sorry, we can't preview this file
plos_one_data.csv.gz (1.29 MB)
plos_one_data.csv.gz
This dataset is a subset of the Audiological and Genetic Database (AudGenDB), a public, de-identified research database derived from electronic health records (EHR). AudGenDB is intended to be a data resource that represents a population enriched for hearing loss research. The complete AudGenDB can be accessed at http://audgendb.chop.edu
This subset is obtained by applying the following exclusion criteria: (1) Only audiogram tests documenting the specific ear tested (‘left’, ‘right’, or ‘soundfield’) are included; (2) Only
bone-conduction tests are included; (3) Tests outside the frequency range of interest (500Hz, 1KHz, 2KHz, and 4KHz) were excluded; (4) Response thresholds recorded as 0 decibels were removed as they typically indicate a failed test attempt; (5) Patients with fewer than 4 tests across all frequencies of interest were also excluded in order to enable
a 4-fold cross-validation procedure (described in more detail below); (6) Finally, test results recorded outside the age range of 0 to 21 were also removed in order to focus on a pediatric population.
This dataset is uploaded for PLOS One publication Latent Patient Cluster Discovery for Robust Future
Forecasting and New-patient Generalization
This subset is obtained by applying the following exclusion criteria: (1) Only audiogram tests documenting the specific ear tested (‘left’, ‘right’, or ‘soundfield’) are included; (2) Only
bone-conduction tests are included; (3) Tests outside the frequency range of interest (500Hz, 1KHz, 2KHz, and 4KHz) were excluded; (4) Response thresholds recorded as 0 decibels were removed as they typically indicate a failed test attempt; (5) Patients with fewer than 4 tests across all frequencies of interest were also excluded in order to enable
a 4-fold cross-validation procedure (described in more detail below); (6) Finally, test results recorded outside the age range of 0 to 21 were also removed in order to focus on a pediatric population.
This dataset is uploaded for PLOS One publication Latent Patient Cluster Discovery for Robust Future
Forecasting and New-patient Generalization
