Experimental data about branching ratios for the products of dissociative recombination of polyatomic ions are presently the unique information source available to modelers of natural or laboratory chemical plasmas. Yet, because of limitations in the measurement techniques, data for many ions are incomplete. In particular, the repartition of hydrogen atoms among the fragments of hydrocarbons ions is often not available. A consequence is that proper implementation of dissociative recombination processes in chemical models is difficult, and many models ignore invaluable data. We propose a novel probabilistic approach based on Dirichlet-type distributions, enabling modelers to fully account for the available information. As an application, we consider the production rate of radicals through dissociative recombination in an ionospheric chemistry model of Titan, the largest moon of Saturn. We show how the complete scheme of dissociative recombination products derived with our method dramatically affects these rates in comparison with the simplistic H-loss mechanism implemented by default in all recent models.