Time Series Analysis/ Statistical Forecasting

Introduction

Time series analysis is linked to the works of Yule and Walker in the 1920s and 1930s. Time series was introduced to obtain an understanding of the underlying forces and structure that produce data. Statistical forecasting is a product of time series. In this regard, time series analysis is used in economic forecasting, sales forecasting, budgeting analysis and yield projections.

The use of time series analysis in different sectors has been promoted by technological advancements. For example, the internet of things and the digitization of various sectors such as healthcare have increased the need of to analyse big data.

Time series analysis tries to answer the question, how did the past influence the future? In this regard, time is one that can be used to predict the future. Time series uses time to analyse possible outcomes of the future. There are various models that are various models or fitting time series. The models include Box-Jenkins ARIMA models, Box-Jenkins Multivariate models and Holt-Winters Exponential Smoothing.

Time is an important factor in any organization. Time series analysis is one of the models that are used by organizations to predict the changes in various aspects that vary with time. Due to continuous innovations in forecasting technologies, many organizations have developed various time series techniques to predict future changes in the organization.

The main aim of forecasting is to understand the market changes at any given time. The time interval to be considered may be annual, monthly, weekly, daily or even hourly.

General concept and trends

Time series depend on past data on various aspects such as monthly sales revenue and weekly overheads. The data is plotted in a chronological manner to yield a statistical inference. Such statistical inference depends on the four components of the time series. The first component is the seasonal variation which tends to repeat over a specific period of time such as weekly, daily, monthly or even annually.

Trend variations tend to move up or down in a predictable pattern. In addition, cyclical variations tend to follow economic cycles in a predictable pattern. The fourth element is the random variation which does not follow any pattern. Random variation is not easily predictable.