<p>Sufficient dimension reduction (SDR) has proven to be a useful tool for data visualization and information retrieval in high dimensional data analysis. Several well-established SDR methods investigate the inverse conditional moments of the predictors given the response. Classical slicing and cumulative slicing are two commonly used framework in the estimation. Motivated by the connection between slicing and imbalanced learning, we propose a SMOTE (synthetic minority oversampling technique) enhanced cumulative estimation framework. Extensive simulation studies and a real data application show the efficacy of the newly proposed method.</p>