# Ensemble Kernel Learning Model for Prediction of Time Series Based on the Support Vector Regression
Time series prediction based on support vector regression
This code was used in my paper whose link is in the following.
Paper link: [Ensemble Kernel Learning Model for Prediction of Time Series Based on the Support Vector Regression and Meta Heuristic Search](https://jsdp.rcisp.ac.ir/article-1-1162-fa.html)
# Background
Time series prediction is a process which predicted future system values based on information obtained from past and present data points.
The main purpose of using different models for time series prediction is making the forecast with the greatest accuracy.
In this repository, a new model for time series prediction is presented based on support vector regression and weighted combination of different kernels and optimization
of parameters and their weights by the optimizer. In successive iterations of the competency function, the optimizer learns the optimal values of the particle array
and uses those values to evaluate the test data.
# Data Set
CO2 (ppm) at Mauna Loa
Monthly closings of the Dow-Jones industrial index
Monthly critical radio frequencies
Sunspot time series
Monthly milk production per cow
# Method
In this method we have preprocessing phase which includes normalizing data and separating data for testing and training. In proposed model, Five kernels were selected as the best kernels by trial and error, and these kernels are applied to data. There may be only a few of the kernels that are useful for the problem, and we are not aware of which kernels are useful for our problem, so kernel outputs aggregate by applying coefficients. This combination creates a new secondary space. The output is given to support vector regression to construct a model that predicts values exactly ɛ accurate, which means the predicted values do not deviate more than ɛ from the original data. This model predicts values by using a leave one out model. Each kernel has parameters that need to be set to optimum values in order to get the best results. Hence in the proposed model, the kernel parameters and their weights are learned by the Gray Wolf Optimizer. By running program in consecutive iterations and examining the different values of the parameters, the optimizer learns the best of them which prediction error has been reduced, and finally returns their best value.
# Result
The proposed model is implemented on five standard time series and compared to other method, test based on the RMSE criterion for DJ time series, improved by 1.58 point, Radio time series, improved by 0.178 point, and Sunspot time series, improved by 1.709 point.
没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
【项目背景】 时间序列预测是根据从过去和现在的数据点获得的信息预测未来系统值的过程。使用不同的模型进行时间序列预测的主要目的是使预测具有最大的准确性。 在该知识库中,提出了一种基于支持向量回归和不同核加权组合以及优化器优化参数及其权重的时间序列预测新模型。在能力函数的连续迭代中, 优化器学习粒子阵列的最优值,并使用这些值来评估测试数据。 【数据集】 Mauna Loa二氧化碳(ppm) 道琼斯工业指数的月度收盘情况 每月临界无线电频率 太阳黑子时间序列 每头奶牛月产奶量 【方法】 在该方法中,我们有预处理阶段,包括归一化数据和分离数据进行测试和训练。在该模型中,通过试错选择了5个核作为最佳核,并将这些核应用于数据。 可能只有少数几个核对问题有用,我们不知道哪些核对我们的问题有用,所以核输出通过应用系数来聚合。这种组合创造了一个新的辅助空间。 输出进行支持向量回归,构建预测值精确ɛaccurate的模型,即预测值与原始数据的偏差不超过ɛ。该模型通过使用一个省略模型来预测值。
资源推荐
资源详情
资源评论
收起资源包目录
基于支持向量回归的时间序列预测_Matlab实现(含完整源码+数据集+项目说明).zip (16个子文件)
基于支持向量回归的时间序列预测_Matlab实现(含完整源码+数据集+项目说明)
func_plot.m 3KB
main.m 5KB
fitness.m 3KB
Get_Functions_details.m 7KB
项目说明.txt 2KB
dataset
monthly-closings-of-the-dowjones.xlsx 11KB
monthly-milk-production-pounds-p.xlsx 9KB
monthly-critical-radio-frequenci.xlsx 10KB
wolfs-sunspot-numbers-1700-1988.xlsx 10KB
co2-ppm-mauna-loa-19651980.xlsx 9KB
kernel_g.m 988B
kernel2.m 3KB
kernel.m 656B
initialization.m 2KB
GWO.m 4KB
README.md 3KB
共 16 条
- 1
资源评论
- 想吃点烧烤2023-01-10资源中能够借鉴的内容很多,值得学习的地方也很多,大家一起进步!
Make程序设计
- 粉丝: 5810
- 资源: 3570
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功