Skip to main content
Fig. 10 | BMC Medical Research Methodology

Fig. 10

From: Treatment of missing data in Bayesian network structure learning: an application to linked biomedical and social survey data

Fig. 10

Distribution of missing values in the real-world data set. (A) Proportion of missing values in each variable (named as in Supplementary Table 1 of Additional file 1), shown as a bar chart. (B) Missing patterns, shown as a heatmap with proportions to the right of the plot. Rows represent a single missing pattern (‘Combinations’) and columns variables, with the variable missing in a given pattern coloured green (blue otherwise). The proportion of each missing pattern is shown as a horizontal bar chart to the right of the heatmap (summing to 0.53 for missing patterns). The very bottom row represents the pattern with no missing values, with its proportion bar in blue with value 0.47

Back to article page