Software for Information Support of Composition and Property Search for Elastomeric Material Formulations

Abstract

The software suite «Rubber C.2» has been developed for information support of composition and property search for elastomeric material formulations. The suite provides centralized management of formulations and data quality control based on Z‑score: completeness of data entry, structural anomalies, and statistical outliers of concentrations are analyzed. Fuzzy search for analogues by composition is implemented with the choice of a similarity metric; the results are ranked and exported into a report. To unify the names of ingredients and physicomechanical properties, synonym thesauri have been created. Using association analysis, stable combinations of rubber compound components are identified with evaluation of support, confidence, and lift; the results are visualized as a graph and a heat map. The machine learning module allows the user to select up to six ingredients, predicts the target property based on their mass fractions using regression models, and evaluates the influence of each component on this property. The program includes reference directories, formulation editing tools, and interactive diagrams of formulation distribution by rubber type. The application of «Rubber C.2» helps to reduce the development cycle of new elastomeric compositions.