Key Quality Characteristic Identification Based on GSCV-RF-RFE

Identifying key quality characteristics in manufacturing processes plays an important role in quality control. This paper proposes a GSCV-RF-RFE-based method for identifying key quality characteristics of unbalanced production data. In the proposed method, the MICE algorithm is employed for multiple interpolation of the dataset, while the parameters of the random forest algorithm are optimized using the grid search method. Comparative analysis demonstrates that the GSCV-RF model outperforms five algorithms, including XGBoost, in terms of anti-over-fitting performance, and achieves higher model accuracy than MNB. Furthermore, the optimized random forest algorithm is combined with a recursive feature elimination algorithm, and adaptive synthetic sampling is employed to balance the input samples. Spearman correlation analysis is conducted to eliminate strong correlations. Experimental results reveal that the proposed combined strategy achieves a low Class II error rate and a high accuracy rate in quality feature identification, indicating its excellent overall performance.