TY - GEN
T1 - Chronic Kidney Disease Survival Prediction with Artificial Neural Networks
AU - Zhang, Hanyu
AU - Hung, Che Lun
AU - Chu, William Cheng Chung
AU - Chiu, Ping Fang
AU - Tang, Chuan Yi
N1 - Publisher Copyright:
© 2018 IEEE.
PY - 2019/1/21
Y1 - 2019/1/21
N2 - The main objective of this work is to investigate the performance of Artificial Neural Network (ANN) models while applying to the survivability prediction on Chronic Kidney Disease (CKD) patients. Chronic kidney disease patients suffer from losing the primary function of the kidney, i.e., blood filtration, gradually over time. At the end stage, regular hemodialysis or kidney transplant are required to survive. During the process of disease development, patients may also suffer from complications of acidosis, anemia, diabetes, high blood pressure or neuropathy, etc., which in turn affects patients' quotidian life. It is reported that the median survival time of late-stage patients is only about 3 years. Evaluating precisely the condition of patients is of great importance as it would greatly help to decide appropriate care, medications or medical interventions needed, which among them have a complex interrelationship and influence the outcome of the individual patient. An accurate prediction model would hopefully be able to fit into that role and may be used to revise current treatment. However, due to the complex nature of the problem, as multiple interrelated factors may influence the patient's survival, finding such a model is a challenging task. Recently, artificial intelligence (AI), especially the deep learning technique has become a thriving field, owing to the rise of computational capability. As the approach needs no human specialist in specifying explicit knowledge in advance, but gathers knowledge automatically from amounts of data by building a hierarchy of concepts with complicated ones upon simpler ones, it has already been applied successfully in intuitive problems, like understanding speeches or images, making diagnoses in medicine or self-driving cars, etc. And conversely, technological advancements have been made available through practical applications, profiting from which we may develop prediction models for chronic kidney disease survivability. In this research, data preprocessing, data transformations, and artificial neural networks are used to establish the mapping from many clinical factors to the patient's survival. The computational results are also reported in the paper.
AB - The main objective of this work is to investigate the performance of Artificial Neural Network (ANN) models while applying to the survivability prediction on Chronic Kidney Disease (CKD) patients. Chronic kidney disease patients suffer from losing the primary function of the kidney, i.e., blood filtration, gradually over time. At the end stage, regular hemodialysis or kidney transplant are required to survive. During the process of disease development, patients may also suffer from complications of acidosis, anemia, diabetes, high blood pressure or neuropathy, etc., which in turn affects patients' quotidian life. It is reported that the median survival time of late-stage patients is only about 3 years. Evaluating precisely the condition of patients is of great importance as it would greatly help to decide appropriate care, medications or medical interventions needed, which among them have a complex interrelationship and influence the outcome of the individual patient. An accurate prediction model would hopefully be able to fit into that role and may be used to revise current treatment. However, due to the complex nature of the problem, as multiple interrelated factors may influence the patient's survival, finding such a model is a challenging task. Recently, artificial intelligence (AI), especially the deep learning technique has become a thriving field, owing to the rise of computational capability. As the approach needs no human specialist in specifying explicit knowledge in advance, but gathers knowledge automatically from amounts of data by building a hierarchy of concepts with complicated ones upon simpler ones, it has already been applied successfully in intuitive problems, like understanding speeches or images, making diagnoses in medicine or self-driving cars, etc. And conversely, technological advancements have been made available through practical applications, profiting from which we may develop prediction models for chronic kidney disease survivability. In this research, data preprocessing, data transformations, and artificial neural networks are used to establish the mapping from many clinical factors to the patient's survival. The computational results are also reported in the paper.
KW - Artificial neural networks
KW - LASSO
KW - chronic kidney disease
KW - classification
KW - deep learning
KW - disease prediction
UR - http://www.scopus.com/inward/record.url?scp=85062550699&partnerID=8YFLogxK
U2 - 10.1109/BIBM.2018.8621294
DO - 10.1109/BIBM.2018.8621294
M3 - Conference contribution
AN - SCOPUS:85062550699
T3 - Proceedings - 2018 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2018
SP - 1351
EP - 1356
BT - Proceedings - 2018 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2018
A2 - Schmidt, Harald
A2 - Griol, David
A2 - Wang, Haiying
A2 - Baumbach, Jan
A2 - Zheng, Huiru
A2 - Callejas, Zoraida
A2 - Hu, Xiaohua
A2 - Dickerson, Julie
A2 - Zhang, Le
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2018 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2018
Y2 - 3 December 2018 through 6 December 2018
ER -