ReliaMatch: Semi-Supervised Classification with Reliable Match

Deep learning has been widely used in various tasks such as computer vision, natural language processing, predictive analysis, and recommendation systems the past decade. However, practical scenarios often lack labeled data, posing challenges for traditional supervised methods. Semi-supervised classification methods address this by leveraging both unlabeled data to enhance model performance, but they face effectively utilizing distinguishing reliable information from unreliable sources. This paper introduced ReliaMatch, a semi-supervised method that addresses these using confidence threshold. It incorporates curriculum stage, feature filtering, pseudo-label filtering improve accuracy reliability. The module eliminates ambiguous semantic features comparing space. removes pseudo-labels with low confidence, enhancing algorithm ReliaMatch employs training mode, gradually increasing dataset difficulty combining selected samples data. approach enhances performance. Experimental results show overcomes associated underutilization of introduction error information, outperforming strategy classification.