A machine learning platform for development of low flammability polymers

Flammability index (FI) and cone calorimetry outcomes, such as maximum heat release rate, null me to igni null on, total smoke release, and fi re growth rate, are cri null cal factors in evalua null ng the fi re safety of polymers. However, predic null ng these proper null es is challenging due to the complexity of material behavior under heat exposure. In this work, we inves null gate the use of machine learning (ML) techniques to predict these fl ammability metrics. We generated synthe null c polymers using Synthe null c Data Vault to augment the experimental dataset. Our comprehensive ML inves null ga null on employed both our polymer descriptors and those generated by the RDkit library. Despite the challenges of limited experimental data, our models demonstrate the poten null al to accurately predict FI and cone calorimetry outcomes, which could be instrumental in designing safer polymers. Addi null onally, we developed POLYCOMPRED, a module integrated into the cloud based MatVerse pla null orm, providing an accessible, web based interface for fl ammability predic null on. This work provides not only the predic null ve modeling of polymer fl ammability but also an interac null ve analysis tool for the discovery and design of new materials with tailored fi re resistant proper null es. 2