Predictive Analysis of Cotton and Turmeric Prices: Understanding its Influential Factors on the National Commodities and Derivatives Exchange.

Abstract

In the modern era of investing, it's become increasingly common for investors to manage diverse portfolios that involve a combination of equities, debt, currencies, and commodities. Recently, the commodities market has emerged as a popular choice among investors, leading to an uptick in research in this area. With many beneficiaries, including farmers, consumers, arbitragers, speculators, and industrialists, the commodities market experiences high traded volumes and constant demand and supply fluctuations. Agricultural commodities traded on the National Commodities and Derivatives Exchange of India have seen an increase in trading volume, contributing to price volatility and risk. Additionally, these commodities are subject to seasonality, as they are only traded during certain times of the year, making investment decisions even more challenging. However, tools such as derivatives can be used to hedge against risk and mitigate potential losses.

In order to hedge their investments effectively, investors must understand and be able to forecast future price trends of commodities and volumes. Fortunately, there are models available to aid investors in predicting future trends, such as Linear Regression, K-Nearest Neighbor (KNN), AutoRegressive Integrated Moving Average (ARIMA), and Long Short-Term Memory (LSTM).

This study anticipates the future prices of certain agricultural goods traded in India. After conducting literature reviews, we have determined that the ARIMA model is a reliable source for precise results. As a result, our study employs the ARIMA model to predict the future prices of Cotton and Turmeric traded on the NCDEX. We accomplished this by utilizing an analytical methodology with the ADF Econometrics model to verify stationarity. Our data was sourced from secondary sources available on the NCDEX website.

Our study anticipates that the prices of these commodities will remain stationary at the first-order difference. Meanwhile, the prices will be stationary at extremes at the second-order difference, indicating the suitability for a further predictive model. Based on information criteria, the models with the lowest AIC and BIC scores are deemed the most suitable for predicting the future prices of Cotton and Turmeric in India.