Discovering pattern in medical audiology data with FP-growth algorithm

There is potential knowledge inherent in vast amounts of untapped and possibly valuable data generated by healthcare providers. So often, clinicians rely in their skills and experience and that of other medical experts as their source of information. The healthcare sector is now capturing more data in the form of digital and non digital format that may potentially be mined to generate valuable insights. In this paper we propose a five step knowledge discovery model to discover patterns in medical audiology records. We use frequent pattern growth (FP-Growth) algorithm in the data processing step to build the FP-tree data structure and mine it for frequents itemsets. Our aim is to discover interesting itemsets that shows connection between hearing thresholds in pure-tone audiometric data and symptoms from diagnosis and other attributes in the medical records. The experimental results are summaries of frequent structures in the data that contains symptoms of tinnitus, vertigo and giddiness with threshold values and other information like gender.

[1]  Roberto J. Bayardo,et al.  Efficiently mining long patterns from databases , 1998, SIGMOD '98.

[2]  Susan Griest,et al.  A 5-Year Prospective Study of Diabetes and Hearing Loss in a Veteran Population , 2006, Otology & neurotology : official publication of the American Otological Society, American Neurotology Society [and] European Academy of Otology and Neurotology.

[3]  He Yun.  Medical Data Mining Based on Association Rules , 2009 .

[4]  M. Davarpanah,et al.  Classifying ear disorders using support vector machines , 2010, 2010 Second International Conference on Computational Intelligence and Natural Computing.

[5]  Gregory Piatetsky-Shapiro,et al.  The KDD process for extracting useful knowledge from volumes of data , 1996, CACM.

[6]  Peter Harrington,et al.  Machine Learning in Action , 2012 .

[7]  Edward Y. Chang,et al.  Pfp: parallel fp-growth for query recommendation , 2008, RecSys '08.

[8]  Zengyou He,et al.  An FP-Tree Based Approach for Mining All Strongly Correlated Item Pairs , 2005, CIS.

[9]  H. Kremer,et al.  Mild and Variable Audiometric and Vestibular Features in a Third DFNA15 Family with a Novel Mutation in POU4F3 , 2009, The Annals of otology, rhinology, and laryngology.

[10]  Muhammad N Anwar,et al.  Data mining of audiology patient records: factors influencing the choice of hearing aid type , 2011, DTMBIO '11.

[11]  Rolf Stadler,et al.  Discovering Data Mining: From Concept to Implementation , 1997 .

[12]  Jian Pei,et al.  Mining Frequent Patterns without Candidate Generation: A Frequent-Pattern Tree Approach , 2006, Sixth IEEE International Conference on Data Mining - Workshops (ICDMW'06).

[13]  Juen-Haur Hwang,et al.  Using cluster analysis to classify audiogram shapes , 2010, International journal of audiology.

[14]  J.H. Hwang,et al.  Diagnostic Value of Combining Bilateral Electrocochleography Results for Unilateral Ménière’s Disease , 2008, Audiology and Neurotology.

[15]  M. Madheswaran,et al.  Hybrid Medical Image Classification Using Association Rule Mining with Decision Tree Algorithm , 2010, ArXiv.

[16]  John K Niparko,et al.  Risk Factors for Hearing Loss in US Adults: Data From the National Health and Nutrition Examination Survey, 1999 to 2002 , 2009, Otology & neurotology : official publication of the American Otological Society, American Neurotology Society [and] European Academy of Otology and Neurotology.

[17]  Rashmi.  Data Mining: A Knowledge Discovery Approach , 2012 .