Assimilation of Principal Component Analysis and Wavelet with Kernel Support Vector Regression for Medium-Term Financial Time Series Forecasting

Abstract

Entities and institutional financiers have gained a lot of growth from financial time series forecasting in recent times. But the major challenges of financial time series data are the high noise and complexity of its nature. Researchers in recent times have successfully engaged the application of support vector regression (SVR) to conquer this challenge. In this study principal component analysis (PCA) is applied to extract the low dimensionality and efficient feature information, while wavelet is used to pre-process the extracted features in other to nu1llify the influence of the noise in the features with a KSVR based forecasting model. The analysis is carried out based on the quarterly tax revenue data of 39 years from the first quarter of 1981 to the last quarter of 2016. The forecasting is made for ten quarters ahead. The initial empirical result shows that the multicollinearity has been reduced to zero (0), and the analytic result reveals that the proposed model PCA-W-KSVR outperforms KSVR, PCA-KSVR, and W-KSVR in terms of MAE, MAPE, MSE and RMSE

Principal component analysis of day‐ahead electricity price forecasting in CAISO and its implications for highly integrated renewable energy markets

“…PCA is often used for studies involving large data sizes, as it can effectively reduce the dimensionality of multivariate analysis (Alhassan et al, 2020; Hopwood et al, 2020; Wang et al, 2019). This reduction in dimensionality is achieved by replacing a number of variables with a smaller number of PCs that effectively summarize a previously large part of the variation of the data.…”

“…Each methodology has its merits and limitations, contingent upon the specific objectives of the forecasting task. For instance, statistical time series analysis leverages models like autoregressive integrated moving average (ARIMA) (Yang et al, 2017), autoregressive (AR) or autoregressive-exogenous ARX(n, m) models (Nowotarski & Weron, 2016), dynamic regression, or Generalized Auto-Regressive Conditional Heteroskedasticity (GARCH)-based models (Fan, 2016;Hong & Wu, 2012), exponential smoothing (Forootan et al, 2022), and vector autoregression (VAR) (Alhassan et al, 2020) to analyze and understand electricity prices' statistical behavior. In contrast, machine learning techniques rely on algorithms and data-driven approaches like PCA, ANN, SVM, decision trees, and random forests to make predictions based on historical data (Bose, 2017;Nikkhah et al, 2019).…”

“…To navigate these complexities, experts employ advanced analytics such as extreme learning machine (ELM), kernel‐based extreme learning machine (KELM), bivariate empirical mode decomposition (BED), and principal component analysis (PCA). For instance, Chen et al (2012) employed ELM alongside bootstrapping for price forecasting in the Australian National Electricity Market, and Alhassan et al (2020) used PCA combined with wavelet and kernel support vector regression methods to forecast financial time series data. Yang et al (2017) combined wavelet transform, ARMA, and KELM methods for enhanced accuracy in their predictions.…”