Application of Data Mining to a Large Hearing-Aid Manufacturer's Dataset to Identify Possible Benefits for Clinicians, Manufacturers, and Users

Joseph Mellor, Michael Stone, John Keane

Research output: Contribution to journalArticlepeer-review

Abstract

Modern hearing instruments contain logging technology to record data, such as the acoustic environments in which the device is being used and how the signal processing is consequently operating. Combined with patient data, such as the audiogram, this information gives a more comprehensive picture of the user and their relationship with the aid. Here, a relatively large, anonymized dataset ( 4300,000 devices, 4150,000 wearers) from a hearing-aid manufacturer was data mined for connections between subsets of the logged varieties of data. Apart from replicating links that have previously been reported in labor-intensive studies, a link between device style (in-the-ear/behind-the-ear) and the sound levels of encountered environments was demonstrated, suggesting that some device types are more successful from a lifestyle perspective. Furthermore, the data also suggested links between the audiogram and the sound environments in which the aid was operated. Modeling the expected link between the environment and the microphone directionality settings revealed patterns of either abnormal fitting or where the aid was not operating correctly—factors that may indicate a failed fitting.
Given the necessarily redacted nature of the dataset, the reported findings represent a proof-of-concept of the use of relatively large-scale data mining to guide and assess hearing-aid fitting procedures for possible benefits to the clinician, manufacturer, and patient.
Original languageEnglish
Number of pages16
JournalTrends in Hearing (Online)
Volume22
Early online date31 May 2018
DOIs
Publication statusPublished - 31 May 2018

Keywords

  • audiogram
  • auditory ecology
  • big data
  • candidature
  • hearing aids

Fingerprint

Dive into the research topics of 'Application of Data Mining to a Large Hearing-Aid Manufacturer's Dataset to Identify Possible Benefits for Clinicians, Manufacturers, and Users'. Together they form a unique fingerprint.

Cite this