Government Scheme Prediction and Identification using K-Means Clustering

Abstract

In an era of digital governance, citizens are often overwhelmed by the sheer volume of welfare schemes, leading to a "discovery gap" where eligible individuals fail to benefit from available resources. This research proposes an intelligent framework designed to bridge this gap by automating the identification and recommendation of government schemes using K-means and Unsupervised Machine Learning. By deploying the K-Means clustering algorithm, we categorize a vast repository of government schemes into distinct thematic clusters based on semantic keyword extraction and feature vectorization. The model processes unstructured scheme descriptions to identify underlying patterns, allowing for precise matching between user-inputted profiles and relevant policy categories. Our experimental results demonstrate that K-Means effectively delineates schemes into actionable domains—such as healthcare, education, agriculture, and financial aid—providing a scalable solution for personalized scheme retrieval. This approach not only enhances administrative transparency but also empowers citizens by transforming complex public policy data into accessible, personalized guidance.