CN106934064A - Network information hotspot prediction system and method based on ELM - Google Patents
Network information hotspot prediction system and method based on ELM Download PDFInfo
- Publication number
- CN106934064A CN106934064A CN201710200772.3A CN201710200772A CN106934064A CN 106934064 A CN106934064 A CN 106934064A CN 201710200772 A CN201710200772 A CN 201710200772A CN 106934064 A CN106934064 A CN 106934064A
- Authority
- CN
- China
- Prior art keywords
- network information
- sample
- focus
- learning machine
- information focus
- 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
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000012549 training Methods 0.000 claims abstract description 36
- 238000012360 testing method Methods 0.000 claims abstract description 28
- 230000008569 process Effects 0.000 claims abstract description 16
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 11
- 238000013519 translation Methods 0.000 claims abstract description 9
- 238000003780 insertion Methods 0.000 claims 1
- 230000037431 insertion Effects 0.000 claims 1
- 239000011159 matrix material Substances 0.000 description 8
- 230000008859 change Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000000739 chaotic effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
- G06Q50/01—Social networking
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Databases & Information Systems (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Tourism & Hospitality (AREA)
- Software Systems (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- General Engineering & Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Operations Research (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a kind of network information hotspot prediction system and method based on ELM, including:The historic click-through rate data of network under test information focus are gathered, the learning sample of network information focus is constituted;Delay time T and embedded dimension m to network information hot spot data are estimated, and enter line translation to network information hot spot data, obtain training sample and test sample;Using extreme learning machine device training network information focus sample, in training process, using Cholesky decomposition methods to the weights β of extreme learning machine deviceLCarry out optimal solution;The weights β of limit of utilization Learning machineL, set up the forecast model of network information focus;Using forecast model, the test sample to network information focus is predicted;When technical scheme causes to be predicted network information focus, real-time is good, and result is preferable.
Description
Technical field
The present invention relates to the Forecasting Methodology of network information focus, and in particular to a kind of network information focus based on ELM is pre-
Examining system and method.
Background technology
The prediction of current network information focus is mainly used:Conventional statistics model and modern statistics model, conventional statistics
Model cannot tracking network information focus changing trend, predict the outcome extremely unreliable.Modern statistics model is higher to obtain
The network information hotspot prediction result of precision, but the data prediction requirement of large scale network information focus cannot be met.Therefore, it is
The accuracy of network information hotspot prediction is improved, the variation tendency of network information focus is preferably described, a kind of prediction is needed badly
Real-time is good and the preferable network information hotspot prediction system and method for result.
The content of the invention
The present invention overcomes the shortcomings of that prior art is present, and technical problem to be solved is:A kind of prediction real-time is provided
Well and result preferably based on ELM network information hotspot prediction system and method.
In order to solve the above-mentioned technical problem, the technical solution adopted by the present invention is:Network information hotspot prediction based on ELM
System, including:Collecting unit:Historic click-through rate data for gathering network under test information focus, constitute network information focus
Learning sample;Estimation unit:Estimate for the delay time T to network information hot spot data and embedded dimension m, and to net
Network information hot spot data enters line translation, obtains training sample and test sample;Training unit:For using extreme learning machine device instruction
Practice network information focus sample, in training process, using Cholesky decomposition methods to the weights β of extreme learning machine deviceLCarry out
Optimal solution;Model sets up unit:For the weights β of limit of utilization Learning machineL, set up the prediction mould of network information focus
Type;Predicting unit:For utilizing forecast model, the test sample to network information focus is predicted.
Preferably, when the estimation unit is estimated the delay time T and embedded dimension m of network information hot spot data,
Optimal delay time T and embedded dimension m are specifically determined using correlation integral algorithm.
Correspondingly, the network information hotspot prediction method based on ELM, comprises the following steps:Collection network under test information heat
The historic click-through rate data of point, constitute the learning sample of network information focus;To the delay time T of network information hot spot data and
Embedded dimension m is estimated, and enters line translation to network information hot spot data, obtains training sample and test sample;Using the limit
Learning machine training network information focus sample, in training process, using Cholesky decomposition methods to extreme learning machine device
Weights βLCarry out optimal solution;The weights β of limit of utilization Learning machineL, set up the forecast model of network information focus;Using pre-
Model is surveyed, the test sample to network information focus is predicted.
Preferably, when the delay time T to network information hot spot data and embedded dimension m estimate, specifically adopt
Optimal delay time T and embedded dimension m are determined with correlation integral algorithm.
The present invention has the advantages that compared with prior art:The present invention is predicted to network information focus
When, the historic click-through rate data of network under test information focus are first gathered, the learning sample of network information focus is constituted, then to net
The delay time T of network information hot spot data and embedded dimension m are estimated, and enter line translation to network information hot spot data, are obtained
Training sample and test sample, then using extreme learning machine device training network information focus sample, recycle extreme learning machine
The weights β of deviceL, the forecast model of network information focus is set up, finally using forecast model, to the test specimens of network information focus
Originally it is predicted;In the training process of extreme learning machine device, it is critical to find weights βLOptimal value, and for existing pole
Limit Learning machine, in βLSolution procedure in, have substantial amounts of matrix inversion operation, cause computation complexity high, to the network information
The training process of hotspot prediction model has a negative impact;Therefore, the present invention is carried out accordingly to existing extreme learning machine device
Improve, introduce weights β of the Cholesky decomposition methods to extreme learning machine deviceLCarry out optimal solution so that βLSolution only pass through
Arithmetic is it is achieved that without matrix inversion operation, calculating is simpler, drastically reduce the area the time of solution so that
Prediction real-time to network information focus is good, as a result also more satisfactory.
Brief description of the drawings
The present invention will be further described in detail below in conjunction with the accompanying drawings;
The structural representation of the embodiment of the network information hotspot prediction system based on ELM that Fig. 1 is provided for the present invention;
The schematic flow sheet of the network information hotspot prediction method based on ELM that Fig. 2 is provided for the present invention;
Fig. 3 is that the sample of gathered data when being predicted to a certain network information focus using embodiments of the invention is illustrated
Figure;
Fig. 4 is the signal of the optimal value obtained after estimating the delay time T of the network information hot spot data in Fig. 3
Figure;
Fig. 5 is the signal of the optimal value obtained after estimating the embedded dimension m of the network information hot spot data in Fig. 3
Figure;
Fig. 6 uses the result schematic diagram after being predicted to the network information focus in Fig. 3 of the invention;
In figure:101 is collecting unit, and 102 is estimation unit, and 103 is training unit, and 104 set up unit, 105 for model
It is predicting unit.
Specific embodiment
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
A part of embodiment of the invention, rather than whole embodiments;Based on the embodiment in the present invention, ordinary skill people
The every other embodiment that member is obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.
The structural representation of the embodiment of the network information hotspot prediction system based on ELM that Fig. 1 is provided for the present invention, such as
Shown in Fig. 1, the network information hotspot prediction system based on ELM, including:
Collecting unit 101:Historic click-through rate data for gathering network under test information focus, constitute network information focus
Learning sample.
Estimation unit 102:Estimate for the delay time T to network information hot spot data and embedded dimension m, and to net
Network information hot spot data enters line translation, obtains training sample and test sample.
Training unit 103:For using extreme learning machine device training network information focus sample, in training process, use
Weights β of the Cholesky decomposition methods to extreme learning machine deviceLCarry out optimal solution.
Model sets up unit 104:For the weights β of limit of utilization Learning machineL, set up the prediction mould of network information focus
Type.
Predicting unit 105:For utilizing forecast model, the test sample to network information focus is predicted.
To a specific forecasting problem, first have to gather historical data, if the sample set of historical data composition is:Wherein:N is the intrinsic dimensionality of data, and k is sample size, training
The regression error of sample is εi;So, the regression forms of standard limit learning machine can be expressed as:
In formula:L represents the number of hidden layer node, ciI-th error of hidden layer node is represented, f represents hidden layer node
Non-thread mapping function, αiAnd βiThe weights of hidden layer node and input node are represented respectively.
Formula (1) is solved, efficiency is at a fairly low and is difficult to obtain globally optimal solution, and Lagrange multiplier is introduced for this
Set up following majorized function:
In formula:HLNode matrix equation is represented, w represents weights, and T represents output result.
Local derviation to variable is calculated, and obtains equation below:
Formula (3) is solved, the weight computing formula of extreme learning machine is
Using weights βLThe forecast model of corresponding problem is built, the expression formula for obtaining output result is
In standard limit machine-learning process, it is critical to find βLOptimal value, in βLSolution procedure, there is substantial amounts of square
Battle array inversion operation, causes computation complexity high, and the training process to network information hotspot prediction model has a negative impact, therefore
The present embodiment is correspondingly improved to the limit machine of standard, Cholesky decomposition methods is introduced, along with extreme learning machine
The training process of device, is quickly found out βLOptimal solution.
Below to using Cholesky decomposition methods to the weights β of extreme learning machine deviceLOptimal solution is carried out to carry out in detail
Thin explanation:
Can be obtained according to formula (3):
Using formula (6) to βLSolved, can obtain corresponding linear equation form is:
ALβL=bL (7)
Meet following constraints simultaneously:
Composite type (6) and formula (8) can be obtained
For v, ALQuadratic form can be described as:
Labor is carried out it can be found that A to formula (10)LIt is a symmetric positive definite matrix, then use Cholesky pairs
It is decomposed, it is possible to obtained:
In formula, SLIt is a triangular matrix.
sijRepresent triangular matrix SLIn nonzero element, then according to ALElement αijCan obtain
In formula, i=1,2 ..., L, j=1,2 ..., L.
Convolution (11) and formula (7), while being multiplied bySo basisWith SLFL=bLEquivalence, can obtain
FLElement fiFor:
In formula, i=1,2 ..., L.
According to SLAnd FLβ can be obtainedLComputing formula be:
The modeling process of improved ELM in contrast standard ELM and the present invention, β in improved ELM in the present inventionLSolution
Only pass through arithmetic it is achieved that without matrix inversion operation, calculating simpler, drastically reduce the area solution when
Between.
Especially under the conditions of the node of hidden layer increases, speed is more accelerated, and can obtain:
So, AL+1With ALBetween relation can be expressed as:
In formula,
According to Cholesky decomposable processes it is recognised that by calculating sL+1,1With sL+1,LIn the element that is not zero just can be with
Obtain SN+1, can now obtain:
So, according to FLCan obtain
Therefore according to fL+1F can be obtainedL+1, f should not be recalculated1,f2,…,fL, accelerate learning efficiency, and can be with
Realize the on-line study of ELM.
The present embodiment first gathers the historic click-through rate of network under test information focus when being predicted to network information focus
Data, constitute the learning sample of network information focus, and then the delay time T to network information hot spot data and embedded dimension m enter
Row is estimated, and enters line translation to network information hot spot data, training sample and test sample is obtained, then using extreme learning machine
Device training network information focus sample, recycles the weights β of extreme learning machine deviceL, the forecast model of network information focus is set up,
Forecast model is finally utilized, the test sample to network information focus is predicted;In the training process of extreme learning machine device,
It is critical to find weights βLOptimal value, and for existing extreme learning machine device, in βLSolution procedure in, have substantial amounts of square
Battle array inversion operation, causes computation complexity high, and the training process to network information hotspot prediction model has a negative impact;Cause
This, the present invention is correspondingly improved to existing extreme learning machine device, introduces Cholesky decomposition methods to extreme learning machine
The weights β of deviceLCarry out optimal solution so that βLSolution only pass through arithmetic it is achieved that without matrix inversion operation,
Calculate simpler, drastically reduce the area the time of solution so that the prediction real-time to network information focus is good, as a result
It is more satisfactory.
Specifically, the estimation unit 102 is estimated the delay time T and embedded dimension m of network information hot spot data
When, optimal delay time T and embedded dimension m are specifically determined using correlation integral algorithm.
Network information focus is typically an one-dimensional data:{ x (i), i=1,2 ... n }, is prolonged according to its chaotic property
It, is then changed a multidimensional data by slow time (τ) and embedded dimension (m):X (t)=x (t), x (i+ τ) ..., x (i+ (m-1)
τ), so as to the eye actual regular data will be changed into random data, so as to find the change for wherein including
Change feature.Optimal τ and m is determined using correlation integral algorithm.If two sample points are:X (i) and X (j), they apart from rij
M () is:
rij(m)=| | X (i)-X (j) | | (19)
Correlation integral can so be obtained is:
In formula, r is the threshold value of distance.
Whole samples are divided into t sequence, ClIt is l-th correlation intergal of sequence, then can obtain:
Minimum point can be obtained is:
WhenWhen reaching minimum value, then it is considered that obtaining optimal τ.
I-th vector is after conversion:Xi(m+1), arest neighbors is Xn(i,m)(m+1), then have
IfIf E (m) is in maximum state of value, then just it is considered that obtaining most
Excellent m.
Correspondingly, the schematic flow sheet of the network information hotspot prediction method based on ELM that Fig. 2 is provided for the present invention, such as
Shown in Fig. 2, the network information hotspot prediction method based on ELM is comprised the following steps:
The historic click-through rate data of network under test information focus are gathered, the learning sample of network information focus is constituted.
Delay time T and embedded dimension m to network information hot spot data are estimated, and network information hot spot data is entered
Line translation, obtains training sample and test sample.
Using extreme learning machine device training network information focus sample, in training process, using Cholesky decomposition methods
To the weights β of extreme learning machine deviceLCarry out optimal solution.
The weights β of limit of utilization Learning machineL, set up the forecast model of network information focus.
Using forecast model, the test sample to network information focus is predicted.
Specifically, when the delay time T to network information hot spot data and embedded dimension m estimate, specifically adopt
Optimal delay time T and embedded dimension m are determined with correlation integral algorithm.
Used as research object, Fig. 3 is using the present invention to selection " Tianjin Chemical Plant blast " this network information focus below
Embodiment gathered data when being predicted to the network information focus sample schematic diagram, Fig. 4 is to the network information in Fig. 3
The schematic diagram of the optimal value that the delay time T of hot spot data is obtained after being estimated, Fig. 5 is to the network information focus in Fig. 3
The schematic diagram of the optimal value that the embedded dimension m of data is obtained after being estimated.
Fig. 3 is analyzed, it can be found that network information focus change is very complicated, is not only become with certain growth
Gesture, while having strong fluctuation.τ and m is estimated using correlation integral algorithm, as a result as shown in Figure 4 and Figure 5, from Fig. 4 and Tu
5 understand that optimal τ and m is respectively:7 and 8, the multidimensional data of network information focus is obtained according to τ=7 and m=8, before selection
Used as training sample, remaining is used to test its prediction effect 200 data.
Using improved ELM in the present invention, to " Tianjin Chemical Plant blast ", this network information much-talked-about topic is predicted,
Result is as shown in Figure 6.Fig. 6 is analyzed it can be found that in the present invention improved ELM network information focus survey high precision,
And it is sufficiently stable to predict the outcome, improved ELM can be used in the prediction of network information much-talked-about topic in showing the present invention, and
Predict the outcome very good.
Select current classical model:Multiple linear regression (MLR), BP neural network (BPNN), SVMs (SVM) are right
10 network information focuses of 2016 are predicted, and they predict the outcome as shown in table 1.
The precision of prediction of table 1 (%) is counted
The precision of prediction of all of network information focus is it can be found that relative to current classical model, this hair in contrast table 1
The network information hotspot prediction precision of improved ELM improves in bright, and it is more excellent to predict the outcome, and this shows, this
Improved ELM can be modeled to network information focus very well in invention, hold its change state and become, and be a kind of highly versatile
Network information hotspot prediction model.
The average modeling time of all models is counted, as a result as shown in table 2.
Averagely modeling time (second) statistics of table 2
It can be found that the average modeling time of improved ELM is minimum in the present invention from table 2, network information heat is accelerated
The modeling efficiency of point, it is possible to achieve network information focus on-line prediction.
Network information focus is influenceed by the thought of people, politics, economy and other factorses, changes sufficiently complex, no
Only there is strong time variation, and with certain chaos change feature.The network information focus of improved ELM in the present invention
Forecast model predict the outcome stabilization, it is credible, modeling and forecasting is in hgher efficiency, can apply to real network public sentiment data analysis,
Predicting the outcome can help pre-control some negative network information hotspot anneals, with actual application value higher.
Finally it should be noted that:Various embodiments above is merely illustrative of the technical solution of the present invention, rather than its limitations;To the greatest extent
Pipe has been described in detail with reference to foregoing embodiments to the present invention, it will be understood by those within the art that:Its according to
The technical scheme described in foregoing embodiments can so be modified, or which part or all technical characteristic are entered
Row equivalent;And these modifications or replacement, the essence of appropriate technical solution is departed from various embodiments of the present invention technology
The scope of scheme.
Claims (4)
1. the network information hotspot prediction system of ELM is based on, it is characterised in that:Including:
Collecting unit (101):Historic click-through rate data for gathering network under test information focus, constitute network information focus
Learning sample;
Estimation unit (102):Estimate for the delay time T to network information hot spot data and embedded dimension m, and to network
Information hot spot data enters line translation, obtains training sample and test sample;
Training unit (103):For using extreme learning machine device training network information focus sample, in training process, use
Weights β of the Cholesky decomposition methods to extreme learning machine deviceLCarry out optimal solution;
Model sets up unit (104):For the weights β of limit of utilization Learning machineL, set up the forecast model of network information focus;
Predicting unit (105):For utilizing forecast model, the test sample to network information focus is predicted.
2. the network information hotspot prediction system based on ELM according to claim 1, it is characterised in that:It is described to estimate single
When first (102) are estimated the delay time T and embedded dimension m of network information hot spot data, specifically calculated using correlation integral
Method determines optimal delay time T and embedded dimension m.
3. the network information hotspot prediction method of ELM is based on, it is characterised in that:Comprise the following steps:
The historic click-through rate data of S101, collection network under test information focus, constitute the learning sample of network information focus;
S102, the delay time T to network information hot spot data and insertion are tieed up m and are estimated, and to network information hot spot data
Enter line translation, obtain training sample and test sample;
S103, using extreme learning machine device training network information focus sample, in training process, using Cholesky decomposition methods
To the weights β of extreme learning machine deviceLCarry out optimal solution;
The weights β of S104, limit of utilization Learning machineL, set up the forecast model of network information focus;
S105, using forecast model, the test sample to network information focus is predicted.
4. the network information hotspot prediction method based on ELM according to claim 3, it is characterised in that:It is described to network
When the delay time T of information hot spot data and embedded dimension m are estimated, specifically determined using correlation integral algorithm optimal
Delay time T and embedded dimension m.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710200772.3A CN106934064A (en) | 2017-03-30 | 2017-03-30 | Network information hotspot prediction system and method based on ELM |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710200772.3A CN106934064A (en) | 2017-03-30 | 2017-03-30 | Network information hotspot prediction system and method based on ELM |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106934064A true CN106934064A (en) | 2017-07-07 |
Family
ID=59424917
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710200772.3A Pending CN106934064A (en) | 2017-03-30 | 2017-03-30 | Network information hotspot prediction system and method based on ELM |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106934064A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111859238A (en) * | 2020-07-27 | 2020-10-30 | 平安科技(深圳)有限公司 | Method and device for predicting data change frequency based on model and computer equipment |
CN112734126A (en) * | 2021-01-18 | 2021-04-30 | 武汉烽火技术服务有限公司 | Hot spot prediction method, device and equipment and readable storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103761254A (en) * | 2013-12-26 | 2014-04-30 | 清华大学 | Method for matching and recommending service themes in various fields |
CN105653645A (en) * | 2015-12-28 | 2016-06-08 | 精硕世纪科技(北京)有限公司 | Network information attention assessment method and device |
-
2017
- 2017-03-30 CN CN201710200772.3A patent/CN106934064A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103761254A (en) * | 2013-12-26 | 2014-04-30 | 清华大学 | Method for matching and recommending service themes in various fields |
CN105653645A (en) * | 2015-12-28 | 2016-06-08 | 精硕世纪科技(北京)有限公司 | Network information attention assessment method and device |
Non-Patent Citations (3)
Title |
---|
姬建新: "捕鱼算法优化核极限学习机的微博热点话题预测", 《激光杂志》 * |
张弦 等: "基于Cholesky 分解的增量式RELM及其在时间序列预测中的应用", 《物理学报》 * |
曹蕾 等: "长江日流量时间序列混沌特性研究—相空间嵌入维数和嵌入滞时的联合确定", 《水利科技与经济》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111859238A (en) * | 2020-07-27 | 2020-10-30 | 平安科技(深圳)有限公司 | Method and device for predicting data change frequency based on model and computer equipment |
CN112734126A (en) * | 2021-01-18 | 2021-04-30 | 武汉烽火技术服务有限公司 | Hot spot prediction method, device and equipment and readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Jin et al. | Multiple model LPV approach to nonlinear process identification with EM algorithm | |
CN110751318B (en) | Ultra-short-term power load prediction method based on IPSO-LSTM | |
CN108776820A (en) | It is a kind of to utilize the improved random forest integrated approach of width neural network | |
CN110175416A (en) | Three Gorges Reservoir water temperature prediction method based on principal component analysis and neural network | |
CN107480815A (en) | A kind of power system taiwan area load forecasting method | |
CN107797931A (en) | A kind of method for evaluating software quality and system based on second evaluation | |
CN106529818A (en) | Water quality evaluation prediction method based on fuzzy wavelet neural network | |
CN106649540A (en) | Video recommendation method and system | |
CN106778838A (en) | A kind of method for predicting air quality | |
CN111416797A (en) | Intrusion detection method for optimizing regularization extreme learning machine by improving longicorn herd algorithm | |
Pu et al. | UFNGBM (1, 1): A novel unbiased fractional grey Bernoulli model with Whale Optimization Algorithm and its application to electricity consumption forecasting in China | |
CN106202377A (en) | A kind of online collaborative sort method based on stochastic gradient descent | |
CN106296434B (en) | Grain yield prediction method based on PSO-LSSVM algorithm | |
CN102629341A (en) | Web service QoS (Quality of Service) on-line prediction method based on geographic position information of user | |
CN106204597A (en) | A kind of based on from the VS dividing method walking the Weakly supervised study of formula | |
CN103020709A (en) | Optimization calculation method based on artificial bee colony algorithm and quantum-behaved particle swarm optimization algorithm | |
CN105184400A (en) | Tobacco field soil moisture prediction method | |
CN106934064A (en) | Network information hotspot prediction system and method based on ELM | |
Gilan et al. | Sustainable building design: A challenge at the intersection of machine learning and design optimization | |
Pulikottil et al. | Onet–a temporal meta embedding network for mooc dropout prediction | |
Niu et al. | Crude oil prices and volatility prediction by a hybrid model based on kernel extreme learning machine | |
CN103605493A (en) | Parallel sorting learning method and system based on graphics processing unit | |
CN103902737A (en) | Projection pursuit classification modeling software and implementation based on swarm intelligence algorithms | |
CN110489616A (en) | A kind of search ordering method based on Ranknet and Lambdamart algorithm | |
JP5970578B2 (en) | Program and apparatus for determining relational model |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170707 |
|
RJ01 | Rejection of invention patent application after publication |