Applicability of Time Series Analysis for Forecasting Senior Secondary Student Mathematics Achievement in Nigeria

摘要

This study modelled time-series data on the Mathematics achievement of senior secondary students for a period of twenty-nine (29) years. Trend analysis and Autoregressive Integrated Moving Average (ARIMA) techniques were employed in this study to determine the trend and forecast values that fitted the series (“linear and quadratic” trend). The result of the analysis revealed that the line graph was not stationary because of the change invariance. Unit Root Stationary Test conducted using the Augmented Dicky-Fuller Test (ADF) later confirms the line graph of the achievement of secondary school students in Mathematics, not stationarity because the p-value of 0.124 is greater than 0.05. The time plot of the first difference became stationary. The stationarity condition of the data series are observed by ACF and PACF plots and then checked using the statistic such as ADF. The model for which the values of the criteria are smallest is considered as the best model. ARIMA (2,1,0) is the best fit model based on the Akaike Information Criterion (AIC). Hence the residual analysis for ARIMA (2,1,0) fitted to the achievement of secondary school students in Mathematics data fits quite well.