基于RBF神经网络的脑卒中后吞咽障碍智能诊断建模应用研究
|
摘要
目的本文基于径向基函数(Radial Basis Function,RBF)神经网络对脑卒中后吞咽障碍的诊断工作进行量化描述与智能化建模。方法随机选取两组具有可比性的各129名的临床患者。选取第一组129名患者进行吞咽障碍床旁临床评估量表的评估,将专家诊断的数据和结果进行量化,构建基于RBF神经网络的诊断模型;第二组129例患者进行量表评估,直接输入计算机,使计算机输出诊断结果;第二组患者量表同时交给专家,诊断出的结果与计算机诊断模型输出的第二组的诊断结果进行比对。结果发现第二组患者计算机诊断结果和专家诊断结果不具有统计学差异。结论本文构建了基于RBF神经网络的脑卒中后吞咽障碍智能诊断模型,并能够学习专家诊断经验,最后与几种典型机器学习方法进行了对比,验证了所建神经网络模型的准确性和优势。
Objective The intelligent model of the diagnosis of post-stroke dysphagia was built and described quantitatively based on Radial Basis Function(RBF) network. Methods Firstly, two groups of 129 patients with no significant difference in age and sex were selected from each group. 129 patients in the first group were assessed by the bedside clinical assessment scale for dysphagia along with quantifying the data and results of expert diagnosis, and the diagnostic model was constructed based on RBF neural network;129 patients in the second group were assessed by the scale and then input directly into the computer so that the computer could output the diagnostic results; the second group was given to experts for diagnosis at the same time, and the results obtained by experts were compared with those obtained by experts. The diagnostic results of the second group of computer outputs were compared. Results We found that there was no statistical difference between the results of computer diagnosis and expert diagnosis in the second group. Conclusion An intelligent diagnosis model of post-stroke dysphagia based on RBF neural network is constructed, and expert diagnosis experience can be learned. Finally, compared with several typical machine learning methods,the accuracy and advantages of the established RBF neural network model is verified.
Objective The intelligent model of the diagnosis of post-stroke dysphagia was built and described quantitatively based on Radial Basis Function(RBF) network. Methods Firstly, two groups of 129 patients with no significant difference in age and sex were selected from each group. 129 patients in the first group were assessed by the bedside clinical assessment scale for dysphagia along with quantifying the data and results of expert diagnosis, and the diagnostic model was constructed based on RBF neural network;129 patients in the second group were assessed by the scale and then input directly into the computer so that the computer could output the diagnostic results; the second group was given to experts for diagnosis at the same time, and the results obtained by experts were compared with those obtained by experts. The diagnostic results of the second group of computer outputs were compared. Results We found that there was no statistical difference between the results of computer diagnosis and expert diagnosis in the second group. Conclusion An intelligent diagnosis model of post-stroke dysphagia based on RBF neural network is constructed, and expert diagnosis experience can be learned. Finally, compared with several typical machine learning methods,the accuracy and advantages of the established RBF neural network model is verified.
引文
[1]汪文婧,孙慧男,陈旭昕,等.吞咽困难的发病原因及治疗研究进展[J].转化医学杂志,2015,4(6):377-381.
[2]大西幸子,孙启良.脑卒中患者摄食-吞咽障碍的评价与训练[J].中国康复医学杂志,1997,12(3):141.
[3]张蜻,王拥军.卒中后吞咽困难的发生机制[J].国外医学脑血管疾病分册,2004,12(4):274-277.
[4]Martino R,Foley N,Bhogal S,et al.Dysphagia after stroke:incidence,diagnosis,and pulmonary complications[J].Stroke,2005,36:2756-2763.
[5]Sung KHA,Ki EK,Hye SL,et al.Interventions for dysphagia following stroke:A systematic review of randomized controlled trials[J].J Korean Dysphagia Society,2019,9(1):26-35
[6]饶明俐.中国脑血管病防治指南[M].北京:人民卫生出版社,2007.
[7]Smithard DG,O'Neill PA,Park C,et al.Complications and Outcome After Acute Stroke Does Dysphagia Matter?[J]Stroke,1996,27(7):1200-1203.
[8]Smithard D G,O'Neill PA,Eng land RE,et al.Thenatural history of dysphagia following astroke[J].Dysphagia,1997,12(4):188-193.
[9]Smithard DG,O'Neill PA,Martin DF,et al.Aspiration following stroke:is it related to the side of the stroke?[J]Clin Rehabil,1997,11(1):73-76.
[10]Hind NP,Wiles CM.Assessment of swallowing and referral to speech and language the rapistin acutestroke[J].QJM,1998,91:829-835.
[11]Yanzhi W,Guixiong L.A Forecasting method based on online self-correcting single model RBF neural network[J].Procedia Eng,2012,29:2516-2520.
[12]Honggui H,Qili C,Junfei Q.An efficient self-organizing RBF neural network for water quality prediction[J].Neural Net,2011,24:717-725.
[13]Passos KO,Cardoso MC,Scheeren B.Association between functionality assessment scales and the severity of dysphagia post-stroke[J].Codas,2017,29(1):239-246.
[14]Reed,Wolf,Cotton D.A visual analogue scale and a Likert上接第23页scale are simple and responsive tools for assessing dysphagia in eosinophili coesophagitis[J].Aliment Pharmacol Ther,2017,45(11):1443-1448.
[15]窦祖林,兰月,于帆,等.吞咽造影数字化分析在脑干卒中后吞咽障碍患者疗效评估中的应用[J].中国康复医学杂志,2013,28(9):799-805.
[16]王利贤.Wilson病构音障碍和吞咽障碍的临床综合评定及相关因素研究[D].广州:广州药科大学,2017.
[17]Teodorescu HL,Chelaru M,Kandel A,et al.Fuzzy methods in tremor assessment,prediction,and rehabilitation[J].Artif Intell Med,2001,21:107-130.
[18]Ahmet Y.Soft computing in medicine[J].Applied Soft Computing,2009,9:1029-1043.
[19]Jamwal PK,Xie SQ,Tsoi YH,et al.Forward kinematics modelling of a parallel ankle rehabilitation robot using modified fuzzy inference[J].Mech Mach Theory,2010,45:1537-1554.
[20]Shamsuddin S,Yussof H,Ismail LI,et al.Humanoid Robot NAOinteracting with autistic children of moderately impaired intelligence to augment communication skills[J].Procedia Eng,2012,41:1533-1538.
[21]Hamet P,Tremblay J.Artificial intelligence in medicine[J].Metabolism,2017,69:36-40.C
[2]大西幸子,孙启良.脑卒中患者摄食-吞咽障碍的评价与训练[J].中国康复医学杂志,1997,12(3):141.
[3]张蜻,王拥军.卒中后吞咽困难的发生机制[J].国外医学脑血管疾病分册,2004,12(4):274-277.
[4]Martino R,Foley N,Bhogal S,et al.Dysphagia after stroke:incidence,diagnosis,and pulmonary complications[J].Stroke,2005,36:2756-2763.
[5]Sung KHA,Ki EK,Hye SL,et al.Interventions for dysphagia following stroke:A systematic review of randomized controlled trials[J].J Korean Dysphagia Society,2019,9(1):26-35
[6]饶明俐.中国脑血管病防治指南[M].北京:人民卫生出版社,2007.
[7]Smithard DG,O'Neill PA,Park C,et al.Complications and Outcome After Acute Stroke Does Dysphagia Matter?[J]Stroke,1996,27(7):1200-1203.
[8]Smithard D G,O'Neill PA,Eng land RE,et al.Thenatural history of dysphagia following astroke[J].Dysphagia,1997,12(4):188-193.
[9]Smithard DG,O'Neill PA,Martin DF,et al.Aspiration following stroke:is it related to the side of the stroke?[J]Clin Rehabil,1997,11(1):73-76.
[10]Hind NP,Wiles CM.Assessment of swallowing and referral to speech and language the rapistin acutestroke[J].QJM,1998,91:829-835.
[11]Yanzhi W,Guixiong L.A Forecasting method based on online self-correcting single model RBF neural network[J].Procedia Eng,2012,29:2516-2520.
[12]Honggui H,Qili C,Junfei Q.An efficient self-organizing RBF neural network for water quality prediction[J].Neural Net,2011,24:717-725.
[13]Passos KO,Cardoso MC,Scheeren B.Association between functionality assessment scales and the severity of dysphagia post-stroke[J].Codas,2017,29(1):239-246.
[14]Reed,Wolf,Cotton D.A visual analogue scale and a Likert上接第23页scale are simple and responsive tools for assessing dysphagia in eosinophili coesophagitis[J].Aliment Pharmacol Ther,2017,45(11):1443-1448.
[15]窦祖林,兰月,于帆,等.吞咽造影数字化分析在脑干卒中后吞咽障碍患者疗效评估中的应用[J].中国康复医学杂志,2013,28(9):799-805.
[16]王利贤.Wilson病构音障碍和吞咽障碍的临床综合评定及相关因素研究[D].广州:广州药科大学,2017.
[17]Teodorescu HL,Chelaru M,Kandel A,et al.Fuzzy methods in tremor assessment,prediction,and rehabilitation[J].Artif Intell Med,2001,21:107-130.
[18]Ahmet Y.Soft computing in medicine[J].Applied Soft Computing,2009,9:1029-1043.
[19]Jamwal PK,Xie SQ,Tsoi YH,et al.Forward kinematics modelling of a parallel ankle rehabilitation robot using modified fuzzy inference[J].Mech Mach Theory,2010,45:1537-1554.
[20]Shamsuddin S,Yussof H,Ismail LI,et al.Humanoid Robot NAOinteracting with autistic children of moderately impaired intelligence to augment communication skills[J].Procedia Eng,2012,41:1533-1538.
[21]Hamet P,Tremblay J.Artificial intelligence in medicine[J].Metabolism,2017,69:36-40.C