In this paper,we propose a framework for evaluating the sensitivity of anticancer drugs based on contrastive learning(SSLGP).In this study,we have designed a deep autoencoder that incorporates contrastive learning strategies to effectively extract information from high-dimensional gene expression features.We have then integrated it into the XGBoost algorithm for training and constructing a drug sensitivity prediction model.To assess the predictive performance of our framework,we have utilized eight publicly available anticancer drug datasets and conducted experiments to compare our method with others.The experimental results demonstrate that the proposed framework achieves a significantly high AUC score of 0.670,exhibiting an average improvement of 5.18%and a maximum improvement of 10.56%compared to other existing methods.These findings highlight the potential value of this framework in clinical practice for facilitating drug selection in patient management.
维普
万方数据