A Study On Stock Price Prediction Using Arima Model
No Thumbnail Available
Files
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
NHCE
Abstract
In financial analytics, stock price prediction is still and difficult tasks, nonlinear, and dynamic, undergoing various changes in every moment. It is ever more a quest by investors, traders and the financial institutions to be accurate forecasting models so make informed investment decisions to manage risk properly. It is within this framework that statistical time series modelling such as ARIMA (Auto Regressive Integrated Moving Average) has come in support as a basic instrument to be used in the study and forecasts of financial time series data. This paper seeks to review of the predicting stock prices, identify its strengths and weaknesses and give its application and implication in relation to financial forecasting.
ARIMA is a family of linear models, and it joins three important factors, autoregression (AR), differencing the data to render it stationary (I), and moving average (MA). It is represented by ARIMA(p, d, q) wherein p is the lag observation ,d is the number of differencing to get the model stationary and moving average window. The ability of the model to predict future points in a time series and use the dependencies between an observation and past values is one of the main strengths of it. Historical stock prices of specific companies with references to mainly the closing date of stock that was of keen interest in the study. The missing data was imputed after the pre-processing of the data and the first or second order differencing was used data steady. Augmented Dickey-Fuller test was tools that were used in confirming stationarity, and plots, including partial autocorrelation (PACF) and autocorrelation (ACF), could show the optimal parameters.