TIME SERIES ANALYSIS OF REAL ESTATE PRICES USING THE BOX-JENKINS APPROACH

Abstract

Forecasting is an important activity in economics, commerce, marketing and various branches of science. This thesis is concerned with forecasting method based on the use of time series analysis. Time series analysis is important to forecast the prices to identify future ups and downs, and turning points in the real estate market. From the purpose of need and importance, the main purpose of this thesis is to forecast the future trends in Pakistan Plot prices. Analyzing the time-oriented data and forecasting the future values of a time series are among the most important problems that analysts face in many fields, ranging from finance and economics, to managing production operations, to the analysis of political and social policy sessions, to investigating the impact of humans and the policy decisions that they make on the environment. Consequently, there is a large group of people in a variety of fields including statistics, science, engineering, and public policy who need to understand some basic concept of time series analysis and forecasting. Data obtained from observations collected sequentially overtime are extremely common. In this thesis, the focus is on univariate time series forecasting where different statistical methods and techniques are useful. A somewhat unique feature of time series and their models is that we usually cannot assume that the observations arise independently from a common population. Successful time series analysis and forecasting requires that the analyst interact with computer software. The techniques and algorithms are just not suitable to manual calculations. The chosen to demonstrate the techniques presented using two packages, R and Excel. However, this thesis focuses more on applied and directed towards model fitting and iv data analysis, for which R is mostly used. This package are selected because it is widely used in practice and has generally good capability for analyzing time series data and generating forecasts. The main features of many time series are trends and seasonal variations that can be modelled deterministically with mathematical function of time. For seasonal time series, the nonstationary ARIMA process is extended to account for a multiplicative seasonal component. These models form the framework for  Expressing various forms of stationary and non-stationary behavior in time series.  Producing optimal forecasts for a time series from its own current and past values. The topics covered include  What an ARIMA forecasting model is used for.  Why stationarity is an important concept for ARIMA processes.  How to select ARIMA models through an iterative three-stage procedure i-e Model identification, Estimation and diagnostic checking.  When, why and how Box-Jenkins methodology can be used for forecasting.