Jonatan Baggman

I’m doing my master thesis here at the ISB group to finish my MSc in Engineering Biology. For my master thesis, I’m looking into imputations for medical datasets and will then use them for machine learning. The machine learning models will hopefully be used in combination with mechanistic systems biology models to predict the risk of different diseases

Common for medical datasets are that they are missing data points. The missing data lowers the quality of the data and a common way to handle them, listwise deletion, removes all observations with data points missing. There are two main problems with this approach: the amount of data is greatly reduced and bias is introduced to the data. Imputations are a way to estimate these missing datapoints to reduce these problems. I will continue with and further develop an imputations framework produced by Tabea Kossen and apply it to the data available to us to compare different imputation methods and determine the most appropriate one to use.