CN106296434A - A kind of Grain Crop Yield Prediction method based on PSO LSSVM algorithm - Google Patents

A kind of Grain Crop Yield Prediction method based on PSO LSSVM algorithm Download PDF

Info

Publication number
CN106296434A
CN106296434A CN201610684306.2A CN201610684306A CN106296434A CN 106296434 A CN106296434 A CN 106296434A CN 201610684306 A CN201610684306 A CN 201610684306A CN 106296434 A CN106296434 A CN 106296434A
Authority
CN
China
Prior art keywords
value
factor
grain yield
grain
year
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
Application number
CN201610684306.2A
Other languages
Chinese (zh)
Other versions
CN106296434B (en
Inventor
杨铁军
杨娜
朱春华
樊超
傅洪亮
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Henan University of Technology
Original Assignee
Henan University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Henan University of Technology filed Critical Henan University of Technology
Priority to CN201610684306.2A priority Critical patent/CN106296434B/en
Publication of CN106296434A publication Critical patent/CN106296434A/en
Application granted granted Critical
Publication of CN106296434B publication Critical patent/CN106296434B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06QDATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/02Agriculture; Fishing; Mining
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06QDATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/04Forecasting or optimisation, e.g. linear programming, "travelling salesman problem" or "cutting stock problem"

Abstract

A kind of Grain Crop Yield Prediction method based on PSO LSSVM algorithm, the method in turn includes the following steps: obtain the grain yield in known time and the value of main affecting factors;The initial data of grain yield and main affecting factors is smoothed;Forecast model is drawn according to least square method supporting vector machine model;The width of penalty factor and kernel function is solved by iterative algorithm;Solve Lagrange multiplier and variate-value b;Solve the value of Radial basis kernel function;The Lagrange multiplier solved, variate-value b and RBF are substituted into forecast model, and calculates the predictive value of the grain yield of 1 year with model.The invention discloses a kind of Grain Crop Yield Prediction method based on PSO LSSVM algorithm, grain yield can be predicted by this Forecasting Methodology, it was predicted that precision is high, can improve the prediction accuracy of China's grain yield.

Description

A kind of Grain Crop Yield Prediction method based on PSO-LSSVM algorithm
Technical field
The invention belongs to Grain Crop Yield Prediction field, be specifically related to a kind of grain yield based on PSO-LSSVM algorithm pre- Survey method.
Background technology
The Accurate Prediction of grain yield can be that government decision provides foundation, and grain yield data are typical small sample numbers According to, it is vulnerable to the impact of the uncertain factor, shows as a complicated nonlinear system.
At present, the method for prediction grain yield is varied both at home and abroad, mainly includes regression analysis, Time Series Method And Artificial Neural Network.The method can be applied to grain yield to carrying out causal analysis between variable by regression analysis During prediction, the main affecting factors of grain yield can be found, but owing to all of factor of influence can not be carried out fully by the method Consideration, be therefore only applicable to short-term forecast.Time Series Method calculates the shortest, relatively low to the quantitative requirement of historical data, The consecutive variations of yield can be reflected.But the model that the method is set up is model based on linear data mostly, and during grain Between sequence data tend to appear as nonlinear characteristic, therefore cause precision of prediction the highest.Artificial Neural Network is a kind of non- Linear prediction method, has parallel processing and the fault-tolerant ability of height, and application is relatively broad at present.But, neutral net requirement Data sample is big, and grain yield data belong to Small Sample Database, so often there is result over-fitting during prediction, extensive The phenomenon such as indifferent.
To sum up, all there is either large or small problem in existing Grain Crop Yield Prediction method, and provides a kind of precision of prediction high Grain Crop Yield Prediction method, grain prediction is had great importance.
Summary of the invention
It is desirable to provide a kind of Grain Crop Yield Prediction method based on PSO-LSSVM algorithm that precision of prediction is high.
For solving above-mentioned technical problem, the invention provides following technical scheme: a kind of based on PSO-LSSVM algorithm Grain Crop Yield Prediction method, the method in turn includes the following steps:
(1) grain yield in known time and the value of main affecting factors are obtained;
(2) grain yield in the known time that step (1) gets and the initial data of main affecting factors are carried out smooth place Reason;
(3) forecast model is drawn according to least square method supporting vector machine model, wherein It is the grain yield of 1 year,It is the value of the factor of influence of 1 year,It it is the value of the factor of influence of 1 year; Being Lagrange multiplier, b is variate-value,For Radial basis kernel function;
(4) penalty factor is solved by iterative algorithmWidth with kernel function
(5) Lagrange multiplier and variate-value b are solved;
(6) Radial basis kernel function is solvedValue;
(7) Lagrange multiplier solved, variate-value b and RBF are substituted into forecast model
, and calculate with this modelThe predictive value of the grain yield of i.e. 1 year;Wherein i is The grain yield that can inquire and the year of factor of influence value, 1 year main affecting factors is it is known that grain yield is to be predicted 's.
The method that draws of main affecting factors is: according to grain yield and the value of factor of influence, calculates factor of influence and grain The degree of association between food yield, compare the degree of association between factor of influence and grain yield, and wherein the degree of association is maximum Factor of influence is main affecting factors.
Described in step (2), the method that the grain yield in step (1) and main affecting factors are smoothed For:
1) respectively the initial data of grain yield and main affecting factors is done difference processing;
2) according to the difference processing of step 1), grain yield and fluctuation meansigma methods p of main affecting factors and py are calculated respectively;
3) obtain grain yield and the value of main affecting factors after smoothing processing, use respectivelyLp-Lp1WithYp-Yp1Calculate adjacent two Undulating value between annual grain yield and main affecting factors valuep1Withpy1
Following three kinds of situations are had for grain yield:
The first situation:P1 > p, andLp < Lp1, then makeLp1=Lp+p
The second situation:P1 > p, andLp > Lp1, then makeLp1=Lp-p
The third situation:P1 < p, thenLp1= Lp1
Following three kinds of situations are had for the value of main affecting factors:
The first situation:Py1 > py, andYp < Yp1, then orderYp1=Yp+py
The second situation:Py1 > py, andYp > Yp1, then orderYp1=Yp-py
The third situation:Py1 < py, thenYp1= Yp1
WhereinLpIt isn-mThe grain yield in year;Lp1Forn-m+1The grain yield in year;YpIt isn-mAnnual main shadow Ring the original value of the factor;Yp1Forn-m+1The original value of annual main affecting factors.
Penalty factor is solved by the method for iteration optimizingWidth with kernel function:
According to forecast model, step 1, show that fitness function is: In formula,WithIt is respectively actual value and the predictive value of 1 year grain yield;WithIt is training sample and inspection sample respectively This number;
Step 2, according to equation below, updates particle rapidity and position;
Wherein,
In formula,tFor iterations,Being nonnegative number, referred to as Inertia Weight, before how many control particle every generation speed renewals has Speed remain.WithIt is nonnegative constant, referred to as accelerated factor.It is the speed of particle,, It is constant, is set by the user the speed for limiting particle,It it is the position of particle.WithIt is random between [0,1] Number.WithBe respectively according to training after LSSVM for the test error that test sample collection obtains determine when the one before Body optimal value and current population optimal value;
Step 3, is analyzed the result after each iteration, the current fitness value obtained by this iterationAnd individuality Extreme valueCompare, if>, then useReplace;With its fitness valueWith overall situation pole ValueCompare, if>, then useReplacement is fallen, simultaneously according to speed and the position of step 3 more new particle;
Step 4, if fitness function value reaches precision of prediction or reaches default maximum iteration time, then output parameter is Excellent solution, optimizing terminates, if the most above-mentioned condition, branches to step (3) and re-starts search.
The method for solving of Lagrange multiplier and variate-value b is: according to formula, solve b, root According to formula, solve;Wherein,It is 1iThe inversion of matrix,It it is main affecting factors composition 'sSquare formation, wherein the value of the i-th row i row is xi×x i ,ForiRank unit matrix, yiIt it is the value of 1 year grain yield;For penalty factor.
The method for solving of RBF is: utilize formulaSolve, its In,It is the value of the factor of influence of 1 year,It it is the value of the factor of influence of 1 year.
By above technical scheme, the invention have the benefit that
1, forecast model of the present invention is by getting up PSO algorithm and least square method supporting vector machine models coupling, simultaneously Initial data is carried out smoothing processing, has been obviously improved so that precision of prediction has had, for improving Grain Crop Yield Prediction Accuracy rate has great significance.
2, being determined by the value of main affecting factors, the prediction for this grain yield has had simplification largely.
3, penalty factorWidth with kernel functionMethod for solving in, by iteration optimization, thus draw optimum, And then improve the precision of prediction of forecast model further.
Accompanying drawing explanation
Fig. 1 is Forecasting Methodology flow chart of the present invention.
Detailed description of the invention
A kind of Grain Crop Yield Prediction method based on PSO-LSSVM algorithm, as it is shown in figure 1, the method includes as follows successively Step:
(1) obtain the grain yield in known time and the value of main affecting factors, it is known that time grain yield and mainly affect because of The value of son can be recorded by existing document and learn.
Wherein judge that whether factor of influence is that the method for main affecting factors is: according to grain yield and factor of influence, meter Calculating the degree of association between factor of influence and grain yield, it neutralizes the maximum factor of influence of the grain yield degree of association is main impact The factor.The computational methods of the degree of association between factor of influence and grain yield are ripe prior art.Wherein, factor of influence and The calculating of the degree of association between grain yield refers to the Chinese patent application of Application No. " 201510985352.1 ".
(2) grain yield in known time and the initial data of main affecting factors to obtaining are smoothed, flat The method of sliding process is:
The first step, does difference processing to the initial data of grain yield and main affecting factors respectively;
Second step, according to the difference processing of step 1), calculates fluctuation meansigma methods p of grain yield and main affecting factors respectively Fluctuation meansigma methods py;
3rd step, obtains grain yield and the value of main affecting factors after smoothing processing: calculate adjacent two year with Lp-Lp1 Undulating value p1 between grain yield, calculates adjacent two annual main affecting factors values py1 with Yp-Yp1;Wherein Lp is m The grain yield in year;Lp1 is the grain yield in m+1 year;Yp is the original value of m year main affecting factors;Yp1 is m+ The original value (m=1,2,3 ... n-2) of 1 annual main affecting factors.
Following three kinds of situations are had for grain yield:
The first situation: if p1 is > p, and Lp < Lp1, then order Lp1=Lp+p;
The second situation: if p1 is > p, and Lp > Lp1, then order Lp1=Lp-p;
The third situation: if p1 < p, then Lp1=Lp1;
Following three kinds of situations are had for the value of main affecting factors:
The first situation: if py1 is > py, and Yp < Yp1, then order Yp1=Yp+py;
The second situation: if py1 is > py, and Yp > Yp1, then order Yp1=Yp-py;
The third situation: if py1 < py, then Yp1=Yp1;
Reduce grain yield and the undulating value of main affecting factors by smoothing processing, thus improve precision of prediction.
(3) forecast model is drawn according to least square method supporting vector machine model, whereinIt is the grain yield of 1 year,It is the value of the factor of influence of 1 year,It it is the value of the factor of influence of 1 year;Being Lagrange multiplier, b is variate-value,For Radial basis kernel function;This forecast model is according to Little square law draws.
(4) penalty factor is solved by iterative algorithmWidth with kernel function
Penalty factorWidth with kernel functionMethod for solving be:
Penalty factor is solved by the method for iteration optimizingWidth with kernel function:
According to forecast model, step 1, show that fitness function is: In formula,WithIt is respectively actual value and the predictive value of 1 year grain yield;WithIt is training sample and inspection sample respectively This number;
Step 2, according to equation below, updates particle rapidity and position;
Wherein,
In formula,tFor iterations,Being nonnegative number, referred to as Inertia Weight, before how many control particle every generation speed renewals has Speed remain.WithIt is nonnegative constant, referred to as accelerated factor.It is the speed of particle,, It is constant, is set by the user the speed for limiting particle,It it is the position of particle.WithIt is random between [0,1] Number.WithBe respectively according to training after LSSVM for the test error that test sample collection obtains determine when the one before Body optimal value and current population optimal value;
Step 3, is analyzed the result after each iteration, the current fitness value obtained by this iterationAnd individuality Extreme valueCompare, if>, then useReplace;With its fitness valueWith overall situation pole ValueCompare, if>, then useReplacement is fallen, simultaneously according to speed and the position of step 3 more new particle;
Step 4, if fitness function value reaches precision of prediction or reaches default maximum iteration time, then output parameter is Excellent solution, optimizing terminates, if the most above-mentioned condition, branches to step (3) and re-starts search.
For random number rand (1,1), automatically generate for computer, not by man's activity during prediction, as long as having The value of known main affecting factors and grain yield, it is possible to draw the cereal product of the unknown according to forecast model.
(5) solving Lagrange multiplier and variate-value b, the method for solving of Lagrange multiplier and variate-value b is: according to public affairs Formula, solve b, according to formula(wherein i=1,2, n-1), solves; Wherein,It is 1iThe inversion of matrix,It it is main affecting factors compositionSquare formation, wherein the value of the i-th row i row is xi , xi It is the value of 1 year main affecting factors, yiIt it is the value of 1 year grain yield;For penalty factor.
(6) Radial basis kernel function is solvedValue, the method for solving of RBF is: utilize formulaSolve, wherein,It is the value of the factor of influence of 1 year,It it is the shadow of 1 year Ringing the value of the factor, wherein n is that the value of main affecting factors is it is known that grain yield time to be predicted.WhereinDraw and can join Examine the Chinese patent application of Application No. " 201510985352.1 ".
(7) Lagrange multiplier solved, variate-value b and RBF are substituted into formula
, and the predictive value of 1 year is calculated with this model;Wherein i is that the grain that can inquire produces In the year of the value of amount and main affecting factors, the value of the main affecting factors of 1 year is it is known that grain yield is to be predicted.
According to above step, the grain yield of, 2012 and in 2011 in 2013 is predicted, utilized existing simultaneously Some LS-SVM, SVM and ARIMA models have been also carried out prediction to this grain yield in 3 years, it was predicted that result is as shown in table 1 below:
Table 1 Grain Crop Yield Prediction results contrast
Predicting the outcome according to above-mentioned, actual value and above-mentioned predicting the outcome are compared, wherein forecast error is as shown in table 2:
Table 2 average relative error compares
By Tables 1 and 2, the precision of ARIMA forecast model is minimum, and average relative error is 1.73%, and SVM model has Non-linear advantage so that improve a bit than ARIMA on precision of prediction, average relative error is 1.56%;LSSVM is owing to needing The model parameter determined is fewer than SVM, has preferably coordinated the relation between the complication of model and generalization ability, it was predicted that precision is also It is obviously improved;And PSO-LSSVM forecast model is by introducing particle swarm optimization algorithm, that finds in LSSVM model is optimal Parameter, improves model accuracy again, it was predicted that the performance that result also show this model is the most superior, and average relative error is 0.8%, the Forecasting Methodology of the present invention by the initial data that PSO-LSSVM forecast model is used is smoothed, average phase Minimum to error, it was predicted that effect is the most optimal.
The invention discloses a kind of Grain Crop Yield Prediction method based on PSO-LSSVM algorithm, can by this Forecasting Methodology To be predicted grain yield, PSO algorithm and least square method supporting vector machine models coupling are got up by the method, simultaneously to former Beginning data have carried out smoothing processing, it was predicted that precision is substantially better than traditional LSSVM, SVM and ARIMA model, can improve me Predicting the outcome of state's grain yield, for grain, plants and purchases offer guidance.

Claims (6)

1. a Grain Crop Yield Prediction method based on PSO-LSSVM algorithm, it is characterised in that: the method includes walking as follows successively Rapid:
Obtain the grain yield in known time and the value of main affecting factors;
The grain yield in known time and the initial data of main affecting factors that get step (1) are smoothed;
Forecast model is drawn according to least square method supporting vector machine model, whereinIt is The grain yield of n,It is the value of the factor of influence of 1 year,It it is the value of the factor of influence of 1 year;It is to draw Ge Lang multiplier, b is variate-value,For Radial basis kernel function;
Penalty factor is solved by iterative algorithmWidth with kernel function
Solve Lagrange multiplier and variate-value b;
Solve Radial basis kernel functionValue;
The Lagrange multiplier solved, variate-value b and RBF are substituted into forecast model
, and calculate with this modelThe predictive value of the grain yield of i.e. 1 year;Wherein i is The grain yield that can inquire and the year of factor of influence value, 1 year main affecting factors is it is known that grain yield is to be predicted 's.
2. Grain Crop Yield Prediction method based on PSO-LSSVM algorithm as claimed in claim 1, it is characterised in that: main shadow The method that draws ringing the factor is: according to grain yield and the value of factor of influence, calculate the pass between factor of influence and grain yield Connection degree, compares the degree of association between factor of influence and grain yield, and the factor of influence that wherein degree of association is maximum is main Factor of influence.
3. Grain Crop Yield Prediction method based on PSO-LSSVM algorithm as claimed in claim 2, it is characterised in that: step (2) Described, the method being smoothed the grain yield in step (1) and main affecting factors is:
Respectively the initial data of grain yield and main affecting factors is done difference processing;
According to the difference processing of step 1), calculate grain yield and fluctuation meansigma methods p of main affecting factors and py respectively;
Obtain grain yield and the value of main affecting factors after smoothing processing, use respectivelyLp-Lp1WithYp-Yp1Calculate adjacent 2 year Undulating value between degree grain yield and main affecting factors valuep1Withpy1
Following three kinds of situations are had for grain yield:
The first situation:P1 > p, andLp < Lp1, then makeLp1=Lp+p
The second situation:P1 > p, andLp > Lp1, then makeLp1=Lp-p
The third situation:P1 < p, thenLp1= Lp1
Following three kinds of situations are had for the value of main affecting factors:
The first situation:Py1 > py, andYp < Yp1, then orderYp1=Yp+py
The second situation:Py1 > py, andYp > Yp1, then orderYp1=Yp-py
The third situation:Py1 < py, thenYp1= Yp1
WhereinLpIt isn-mThe grain yield in year;Lp1Forn-m+1The grain yield in year;YpIt isn-mAnnual main shadow Ring the original value of the factor;Yp1Forn-m+1The original value of annual main affecting factors.
4. Grain Crop Yield Prediction method based on PSO-LSSVM algorithm as claimed in claim 3, it is characterised in that: by repeatedly Method for optimizing solves the width of penalty factor and kernel function:
According to forecast model, step 1, show that fitness function is: In formula,WithIt is respectively actual value and the predictive value of 1 year grain yield;WithIt is training sample and inspection sample respectively This number;
Step 2, according to equation below, updates particle rapidity and position;
Wherein,
In formula,tFor iterations,Being nonnegative number, referred to as Inertia Weight, before how many control particle every generation speed renewals has Speed remain;
WithIt is nonnegative constant, referred to as accelerated factor;
It is the speed of particle,,It is constant,It it is the position of particle;
WithIt it is the random number between [0,1];
WithBe respectively according to training after LSSVM current individual that the test error that test sample collection obtains is determined The figure of merit and current population optimal value;
Step 3, is analyzed the result after each iteration, the current fitness value obtained by this iterationWith individual pole ValueCompare, if>, then useReplace;With its fitness valueAnd global extremumCompare, if>, then useReplacement is fallen, simultaneously according to speed and the position of step 3 more new particle;
Step 4, if fitness function value reaches precision of prediction or reaches default maximum iteration time, then output parameter is Excellent solution, optimizing terminates, if the most above-mentioned condition, branches to step (3) and re-starts search.
5. Grain Crop Yield Prediction method based on PSO-LSSVM algorithm as claimed in claim 4, it is characterised in that: glug is bright The method for solving of day multiplier and variate-value b is: according to formula, solve b, according to formula, solve;Wherein,It is 1iThe inversion of matrix,It it is main affecting factors compositionSide Battle array, wherein the value of the i-th row i row is xi×x i,ForiRank unit matrix, yiIt it is the value of 1 year grain yield;For punishment The factor.
6. Grain Crop Yield Prediction method based on PSO-LSSVM algorithm as claimed in claim 5, it is characterised in that: radially base The method for solving of function is: utilize formulaSolve, wherein,It it is the shadow of 1 year Ring the value of the factor,It it is the value of the factor of influence of 1 year.
CN201610684306.2A 2016-08-18 2016-08-18 Grain yield prediction method based on PSO-LSSVM algorithm Active CN106296434B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610684306.2A CN106296434B (en) 2016-08-18 2016-08-18 Grain yield prediction method based on PSO-LSSVM algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610684306.2A CN106296434B (en) 2016-08-18 2016-08-18 Grain yield prediction method based on PSO-LSSVM algorithm

Publications (2)

Publication Number Publication Date
CN106296434A true CN106296434A (en) 2017-01-04
CN106296434B CN106296434B (en) 2022-02-11

Family

ID=57678788

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610684306.2A Active CN106296434B (en) 2016-08-18 2016-08-18 Grain yield prediction method based on PSO-LSSVM algorithm

Country Status (1)

Country Link
CN (1) CN106296434B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108090628A (en) * 2018-01-16 2018-05-29 浙江大学 A kind of grain feelings security detection and analysis method based on PSO-LSSVM algorithms
CN112541296A (en) * 2020-07-22 2021-03-23 华北电力大学(保定) SO2 prediction method based on PSO-LSSVM

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103400310A (en) * 2013-08-08 2013-11-20 华北电力大学(保定) Method for evaluating power distribution network electrical equipment state based on historical data trend prediction
CN105512751A (en) * 2015-11-30 2016-04-20 国家电网公司 Electricity consumption prediction method and device

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103400310A (en) * 2013-08-08 2013-11-20 华北电力大学(保定) Method for evaluating power distribution network electrical equipment state based on historical data trend prediction
CN105512751A (en) * 2015-11-30 2016-04-20 国家电网公司 Electricity consumption prediction method and device

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
冯贵阳 等: "基于主成分分析和最小二乘支持向量机的油田产量预测模型", 《电脑知识与技术》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108090628A (en) * 2018-01-16 2018-05-29 浙江大学 A kind of grain feelings security detection and analysis method based on PSO-LSSVM algorithms
CN112541296A (en) * 2020-07-22 2021-03-23 华北电力大学(保定) SO2 prediction method based on PSO-LSSVM

Also Published As

Publication number Publication date
CN106296434B (en) 2022-02-11

Similar Documents

Publication Publication Date Title
Ding et al. Forecasting China's electricity consumption using a new grey prediction model
Niu et al. Uncertainty modeling for chaotic time series based on optimal multi-input multi-output architecture: Application to offshore wind speed
CN106650784A (en) Feature clustering comparison-based power prediction method and device for photovoltaic power station
CN109299812B (en) Flood prediction method based on deep learning model and KNN real-time correction
CN108876021B (en) Medium-and-long-term runoff forecasting method and system
CN103106331B (en) Based on the lithographic line width Intelligent Forecasting of dimensionality reduction and increment type extreme learning machine
CN107480815A (en) A kind of power system taiwan area load forecasting method
CN110047015A (en) A kind of water total amount prediction technique merging KPCA and thinking Optimized BP Neural Network
CN108921359A (en) A kind of distribution gas density prediction technique and device
CN109754122A (en) A kind of Numerical Predicting Method of the BP neural network based on random forest feature extraction
CN107909221A (en) Power-system short-term load forecasting method based on combination neural net
CN107886160A (en) A kind of BP neural network section water demand prediction method
CN110751318A (en) IPSO-LSTM-based ultra-short-term power load prediction method
CN106296434A (en) A kind of Grain Crop Yield Prediction method based on PSO LSSVM algorithm
Zhang et al. Short‐term load forecasting based on wavelet neural network with adaptive mutation bat optimization algorithm
CN111292124A (en) Water demand prediction method based on optimized combined neural network
CN110851959A (en) Wind speed interval prediction method integrating deep learning and quantile regression
Gensler et al. An analog ensemble-based similarity search technique for solar power forecasting
CN108171379A (en) A kind of electro-load forecast method
Shao et al. An advanced weighted system based on swarm intelligence optimization for wind speed prediction
CN112307672A (en) BP neural network short-term wind power prediction method based on cuckoo algorithm optimization
CN110516835A (en) A kind of Multi-variable Grey Model optimization method based on artificial fish-swarm algorithm
CN112288164A (en) Wind power combined prediction method considering spatial correlation and correcting numerical weather forecast
CN103353895A (en) Pre-processing method of power distribution network line loss data
CN112149883A (en) Photovoltaic power prediction method based on FWA-BP neural network

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant