Local density comparison sampling inspection method and device for quality detection and medium

The invention discloses a local density comparison sampling inspection method and device for quality detection, and a medium. The method comprises the following steps: S1, carrying out feature extraction on collected batch data of a semiconductor production etching process; s2, performing data standardization on the data after feature extraction; s3, performing feature selection on the standardized feature data; s4, clustering the data after feature selection to obtain a center sample of each class; s5, scoring the data after feature selection, calculating an outlier score of each sample by adopting a local density comparison algorithm, and sorting the outlier scores from large to small to obtain an outlier score sorting table; and S6, performing sampling inspection on the samples, and extracting the central sample obtained in the step S4 and k samples with the highest outlier score in the sorting table obtained in the step S5 for quality detection. According to the method, the data analysis process is simple