Time Series Prediction Using Support Vector Machines: A Survey

By Nicholas I. Sapankevych, Ravi Sankar. IEEE Computational Intelligence Magazine, 2009

1,053 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Electrical and Electronic Engineering, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/40927c5d81988a1151639fad150cbc74f64e0d68

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Electrical and Electronic Engineering

In Management Science and Operations Research

Citations in the corpus

Deep Learning with Long Short-Term Memory for Time Series Prediction (2018, background): “Time series analysis and prediction have been intensively studied for 40 years [4].”
A Strategy for Short-Term Load Forecasting by Support Vector Regression Machines (2013, background): “A detailed analysis why SVR usually performs better than other modeling methods in time series prediction can be found in [27], [32], and [33].”
Deep Transfer Learning for Intelligent Cellular Traffic Prediction Based on Cross-Domain Big Data (2019, methodology): “The SVR method, which is a nonlinear model, can deal with the nonlinearities of cellular traffic, thus obtains better results than LR method.”
Spatiotemporal modeling and prediction in cellular networks: A big data enabled deep learning approach (2017, neutral): “It has been also applied for time series analysis in many applications [21], [23].”
Deep Learning Based Forecasting of Critical Infrastructure Data (2017, methodology): “This model computes a linear regression function in a high dimensional feature space mapping the input data by means of a nonlinear function [38].”
Enhancing the momentum strategy through deep regression (2019, background): “Support vector regression (SVR) has been used extensively in the past to make predictions of diverse set of time-series variables (Sapankevych and Sankar 2009), and in particular, it has also found success in ﬁnancial applications for, say, predicting the stock prices (Kim and Heo 2017) and…”
Chaotic Time Series Prediction Based on a Novel Robust Echo State Network (2012, methodology): “Several neural network models and kernel-based methods have been applied for time series forecasting, such as multilayer perceptron [4], radial basis function neural network [5], [6], support vector machine [7]–[9], relevant vector machine (RVM) [10], fuzzy model [11], extended Kalman filter [12], r”
Short-term load forecasting based on multivariate time series prediction and weighted neural network with random weights and kernels (2018, background): “However, most of the published literatures are concerned with univariate time series other than multivariate time series [13,14].”
Multivariate Temporal Convolutional Network: A Deep Neural Networks Approach for Multivariate Time Series Forecasting (2019, methodology): “Advantages Challenges AUTOREGRESSIVE [8] Simple and efﬁcient for lower order models Nonlinear, multivariable and non-stationary SVR [18] Nonlinear and high-dimensional Selection of free parameters, NOT suitable for big data Hybrid VMD and ANN [30] Strong explanatory power of mathematics…”
Intraday prediction of Borsa Istanbul using convolutional neural networks and feature correlations (2017, background): “For this purpose, filter-based feature selection methods were combined, SVM classifier was trained and finally the voting scheme was applied.”
A Bias and Variance Analysis for Multistep-Ahead Time Series Forecasting (2016, neutral): “Examples include K -nearest neighbor (KNN) [6], [7], neural networks (NNs) [8]–[10], support vector machines [11], boosting [12], and fuzzy NNs [13]–[15].”
A novel wavelet-SVM short-time passenger flow prediction in Beijing subway system (2015, background): “As stated by Sapankevych and Sankar [22], the underlying motivation for using SVM is the ability of this methodology to accurately forecast time series data when the underlying system processes are typically nonlinear, nonstationary and not defined a priori.”
Position-Based Content Attention for Time Series Forecasting with Sequence-to-Sequence RNNs (2017, background): “More recently, based on the development of statistical machine learning, time series prediction has been formulated as a regression problem typically solved with Support Vector Machines (SVM) [13] and, even more recently, with Random Forests (RF) [14,15].”