And duration.User information classification can be adjusted according to specific circumstances.In this regard, the embodiment of the present invention is without limiting.
Coding unit 120, for being encoded to the user information.
Specifically, carrying out one-hot coding using the method for enumerated value to discrete type user information;To continuous type user information
Using etc. frequency divisions case method carry out interval division, each section is encoded according to the interval division.
Wherein, the discrete type user information includes: terminal brand, terminal models, operating system, user's gender, user
Grade, user's permanent residence on daytime, user's night permanent residence, user's set meal type, user often use application program, user active daily
Time and user pay close attention to classification;The continuous type user information includes: that terminal price, Time To Market, age of user, user month are flat
It consumes, the tracking area list TAList that user's monthly average is passed through daily using flow, user's monthly average duration of call, user
Quantity, user browse the number and duration of mobile phone sales platform.
Such as: user's gender is encoded, male can be encoded to 0, women is encoded to 1;Terminal brand is carried out
Coding, can be by the common brand of TOP20 according to 1~20 coding, other brands are encoded to 0;To user daytime/night permanent residence into
Row coding can be encoded for the different zones in different cities;Terminal price is encoded, it can be by terminal price
Be divided into section: (0,2000], (2000,4000], (4000,6000] and (6000 ,+∞), by above-mentioned section be separately encoded for
1,2,3,4.The embodiment of the present invention is to specific coding mode and method without limiting.
It is special to obtain user for optimizing the coding according to complex parameter quantum particle swarm optimization for analytical unit 130
Levy label;And
According to the decision tree training pattern training user characteristics label, obtains user and change planes prediction result.
Specifically, described optimize the coding according to complex parameter quantum particle swarm optimization, user characteristics mark is obtained
The detailed process of label, comprising:
Analytical unit 130 according to complex parameter quantum particle swarm optimization, initialize the position of quanta particle, speed and
Personal best particle;Wherein, the position x of the initialization quanta particleiIt is 0 or 1, initializes the speed v of quanta particlei=
(αi1 αi2 … αil)=(vi1 vi2 … vil), initialize the personal best particle p of quanta particlei=xi;Wherein, i refers to i-th
A quanta particle, l refer to the dimension of quanta particle, and the dimension is for identifying the user information.
Wherein, xi=(xi1,xi2,…,xil), 1≤i≤P;pi=(pi1 pi2 … pil)=(xi1,xi2,…,xil),
In, P indicates the scale of quantum particle swarm, i.e. the quanta particle number of quantum particle swarm.
Specifically, for conventional discrete particle swarm algorithm, since the bit of quanta particle can be expressed as a pair of of plural number
(α, β), wherein α and β is respectively indicated to after user information coding, obtains the probability amplitude of different coding, wherein 0≤α≤1, | α
|2+|β|2=1, therefore the speed of quanta particle can indicate are as follows:
vi=(αi1 αi2 … αil)=(vi1 vi2 … vil)。
Invention introduces the concepts of Quantum rotating gate, are updated with speed of the quantum rotation angle θ to quanta particle, rotation
The thought turned are as follows: in an iterative process, particle is moved by the rotation angle θ, the direction bigger to objective function, it is assumed that quanta particle
Position takes " 1 " that objective function can be made to obtain greater value, then | α |2It should reduce, | β |2It should increase, if | αi+1|2<|αi|2, | β i+1 |2
>|βi|2, then Quantum rotating gate indicates are as follows:
Wherein, j-th of quantum bit of j expression particle, j-th of i-th of particle
The speed vij of quantum bit can be indicated are as follows:
Wherein, t indicates the t times iteration;
It can also be expressed as:
Wherein,
Wherein, a1And a2It is constant, for identifying the individual of quanta particle
Optimal location, global optimum position and influence journey of the position to the speed of quanta particle in current iteration for initializing quanta particle
Degree.
Further, locally optimal solution is fallen into order to prevent, introduces mutation probability c,
Wherein, μ is decay factor, meets 0 < μ < 1, and T is maximum number of iterations, and t indicates the t times iteration.The variation is general
Rate c reduces with the increase of the number of iterations and levels off to 0.
At this point, if the current location of some bit of quanta particle, personal best particle and global optimum position are identical,
And meet random quantity r < c, the speed v of j-th of quantum bit of i-th of particleijIt can indicate are as follows:
Wherein, it is equally distributed random between [0,1] to be that each iteration generates at random by r
Number, N is quantum non-gate.The formula can simplify are as follows:
Therefore, the speed of d-th of bit, the t times iteration of i-th of quanta particle can indicate are as follows:
Wherein,It is the rotation angle of d-th of bit, the t times iteration of i-th of quanta particle;It is i-th of quantum grain
The personal best particle of d-th of bit, the t-1 times iteration of son,It is d-th of bit t-1 of i-th of quanta particle
The global optimum position of secondary iteration.
Analytical unit 130 calculates the objective function of the quanta particle, obtains the global optimum position p of populationg。
Wherein, the global optimum position p of the populationgIndicate entire population in hyperspace motion process, with
The calculating to the target function value of each quanta particle, and iteration each time obtain in all populations of memory optimal
The particle of target function value optimal location achieved.
Analytical unit 130 updates the position and speed of the quanta particle.
Specifically, according to formulaUpdate the amount
The speed of seed, whereinIt is the rotation angle of d-th of bit, the t times iteration of i-th of quanta particle;It is i-th
The personal best particle of d-th of bit, the t-1 times iteration of quanta particle,It is d-th of bit of i-th of quanta particle
The global optimum position of the t-1 times iteration in position.
According to formulaUpdate the position of the quanta particle, whereinIt is i-th of quantum
The position of d-th of bit, the t times iteration of particle, it is equally distributed random between [0,1] that r is that each iteration generates at random
Number,It is the speed of d-th of bit, the t times iteration of i-th of quanta particle.
Analytical unit 130 calculates the target of the quanta particle according to the position and speed of the updated quanta particle
Function obtains personal best particle, global optimum position and the mean place of the quanta particle.
Wherein, each particle is in hyperspace motion process, can to make its objective function obtain the position of maximum value into
Row memory, the personal best particle of the quanta particle is for characterizing particle optimal location obtained itself.
The global optimum position p of the populationgEntire population is indicated in hyperspace motion process, with to every
The calculating of the target function value of a quanta particle, and iteration each time obtain optimal objective letter in all populations of memory
The particle of numerical value optimal location achieved.
Mean place z=[the z1,z2,...,zM], whereinJ=1,2 ..., M.
Analytical unit 130 is determined according to the personal best particle of the quanta particle, global optimum position and mean place
User characteristics label;
Wherein, the objective function F (x) of the quanta particle=(TP+TN)/(TP+TN+FN+FP);Wherein, TP indicates pre-
It surveys result to change planes for user, and user is practical changes planes;FP indicates that prediction result is changed planes for user, and actual user does not change planes;TN table
Show that prediction user is to change planes, actual user does not change planes;FN indicates that prediction user does not change planes, and actual user changes planes.
If the number of iterations is not up to preset maximum number of iterations, repeatedly step S1033, S1034;If institute
It states the number of iterations and has reached preset maximum number of iterations, then algorithm terminates.
Further, described according to the decision tree training pattern training user characteristics label, it obtains user and changes planes prediction
As a result, comprising:
Analytical unit 130 determines the input of the decision tree according to the user characteristics label;
Specifically, according toDetermine user characteristics label;Wherein,It is i-th of quanta particle
D-th of bit, the t times iteration position, r is that each iteration generates equally distributed random number between [0,1] at random,It is the speed of d-th of bit, the t times iteration of i-th of quanta particle;Wherein,
Wherein,It is i-th of quanta particle
D-th of bit, the t times iteration rotation angle;It is of d-th of bit, the t-1 times iteration of i-th of quanta particle
Body optimal location,It is the global optimum position of d-th of bit, the t-1 times iteration of i-th of quanta particle;Wherein
Wherein, a1And a2It is constant, for identifying the individual of quanta particle
Optimal location, global optimum position and influence journey of the position to the speed of quanta particle in current iteration for initializing quanta particle
Degree.
Specifically, if the position of d-th of bit, the t times iteration of i-th of quanta particleIt is 1, then it is its is corresponding
Input of the user characteristics label as decision tree, if the position of d-th of bit, the t times iteration of i-th of quanta particleFor
0, then give up the user characteristics label.
Analytical unit 130 is preset according to the input of the decision tree training pattern training decision tree if the number of iterations is greater than
Thresholding, determines optimal objective function value, and output user changes planes prediction result.
The present invention is iterated optimization to user information coding by using complex parameter quantum particle swarm optimization, obtains
The input of decision tree training pattern, i.e. user characteristics label are taken, the user characteristics label is trained according to decision tree training pattern,
Determine optimal objective function value, output user changes planes prediction result.It not only can handle high-dimensional data, use can also be improved
Family is changed planes the accuracy of prediction result.
It changes planes prediction meanss the embodiment of the invention also provides a kind of user, is the use of the embodiment of the present invention as described in Figure 5
Family is changed planes the structural schematic diagrams of prediction meanss.The user change planes prediction meanss 20 include memory 210 and processor 220.Fig. 5
Shown in be only schematic diagram, other modules for prediction meanss of not changing planes to the user and the structural relation structure of modules
At restriction.Wherein,
Memory 210, for storing computer program.
Processor 220 realizes that user changes planes prediction technique for executing the computer program.
In several embodiments provided herein, it should be understood that disclosed method and system can pass through it
Its mode is realized.For example, system embodiment described above is only schematical, for example, the functional module is drawn
Point, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can
To combine or be desirably integrated into another system, or some features can be ignored or not executed.Another point, it is shown or beg for
The mutual coupling, direct-coupling or communication connection of opinion can be through some interfaces, the INDIRECT COUPLING of device or unit
Or communication connection, it can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, each functional module in each embodiment of the present invention can integrate in one processing unit
It is that the independent physics of each unit includes, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or the network equipment etc.) executes the portion of each embodiment the method for the present invention
Step by step.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (Read-Only Memory, abbreviation
ROM), random access memory (Random Access Memory, abbreviation RAM), magnetic or disk etc. are various can store
The medium of program code.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, protection scope of the present invention should be based on the protection scope of the described claims.
Claims (14)
- The prediction technique 1. a kind of user changes planes, which is characterized in that the described method includes:User information is obtained, the user information includes: terminal parameter information, terminal behavior information, user information, position letter Breath, consumption information, user preference information and user browse information;The user information is encoded;Optimize the coding according to complex parameter quantum particle swarm optimization, obtains user characteristics label;According to the decision tree training pattern training user characteristics label, obtains user and change planes prediction result.
- The prediction technique 2. user according to claim 1 changes planes, which is characterized in that the terminal parameter information includes: end Hold brand, terminal models, terminal price, Time To Market and operating system;Whether it using duration and terminal is contract machine that the terminal behavior information includes: terminal;The user information includes: age of user, user's gender and user gradation;The location information includes: user's permanent residence on daytime and user's night permanent residence;The consumption information includes: that user's monthly average is consumed, user's monthly average uses flow, user's monthly average duration of call and use Family set meal type;The user preference information include: user often use that the daily active time of application program, user, user passed through daily with Track area list TA List quantity and user pay close attention to classification;It includes: the number and duration that user browses mobile phone sales platform that the user, which browses information,.
- The prediction technique 3. user according to claim 2 changes planes, which is characterized in that described that the user information is compiled Code include:One-hot coding is carried out using the method for enumerated value to discrete type user information;To frequency divisions casees such as continuous type user information uses Method carries out interval division, is encoded according to the interval division to each section;Wherein, the discrete type user information include: terminal brand, terminal models, operating system, user's gender, user gradation, User's permanent residence on daytime, user's night permanent residence, user's set meal type, user often use the daily active time of application program, user Classification is paid close attention to user;The continuous type user information includes: that terminal price, Time To Market, age of user, user's monthly average disappear The tracking area list TA List number that expense, user's monthly average are passed through daily using flow, user's monthly average duration of call, user Amount, user browse the number and duration of mobile phone sales platform.
- The prediction technique 4. user according to claim 3 changes planes, which is characterized in that described according to complex parameter quanta particle Colony optimization algorithm optimizes the coding, obtains user characteristics label, comprising:According to complex parameter quantum particle swarm optimization, position, speed and the personal best particle of quanta particle are initialized;Its In, the position x of the initialization quanta particleiIt is 0 or 1, initializes the speed v of quanta particlei=(αi1 αi2 … αil)= (vi1 vi2 … vil), initialize the personal best particle p of quanta particlei=xi;Wherein, i refers to i-th of quanta particle, the l amount of finger The dimension of seed, the dimension is for identifying the user information;The objective function for calculating the quanta particle obtains the global optimum position p of populationg;Update the position and speed of the quanta particle;The objective function that the quanta particle is calculated according to the position and speed of the updated quanta particle, obtains the amount Personal best particle, global optimum position and the mean place of seed;User characteristics label is determined according to the personal best particle of the quanta particle, global optimum position and mean place;Wherein, the objective function F (x) of the quanta particle=(TP+TN)/(TP+TN+FN+FP);Wherein, TP indicates prediction knot Fruit is changed planes for user, and user is practical changes planes;FP indicates that prediction result is changed planes for user, and actual user does not change planes;TN indicates pre- Surveying user is to change planes, and actual user does not change planes;FN indicates that prediction user does not change planes, and actual user changes planes.
- The prediction technique 5. user according to claim 4 changes planes, which is characterized in that described to be instructed according to decision tree training pattern Practice the user characteristics label, obtain user and change planes prediction result, comprising:The input of the decision tree is determined according to the user characteristics label;Optimal mesh is determined if the number of iterations is greater than pre-determined threshold according to the input of the decision tree training pattern training decision tree Offer of tender numerical value, output user change planes prediction result.
- The prediction technique 6. user according to claim 4 or 5 changes planes, which is characterized in that described according to the quanta particle Personal best particle, global optimum position and mean place determine user characteristics label, comprising:According toDetermine user characteristics label;Wherein,It is d-th of bit of i-th of quanta particle The position of the t times iteration in position, r is that each iteration generates equally distributed random number between [0,1] at random,It is i-th of amount The speed of d-th of bit, the t times iteration of seed;Wherein,Wherein,It is the d of i-th of quanta particle The rotation angle of the t times iteration of a bit;Be d-th of bit, the t-1 times iteration of i-th of quanta particle individual most Excellent position,It is the global optimum position of d-th of bit, the t-1 times iteration of i-th of quanta particle;WhereinWherein, a1And a2It is constant, the individual for identifying quanta particle is optimal Position, global optimum position and influence degree of the position to the speed of quanta particle in current iteration for initializing quanta particle.
- The forecasting system 7. a kind of user changes planes, which is characterized in that the system comprises:Information acquisition unit, for obtaining user information, the user information include: terminal parameter information, terminal behavior information, User information, location information, consumption information, user preference information and user browse information;Coding unit, for being encoded to the user information;Analytical unit obtains user characteristics label for optimizing the coding according to complex parameter quantum particle swarm optimization; AndAccording to the decision tree training pattern training user characteristics label, obtains user and change planes prediction result.
- The forecasting system 8. user according to claim 7 changes planes, which is characterized in that the terminal parameter information includes: end Hold brand, terminal models, terminal price, Time To Market and operating system;Whether it using duration and terminal is contract machine that the terminal behavior information includes: terminal;The user information includes: age of user, user's gender and user gradation;The location information includes: user's permanent residence on daytime and user's night permanent residence;The consumption information includes: that user's monthly average is consumed, user's monthly average uses flow, user's monthly average duration of call and use Family set meal type;The user preference information include: user often use that the daily active time of application program, user, user passed through daily with Track area list TA List quantity and user pay close attention to classification;It includes: the number and duration that user browses mobile phone sales platform that the user, which browses information,.
- The forecasting system 9. user according to claim 8 changes planes, which is characterized in that described that the user information is compiled Code include:One-hot coding is carried out using the method for enumerated value to discrete type user information;To frequency divisions casees such as continuous type user information uses Method carries out interval division, is encoded according to the interval division to each section;Wherein, the discrete type user information include: terminal brand, terminal models, operating system, user's gender, user gradation, User's permanent residence on daytime, user's night permanent residence, user's set meal type, user often use the daily active time of application program, user Classification is paid close attention to user;The continuous type user information includes: that terminal price, Time To Market, age of user, user's monthly average disappear The tracking area list TA List number that expense, user's monthly average are passed through daily using flow, user's monthly average duration of call, user Amount, user browse the number and duration of mobile phone sales platform.
- The forecasting system 10. user according to claim 9 changes planes, which is characterized in that described according to complex parameter quantum grain Subgroup optimization algorithm optimizes the coding, obtains user characteristics label, comprising:According to complex parameter quantum particle swarm optimization, position, speed and the personal best particle of quanta particle are initialized;Its In, the position x of the initialization quanta particleiIt is 0 or 1, initializes the speed v of quanta particlei=(αi1 αi2 … αil)= (vi1 vi2 … vil), initialize the personal best particle p of quanta particlei=xi;Wherein, i refers to i-th of quanta particle, the l amount of finger The dimension of seed, the dimension is for identifying the user information;The objective function for calculating the quanta particle obtains the global optimum position p of populationg;Update the position and speed of the quanta particle;The objective function that the quanta particle is calculated according to the position and speed of the updated quanta particle, obtains the amount Personal best particle, global optimum position and the mean place of seed;User characteristics label is determined according to the personal best particle of the quanta particle, global optimum position and mean place;Wherein, the objective function F (x) of the quanta particle=(TP+TN)/(TP+TN+FN+FP);Wherein, TP indicates prediction knot Fruit is changed planes for user, and user is practical changes planes;FP indicates that prediction result is changed planes for user, and actual user does not change planes;TN indicates pre- Surveying user is to change planes, and actual user does not change planes;FN indicates that prediction user does not change planes, and actual user changes planes.
- The forecasting system 11. user according to claim 10 changes planes, which is characterized in that described according to decision tree training pattern The training user characteristics label obtains user and changes planes prediction result, comprising:The input of the decision tree is determined according to the user characteristics label;Optimal mesh is determined if the number of iterations is greater than pre-determined threshold according to the input of the decision tree training pattern training decision tree Offer of tender numerical value, output user change planes prediction result.
- 12. user described in 0 or 11 changes planes forecasting system according to claim 1, which is characterized in that described according to the quantum grain Personal best particle, global optimum position and the mean place of son determine user characteristics label, comprising:According toDetermine user characteristics label;Wherein,It is d-th of bit of i-th of quanta particle The position of the t times iteration in position, r is that each iteration generates equally distributed random number between [0,1] at random,It is i-th of amount The speed of d-th of bit, the t times iteration of seed;Wherein,Wherein,It is the d of i-th of quanta particle The rotation angle of the t times iteration of a bit;Be d-th of bit, the t-1 times iteration of i-th of quanta particle individual most Excellent position,It is the global optimum position of d-th of bit, the t-1 times iteration of i-th of quanta particle;WhereinWherein, a1And a2It is constant, the individual for identifying quanta particle is optimal Position, global optimum position and influence degree of the position to the speed of quanta particle in current iteration for initializing quanta particle.
- The prediction meanss 13. a kind of user changes planes, which is characterized in that described device includes:Memory, for storing computer program;Processor realizes that user as claimed in any one of claims 1 to 6 changes planes the side of prediction for executing the computer program Method.
- 14. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium Program, the computer program realize that user as claimed in any one of claims 1 to 6 changes planes the side of prediction when being executed by processor Method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811400416.7A CN109583949B (en) | 2018-11-22 | 2018-11-22 | User switching prediction method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811400416.7A CN109583949B (en) | 2018-11-22 | 2018-11-22 | User switching prediction method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109583949A true CN109583949A (en) | 2019-04-05 |
CN109583949B CN109583949B (en) | 2021-03-12 |
Family
ID=65923466
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811400416.7A Active CN109583949B (en) | 2018-11-22 | 2018-11-22 | User switching prediction method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109583949B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110175170A (en) * | 2019-04-19 | 2019-08-27 | 平安科技(深圳)有限公司 | Data-optimized processing method, device, computer equipment and storage medium |
CN110991711A (en) * | 2019-11-19 | 2020-04-10 | 浙江鸿程计算机系统有限公司 | Multi-factor perception terminal switching prediction method based on deep neural network |
CN111047425A (en) * | 2019-11-25 | 2020-04-21 | 中国联合网络通信集团有限公司 | Behavior prediction method and device |
CN111913940A (en) * | 2020-06-20 | 2020-11-10 | 武汉海云健康科技股份有限公司 | Temperature member label prediction method and device, electronic equipment and storage medium |
CN113538025A (en) * | 2020-04-14 | 2021-10-22 | 中国移动通信集团浙江有限公司 | Method and device for predicting replacement of terminal equipment |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104794195A (en) * | 2015-04-17 | 2015-07-22 | 南京大学 | Data mining method for finding potential telecommunication users changing cell phones |
CN105678596A (en) * | 2016-03-23 | 2016-06-15 | 中国联合网络通信集团有限公司 | Mobile terminal replacement prediction method, and replacement prediction system thereof |
CN107016460A (en) * | 2017-03-27 | 2017-08-04 | 中国联合网络通信集团有限公司广西壮族自治区分公司 | User changes planes Forecasting Methodology and device |
CN107590569A (en) * | 2017-09-25 | 2018-01-16 | 山东浪潮云服务信息科技有限公司 | A kind of data predication method and device |
CN107657267A (en) * | 2017-08-11 | 2018-02-02 | 百度在线网络技术(北京)有限公司 | Product potential user method for digging and device |
CN107783998A (en) * | 2016-08-26 | 2018-03-09 | 华为技术有限公司 | The method and device of a kind of data processing |
CN107798379A (en) * | 2017-11-23 | 2018-03-13 | 东北大学 | Improve the method for quantum particle swarm optimization and the application based on innovatory algorithm |
CN108256907A (en) * | 2018-01-09 | 2018-07-06 | 北京腾云天下科技有限公司 | A kind of construction method and computing device of customer grouping model |
CN108647364A (en) * | 2018-05-21 | 2018-10-12 | 广东工业大学 | A kind of prediction recommendation method based on mobile terminal application data |
US20180300796A1 (en) * | 2017-04-14 | 2018-10-18 | International Business Machines Corporation | Cognitive order processing by predicting resalable returns |
CN108846695A (en) * | 2018-06-07 | 2018-11-20 | 中国联合网络通信集团有限公司 | The prediction technique and device of terminal replacement cycle |
-
2018
- 2018-11-22 CN CN201811400416.7A patent/CN109583949B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104794195A (en) * | 2015-04-17 | 2015-07-22 | 南京大学 | Data mining method for finding potential telecommunication users changing cell phones |
CN105678596A (en) * | 2016-03-23 | 2016-06-15 | 中国联合网络通信集团有限公司 | Mobile terminal replacement prediction method, and replacement prediction system thereof |
CN107783998A (en) * | 2016-08-26 | 2018-03-09 | 华为技术有限公司 | The method and device of a kind of data processing |
CN107016460A (en) * | 2017-03-27 | 2017-08-04 | 中国联合网络通信集团有限公司广西壮族自治区分公司 | User changes planes Forecasting Methodology and device |
US20180300796A1 (en) * | 2017-04-14 | 2018-10-18 | International Business Machines Corporation | Cognitive order processing by predicting resalable returns |
CN107657267A (en) * | 2017-08-11 | 2018-02-02 | 百度在线网络技术(北京)有限公司 | Product potential user method for digging and device |
CN107590569A (en) * | 2017-09-25 | 2018-01-16 | 山东浪潮云服务信息科技有限公司 | A kind of data predication method and device |
CN107798379A (en) * | 2017-11-23 | 2018-03-13 | 东北大学 | Improve the method for quantum particle swarm optimization and the application based on innovatory algorithm |
CN108256907A (en) * | 2018-01-09 | 2018-07-06 | 北京腾云天下科技有限公司 | A kind of construction method and computing device of customer grouping model |
CN108647364A (en) * | 2018-05-21 | 2018-10-12 | 广东工业大学 | A kind of prediction recommendation method based on mobile terminal application data |
CN108846695A (en) * | 2018-06-07 | 2018-11-20 | 中国联合网络通信集团有限公司 | The prediction technique and device of terminal replacement cycle |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110175170A (en) * | 2019-04-19 | 2019-08-27 | 平安科技(深圳)有限公司 | Data-optimized processing method, device, computer equipment and storage medium |
CN110175170B (en) * | 2019-04-19 | 2023-11-21 | 平安科技(深圳)有限公司 | Data optimization processing method, device, computer equipment and storage medium |
CN110991711A (en) * | 2019-11-19 | 2020-04-10 | 浙江鸿程计算机系统有限公司 | Multi-factor perception terminal switching prediction method based on deep neural network |
CN111047425A (en) * | 2019-11-25 | 2020-04-21 | 中国联合网络通信集团有限公司 | Behavior prediction method and device |
CN111047425B (en) * | 2019-11-25 | 2023-10-24 | 中国联合网络通信集团有限公司 | Behavior prediction method and device |
CN113538025A (en) * | 2020-04-14 | 2021-10-22 | 中国移动通信集团浙江有限公司 | Method and device for predicting replacement of terminal equipment |
CN113538025B (en) * | 2020-04-14 | 2024-03-22 | 中国移动通信集团浙江有限公司 | Replacement prediction method and device for terminal equipment |
CN111913940A (en) * | 2020-06-20 | 2020-11-10 | 武汉海云健康科技股份有限公司 | Temperature member label prediction method and device, electronic equipment and storage medium |
CN111913940B (en) * | 2020-06-20 | 2024-04-26 | 武汉海云健康科技股份有限公司 | Temperature membership tag prediction method and device, electronic equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN109583949B (en) | 2021-03-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109583949A (en) | A kind of user changes planes prediction technique and system | |
US20220027359A1 (en) | Online hyperparameter tuning in distributed machine learning | |
Junior et al. | An iterative boosting-based ensemble for streaming data classification | |
US8392153B2 (en) | Method for constructing segmentation-based predictive models | |
CN110968701A (en) | Relationship map establishing method, device and equipment for graph neural network | |
CN109766454A (en) | A kind of investor's classification method, device, equipment and medium | |
WO2001002986A2 (en) | Automated web-based targeted advertising with quotas | |
CN104462594A (en) | Method and device for providing user personalized resource message pushing | |
CN113468227A (en) | Information recommendation method, system, device and storage medium based on graph neural network | |
CN108734216A (en) | Classification of power customers method, apparatus and storage medium based on load curve form | |
CN110046981A (en) | A kind of credit estimation method, device and storage medium | |
CN106909560A (en) | Point of interest sort method | |
CN107870810A (en) | Using method for cleaning, device, storage medium and electronic equipment | |
Jahangir et al. | Comparing performance of genetic and discrete invasive weed optimization algorithms for solving the inventory routing problem with an incremental delivery | |
CN113065882A (en) | Commodity processing method and device and electronic equipment | |
CN112633690A (en) | Service personnel information distribution method, service personnel information distribution device, computer equipment and storage medium | |
CN108073672A (en) | Business object recommends method, apparatus and electronic equipment | |
CN114511387A (en) | Product recommendation method and device, electronic equipment and storage medium | |
CN116739665A (en) | Information delivery method and device, electronic equipment and storage medium | |
Osojnik et al. | Incremental predictive clustering trees for online semi-supervised multi-target regression | |
CN113822734A (en) | Method and apparatus for generating information | |
Hain et al. | The promises of Machine Learning and Big Data in entrepreneurship research | |
Roy et al. | A short review on applications of big data analytics | |
Hu et al. | Learning mixed multinomial logits with provable guarantees | |
CN116051196A (en) | Advertisement recommendation method, advertisement recommendation device, advertisement recommendation equipment, advertisement recommendation storage medium and advertisement recommendation computer product |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |