Stock trend prediction using time series analysis in the nairobi securities exchange

Abstract

Statistical models have been used to come up with with solutions to explore large data sets. In this research I hope to find a solution for investment decision using time series analysis.A time series is a sequence of data points, typically measured at uniform time intervals, it is composed of methods of analysis of data collected in time and obtaining characteristics that can enable to forecast future values. There are many predictive mathematical models, for my research I will use an Autoregressive Integrated Moving Average Filter (ARIMA) model to analyze the share price performance of a firm operating in telecommunication and technology sector of the economy,Safaricom Kenya. Time plot will be used to detect the presence of time series and check the stationarity of the time series, the nature or structure of this dependency will be measured by using auto-covariance, auto-correlation and partial autocorrelation and finally an autoregressive model and moving average model will be fitted to stationary series to predict or forecast the future stock prices.The data on stock prices from year 2009 to 2015 will be collected from the Nairobi Securities Exchange (NSE) trading records. Data entry will be done using Microsoft EXCEL and analyzed using Minitab and Gretl.