<![CDATA[Small-scale industries play a crucial role in India’s economic development by generating employment, increasing productivity, and contributing significantly to GDP. However, predicting investment trends in this sector remains challenging due to fluctuating economic conditions and policy shifts. This study analyzes the behavior and predictability of small-scale industrial investment in India using the ARIMA (Auto-Regressive Integrated Moving Average) model. It focuses on assessing the statistical properties of the investment variable namely normality, stationarity, and transformation as the basis for accurate forecasting. A quantitative time series approach was applied using secondary data. The Jarque–Bera test assessed normality, while the Augmented Dickey–Fuller (ADF) and Auto-Correlation Function (ACF) tests examined stationarity and structure. The Jarque–Bera statistic of 4.2314 (p = 0.1205) confirmed that the data is normally distributed. However, the ADF test (p = 0.9992) failed to reject the null hypothesis, indicating non-stationarity. Time series plots showed an exponential upward trend, reflecting steady growth in small-scale investments. To stabilize variance and linearize this trend, a natural logarithmic transformation was applied, producing more consistent and interpretable data for ARIMA forecasting. The findings reveal that while the data meets econometric assumptions, its non-stationarity highlights the dynamic, policy-sensitive nature of India’s small-scale industrial sector. The exponential growth aligns with industrial expansion and supportive government initiatives for MSMEs. The study concludes that despite normal distribution, the variable’s non-stationarity requires transformation for reliable analysis. Future research should employ multivariate and machine learning models to capture external influences and enhance prediction accuracy.]]>
