Developing a CR-SCAD algorithm for fibrosis and inflammatory activity analysis of chronic hepatitis C
摘要
Analyzing the degree of hepatic fibrosis and the level of inflammatory activity plays an important role in the clinical management of chronic hepatitis C. This study proposes a classification method, termed CR-SCAD, which integrates collaborative representation (CR) with the smoothly clipped absolute deviation (SCAD) penalty to achieve dense representation and sparse variable selection simultaneously. The method was evaluated on 123 clinical cases using 14 serological indices. Experimental results demonstrated that CR-SCAD achieved 69.92% accuracy for fibrosis staging and 65.04% for inflammatory activity grading, substantially outperforming baselines such as SCAD (44.72% and 62.60%) and CR (3.25% and 32.52%). These findings indicate that CR-SCAD provides an effective auxiliary technique for automatically assessing fibrosis and inflammatory activity in chronic hepatitis C patients.