Methodological Evaluation of Smallholder Farms Systems in Ethiopia: Time-Series Forecasting Model for System Reliability Analysis

The study focuses on evaluating smallholder farms in Ethiopia, a region experiencing significant land degradation. A comparative analysis using statistical methods to evaluate the performance of various forecasting models under different conditions was conducted. The findings indicate that a hybrid ARIMA-GARCH model provided the most reliable forecasts, with an accuracy rate exceeding 80% in predicting system reliability. The hybrid ARIMA-GARCH model demonstrated superior predictive capabilities for smallholder farms' systems in Ethiopia. Further research should explore the application of these models in other regions and incorporate additional variables to enhance forecasting precision. Smallholder Farms, Time-Series Forecasting, Reliability Analysis, Hybrid Model, ARIMA-GARCH The empirical specification follows Y=₀+^ X+, and inference is reported with uncertainty-aware statistical criteria.