This work addresses the computational difficulties involved with the simulation of the nonlinear ring oscillator system in the offline and online phase. Conventional POD has offline basis extraction burden along with online computation in nonlinear systems. The computation of the nonlinear term can be improved with a discrete empirical interpolation method (DEIM). However, the issue of offline computation in POD still persists. This work proposed approximate snapshot ensemble generation for basis extraction in nonlinear model order reduction to improve offline computation. Hence, proposed approach reduces offline basis computation, and DEIM reduces the online computational burden. This approach is tested on a large ring oscillator model with two different sets of inputs.