E Axiotis, A Kontogiannis, E Kalpoutzakis, G Giannakopoulos

Abstract:
Traditional Medicine (TM) is an interdisciplinary field of research based both on anthropological and scientific approaches. The database derived from TM prescriptions contains plant extracts and ingredients with unique pharmacological activities. New technologies are nowadays implemented to decode information entangled in natural product metabolites, selected over millions of years of evolution, and over thousands of years of written records of our civilization. However, the volume and the complexity of sources that need to be monitored and classified, present the so called “big data” challenge. Artificial intelligence (AI) and algorithms of Machine Learning (ML), improve the related research workflows, increasing velocity and precision. In addition to bibliographic databases the necessary information sourced from healers needs to be filtered to obtain fruitful data. The use of personalized data mining techniques that is adaptive in real-time by the experts of ethnopharmacology, is still under evolution. In our model, a ML algorithm is used as an AI system in order to discriminate between relevant and irrelevant documents, biased by decisions provided by a human expert, improving the effectiveness and the efficiency of the identification and retrieval of the appropriate information. The evolution of such a system will contribute to knowledge sharing and scientific advancement in ethnopharmacology.