DEVELOPMENT OF THE SOFTWARE FOR IMPLEMENTATION OF INTELLECTUAL TECHNOLOGY FOR FORECASTING DEPENDENCE «STRUCTURE-PROPERTIES» OF MEDICINAL COMPOUNDS BASED ON MODIFIED ARTIFICIAL IMMUNE SYSTEMS ALGORITHM

Abstract

Nowadays the pharmacological industry is actively developing due to the introduc-
tion of the latest achievements in the field of artificial intelligence (AI).The world's leading drug
manufacturers are collaborating with well-known companies involved in the development of new
innovative artificial intelligence algorithms in order to reduce the cost of the drug production
process and time resources.In this connection, the development of modern intellectual technology
that allows processing a large array of chemical information and predicting the pharmacological
properties of medicinal compounds is an actual task.The prediction of the Quantitative Structure-
Activity Relationship (QSAR) of drug compounds is carried out using a modified Artificial Immune
Systems (AIS) and the Flower Pollination Algorithm (FPA).The FPA algorithm is used to solve the
problem of extracting informative descriptors describing the structure of chemical compounds. The
software architecture for the implementation of the proposed intellectual technology using agent-
based and ontological approaches for the multi-agent platform JADE is presented.The ontological
model of the modified FPA-AIS algorithm is constructed in the Protégé ontology editor, taking into
account the mechanism for evaluating the effectiveness of the modified FPA-AIS algorithm based
on the characteristics: classification error, accuracy, precision, recall and f-measure.