Time Series for Data Sciences: Analysis and Forecasting
Anoop Kumar Chaturvedi · Journal of the Royal Statistical Society Series A (Statistics in Society) · 2023
The practitioners of data science often have to deal the time series data with the objective of developing models and forecasts. The book is aimed at providing a practical and accessible guide of time series analysis, modelling, and forecasting methods with main emphasis on applications using R programming language. The book consists of 11 chapters with first chapter discussing the motivation behind analysing time series data, the tools to access and manipulate time series data to make it compatible with R. The chapter also discusses various time series datasets, which have been used in successive chapters. Chapter two illustrates various preliminary analysis techniques such as decomposing it into various components, and forecasting methods based on exponential smoothing. Chapter three discusses statistical fundamentals which form the foundation for time series models. The focus of chapter four is the frequency domain analysis and smoothing of time series data. Chapter five and six introduce stationary autoregressive (AR), moving average (MA), and mixed autoregressive moving average (ARMA) models, model identification and fitting, and obtaining forecasting using these models. Chapter seven considers fitting and forecasting using autoregressibe integrated moving average models, seasonally differenced ARMA models, and autoregressive conditional heteroscedastic and generalized autoregressive conditional heteroscedastic processes. The mainstay of chapter eight is linear time series regression involving correlated errors and explain procedures for fitting and forecasting. Chapter nine explains diagnostics checking for the adequacy of a candidate model and residual analysis. Chapter 10 presents multivariate time series models including vector autoregressive (VAR) and seasonal VAR processes. The last chapter 11 explores the neural network-based time series models, their architecture, fitting, and forecasting.