Elicitation of neurological knowledge with argument-based machine learning

OBJECTIVE The paper describes the use of expert's knowledge in practice and the efficiency of a recently developed technique called argument-based machine learning (ABML) in the knowledge elicitation process. We are developing a neurological decision support system to help the neurologists differentiate between three types of tremors: Parkinsonian, essential, and mixed tremor (comorbidity). The system is intended to act as a second opinion for the neurologists, and most importantly to help them reduce the number of patients in the "gray area" that require a very costly further examination (DaTSCAN). We strive to elicit comprehensible and medically meaningful knowledge in such a way that it does not come at the cost of diagnostic accuracy. MATERIALS AND METHODS To alleviate the difficult problem of knowledge elicitation from data and domain experts, we used ABML. ABML guides the expert to explain critical special cases which cannot be handled automatically by machine learning. This very efficiently reduces the expert's workload, and combines expert's knowledge with learning data. 122 patients were enrolled into the study. RESULTS The classification accuracy of the final model was 91%. Equally important, the initial and the final models were also evaluated for their comprehensibility by the neurologists. All 13 rules of the final model were deemed as appropriate to be able to support its decisions with good explanations. CONCLUSION The paper demonstrates ABML's advantage in combining machine learning and expert knowledge. The accuracy of the system is very high with respect to the current state-of-the-art in clinical practice, and the system's knowledge base is assessed to be very consistent from a medical point of view. This opens up the possibility to use the system also as a teaching tool.

[1]  Saso Dzeroski,et al.  Weakening the language bias in LINUS , 1994, J. Exp. Theor. Artif. Intell..

[2]  K. S. Nijran,et al.  Automatic classification of 123I-FP-CIT (DaTSCAN) SPECT images , 2011, Nuclear medicine communications.

[3]  G. Deuschl,et al.  Essential tremor and cerebellar dysfunction clinical and kinematic analysis of intention tremor. , 2000, Brain : a journal of neurology.

[4]  Ivan Bratko,et al.  Argument based machine learning , 2006, Artif. Intell..

[5]  Saso Dzeroski,et al.  Learning Nonrecursive Definitions of Relations with LINUS , 1991, EWSL.

[6]  Saso Dzeroski,et al.  The utility of background knowledge in learning medical diagnostic rules , 1993, Appl. Artif. Intell..

[7]  Bojan Cestnik,et al.  Estimating Probabilities: A Crucial Task in Machine Learning , 1990, ECAI.

[8]  Heng-Li Yang.  Information/knowledge acquisition methods for decision support systems and expert systems , 1995 .

[9]  G Edwards,et al.  Peirs: A pathologist‐maintained expert system for the interpretation of chemical pathology reports , 1993, Pathology.

[10]  Giuliana Grimaldi,et al.  Tremor: From Pathogenesis to Treatment , 2008, Synthesis Lectures on Biomedical Engineering.

[11]  Herbert A. Simon,et al.  Applications of machine learning and rule induction , 1995, CACM.

[12]  Tom Robinson,et al.  Essential tremor-the most common movement disorder in older people. , 2006, Age and ageing.

[13]  John H. Boose,et al.  A survey of knowledge acquisition techniques and tools , 1993 .

[14]  Nancy J. Cooke,et al.  Varieties of knowledge elicitation techniques , 1994, Int. J. Hum. Comput. Stud..

[15]  Ashwin Srinivasan,et al.  Ripple down rules: Turning knowledge acquisition into knowledge maintenance , 1992, Artif. Intell. Medicine.

[16]  Ivan Bratko,et al.  Elicitation of Neurological Knowledge with ABML , 2011, AIME.

[17]  Peter Clark,et al.  Rule Induction with CN2: Some Recent Improvements , 1991, EWSL.

[18]  Roy Rada,et al.  Machine learning - applications in expert systems and information retrieval , 1986, Ellis Horwood series in artificial intelligence.

[19]  Alcimar B. Soares,et al.  MEASUREMENT OF TREMOR USING DIGITIZING TABLETS , 2007 .

[20]  J. Hughes,et al.  Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico-pathological study of 100 cases. , 1992, Journal of neurology, neurosurgery, and psychiatry.

[21]  Randolph A. Miller,et al.  Towards computer-assisted maintenance of medical knowledge bases , 1990, Artif. Intell. Medicine.

[22]  K. Lyons,et al.  Essential tremor: differential diagnosis and current therapy. , 2003, The American journal of medicine.

[23]  Charles C. Osuagwu,et al.  The Underlying Issues in Knowledge Elicitation , 2006 .

[24]  Ivan Bratko,et al.  Argument based machine learning , 2006, Artif. Intell..

[25]  Geoffrey I. Webb,et al.  An Experimental Evaluation of Integrating Machine Learning with Knowledge Acquisition , 1999, Machine Learning.

[26]  P. Kraus,et al.  Spiralometry: Computerized assessment of tremor amplitude on the basis of spiral drawing , 2010, Movement disorders : official journal of the Movement Disorder Society.

[27]  Edward A. Feigenbaum,et al.  Some challenges and grand challenges for computational intelligence , 2003, JACM.

[28]  Susanne A Schneider,et al.  Tremor—some controversial aspects , 2011, Movement disorders : official journal of the Movement Disorder Society.

[29]  Byeong Ho Kang,et al.  Multiple Classification Ripple Down Rules : Evaluation and Possibilities , 2000 .

[30]  Ian M. Neale,et al.  First generation expert systems: a review of knowledge acquisition methodologies , 1988, The Knowledge Engineering Review.

[31]  Edward A. Felgenbaum.  The art of artificial intelligence: themes and case studies of knowledge engineering , 1977, IJCAI 1977.

[32]  Ivan Bratko,et al.  Fighting Knowledge Acquisition Bottleneck with Argument Based Machine Learning , 2008, ECAI.

[33]  Andrew W. H. Ip,et al.  Knowledge elicitation approach in enhancing tacit knowledge sharing , 2011, Ind. Manag. Data Syst..