ASAYA Machine Learning
We use a decision-tree based on modeled knowledge gathered from therapists, academic researches, and official guidelines. This decision-tree takes the patient’s data through a series of queries, and based on the gathered knowledge, then outputs the suitable chemovars, titration protocols and strains for the patient.
Mass-data from patients’ feedback trigger a collaborative filtering algorithm that is used to perform a multi-patient comparison of personal details and experiences, in order to provide the patient with the most suitable treatment, based on other patients’ collective experiences. We also integrate boosting algorithms, which can rank features based on their influence on the treatment recommendation and the between-patient comparison
We build a robust architecture that enables easy integration with variable sources of data such as: Connected Medical Devices transmitting dosage details and time-stamps, wearables recording physio-metrics and more. Such data becomes another module in Asaya’s Machine Learning infrastructure.