Hi Statalisters,
I want to compare the distributional fit of predicted values with the original data. The user written Stata command 'ppplot' by Nicholas J. Cox plots the percentiles of the predictions with the percentiles of the original data. A diagonal line is a perfect distributional fit. In 'Hinloopen, Jeroen; van Marrewijk, Charles (2005) : Comparing Distributions: The Harmonic Mass Index, Tinbergen Institute Discussion Paper, No. 05-122/1' the authors introduce the harmonic mass index which measures the area between the diagonal (e.g. the perfect fit) and the P-P-plot. Besides the graph I want to incorporate this measure in my study, but I am fairly new to Stata and haven't figured out a way how to calculate the HMI.
How can I calculate this measure and are there others measures helping to compare the distributional fit of predictions and original data (besides the ppplot and ksmirnov)?
Kind regards,
Steffen
Related Posts with Measures to compare distributions and the Probability-Probability-Plot
Storing Beta values by running regression loops on panel data- the smae code and the same data, but the result is different using STATA 13,15 and 16
- Identification of gender from names through machine learning methods
- Line graph for multiple time points
- bootstrap: insufficient observations to compute bootstrap standard errors
Subscribe to:
Post Comments (Atom)
0 Response to Measures to compare distributions and the Probability-Probability-Plot
Post a Comment