Computational simulations of materials have advanced significantly in recent years, shifting from analyzing experimental observation to providing capability to predict of mechanical behaviors for use in advanced material development. Machine learning has been a topic of great interest in recent years within the engineering community. In this study, we will present an efficient machine learning framework that has potential to significantly improve predictive capability of computational modeling, which in turn would reduce the number of prototypical experimental validation. Based on data obtained from dislocation dynamics modeling, we parameterize intrinsically and predict characterization of small-scale plasticity in single crystalline micropillars. The algorithm is fed Cu micropillar simulation data from three-dimensional dislocation dynamics simulations with varying conditions of sample size, initial dislocation density, and other nucleation criterion.