Function to adjust any negative predicted features

Sometimes PC model can generate negative values for some features. E.g. South Africa where GINI is much too high. Here, the function adjusts negative predicted features.

adjust_negative_predicted_features(
  df,
  min_lowest_feature_val = 0.006,
  grouping_variables = c("country", "year")
)

Arguments

df	dataframe with decile data
min_lowest_feature_val	Lowest possible value for lowest feature (e.g. d1).
grouping_variables	variables to group by