TRIS is a fast and accurate approach for splitting and expanding source code identifiers. It treats identifier splitting as an optimization problem and uses a two-phase strategy: 1) building a dictionary of terms and transformation rules, and 2) applying these to split identifiers. A case study found TRIS outperformed other approaches like Camel Case splitting, Samurai, and TIDIER in accuracy on two software projects. TRIS also had higher accuracy than Samurai and GenTest on a large dataset of identifier split correctness.