Price Variation Analysis for Food Security Through Clustering Techniques and ARIMA Forecasting Using R Language

Food prices, essential economic stability, and consumer welfare indicators fluctuate globally due to various economic factors. Inflation has driven food price volatility, impacting purchasing power and food accessibility. This study aims to implement a clustering model to analyze food price patterns by grouping food items with similar pricing characteristics and assessing the impact of inflation. Clustering methods K-means, Partition Around Medoids (PAM), and fuzzy C-means (FCM) were evaluated using average silhouette coefficient, gap statistic, and sum of squared distances. Results revealed that the fuzzy C-means algorithm provided optimal clustering with a silhouette score of 93.2% using five clusters, indicating a strong separation and cohesion within clusters. An Autoregressive Integrated Moving Average (ARIMA) (2,0,2) model with a non-zero mean was then used to forecast regional food prices in the Philippines, revealing moderate autoregressive and moving average influences, a stabilized mean price of approximately 31.32, and a trend projecting increases followed by steadiness. This study introduces a combined use of clustering and time series forecasting to analyze and project food price behaviors across Philippine regions, offering an innovative framework for segment-based market insights and policy-focused intervention. The findings include a Data-Driven Food Security Framework that combines price clustering, demand analysis, and regional forecasting to provide valuable information for enhancing food security and informing pricing strategies in the country.