Letter to the Editor
Imputation is beneficial for handling missing data in predictive models

https://doi.org/10.1016/j.jclinepi.2007.03.003Get rights and content

References (4)

There are more references available in the full text version of this article.

Cited by (79)

  • Diffusion basis spectrum imaging predicts long-term clinical outcomes following surgery in cervical spondylotic myelopathy

    2023, Spine Journal
    Citation Excerpt :

    Nonoverlapping 95% confidence intervals between feature sets for a given performance metric (eg, accuracy) were considered statistically significantly different [33]. Missing data from patient-reported outcome questionnaires (<0.5%) were statistically imputed via single regression imputation [34] based on predictive mean matching (mice package R) [34–37]. All statistical analyses were performed in Python and R, version 4.1.4.

  • Risk factor analysis for “diagnosis not reached” results from bovine samples submitted to British veterinary laboratories in 2013–2017

    2020, Preventive Veterinary Medicine
    Citation Excerpt :

    The main concern in the present study was related to the loss of precision and power due to the exclusion of those records with unknown values. But considering the high number of submissions included, it may not have substantially impacted upon statistical power (Allison, 2000; Sterne et al., 2009; Steyerberg and van Veen, 2007). It was not possible to explore any biologically plausible interaction due to the characteristics of the dataset.

  • Investigating Risk Factors and Predicting Complications in Deep Brain Stimulation Surgery with Machine Learning Algorithms

    2020, World Neurosurgery
    Citation Excerpt :

    Given the scarce nature of DBS case data and the resources required to collect it, the research team was motivated to retain as many cases as possible for analysis. Data imputation addresses this issue and various methods can be used.61-65 Four neural network regression models were developed to impute BMI for the cases with missing data.

View all citing articles on Scopus
View full text