Paper Title
Investigating The Overall Mortality Risk Linked to Acute Myeloid Leukemiathe Utilization of Deep Learning-Based Fuzzy Systems

Abstract
Leukemia is a prominent cause of cancer-related deaths globally, ranking as the 11th most deadly cancer. Among the various types of leukemia, Acute Myeloid Leukemia (AML) stands out with a five-year mortality rate of 68.3%. The impact of leukemia on cancer-related mortality is substantial.AML usually involves multiple genomic mutations, some of may be key regulatory factors in hematopoietic cell differentiation and proliferation.So, this research introduces the Fuzzybased RNNCoxPH analytic approach for identifying missense variants associated with a high risk of all-cause mortality in Acute Myeloid Leukemia (AML). The proposed approach combines fuzzy logic withRecurrent Neural Networks (RNNs) and Cox proportional hazards regression (CoxPH) to address the challenges of high-throughput data variability. Fuzzy logic enhances risk estimation by classifying the membership grade of missense variants.The study utilizes the TCGA-LAML clinicopathological information and Mutation dataset toderive four risk score models: RNN, CoxPH, RNNCoxPH Addition, and RNNCoxPH Multiplication to analyzeeight risk factors.The Fuzzy-based RNNCoxPH model achieves a balanced accuracy of 93.04%, outperforming other methods. This approach demonstrates efficacy in identifying and classifying missense variants associated with mortality risk in AML, potentiallyadvancing cancer research. Keywords - Acute Myeloid Leukemia,RNN, Fuzzy logic, Hybrid model,The Cancer Genome Atlas Program