Fig. 9From: A novel algorithm for fast and scalable subspace clustering of high-dimensional dataRuntime comparison for fixed data size=1595. Parameters for CLIQUE: τ=0.003, g r i d S i z e=3. Parameters for SUBSCALE, INSCY and SUBCLU (similar values): ε=0.001, τ=3, m i n S i z e=4. CLIQUE crashed after 10 dimensions. SUBCLU didn’t give meaningful clusters. Dataset used: D05, D10, D15, D20, D25Back to article page