Computational Analysis of Traditional Ayurvedic Knowledge: A Study of Medicinal Plants from the Marathwada Region

Abstract: Medicinal plants are a part of traditional healthcare especially in areas where old knowledge about plants has been kept alive through Ayurvedic books and local customs. Much of this information is spread out across many books, government websites and traditional medical sources which makes it hard for computers to analyze. The study expands a previously developed dataset from 21 to 100 medicinal plants associated with the Marathwada region of Maharashtra, India. Each plant entry was organized using 21 attributes to turn mixed up text information into a format that computers can understand. Exploratory analysis was conducted to examine relationships among Ayurvedic properties including rasa, virya, and vipaka. Preliminary observations show that there are connections between taste categories and how plants affect the body. For instance, the analysis indicated that bitter taste frequently co-occurs with astringent attributes and that pungent post digestive effect is very common among the selected plants. The results provide an initial view of the relationships among Ayurvedic functional properties in the selected plants. Although the dataset is limited in size, this study shows how organizing knowledge in a systematic way can make it possible for computers to analyze and can help lay the groundwork for future research on medicinal plants using AI. The study is focused on computational analysis of Ayurvedic functional properties and bridging the gap between scattered textual knowledge into AI driven healthcare research.