Classification Techniques to Predict the Risk of Myelodysplastic Syndromes using Deep Learning

This paper presents a new hybrid deep learning model of CNN and Logistic Regression (LR) for risk prediction of Myelodysplastic syndromes (MDS). This model uses image processing procedures for patient's risk factor detection. The goal is to predict MDS risk with high specificity and compassion. The hybrid neural network model first obtains the multilevel feature representation of medical imaging data. Then these features are inputted into the logistic regression classifier to predict the risk. Experiments are run on the real-world MDS dataset to evaluate the model performance. Evidence demonstrates that the suggested approach surpasses traditional machine learning techniques and other deep learning methods for MDS risk prediction. The results of this study demonstrate the promise of the proposed hybrid model for risk prediction in medical applications. It is expected that the hybrid deep learning model suggested here paper will provide an effective way for early diagnosis of MDS and reduce the death hazard of human beings.