Enhancing density-based clustering: Parameter reduction and outlier detection