Necessity of Incorporating Break-in-trend in Aggregate Time Series Data: Evidence from Agricultural GDP in India

This paper examines the importance of break analysis in time series data for obtaining the best-fitted curve, using agricultural GDP data in India as a case study. Through visual assessment and statistical analysis, it is observed that the curve with breaks demonstrates a closer fit to the actual line and captures the underlying data patterns more accurately. Statistical measures, including R-squared values, Bayesian Information Criterion, average absolute residuals, and Normality test, further support the superiority of the curve with breaks in terms of explaining data variation, trade-off between fit and complexity, accuracy in prediction, and adherence to normality assumptions. Therefore, incorporating break analysis in curve fitting proves essential for achieving optimal results and improving understanding of time series patterns. Policymakers should focus on using curve-fitting models with breaks for analyzing agricultural GDP data because they perform better at capturing fluctuations and patterns, aiding informed policy decisions.