Versions

VEEL 0.04

• A fit assuming ascending and descending data are now both generated and the best one (RMSE criterion) is chosen.

VEEL 0.02

• Pandas tables are now used to read in data. This allows the grouping variables to be words as well as numbers. It also prevents easy py2exe 'compiling' on Windows 7 (64 bit).
• Data are no longer assumed to be errors. This means that previously, the data were assumed to approach zero over time, but now, it can also deviate from zero over time. This is useful if the data are some measure of performance. The direction of the curve is now estimated using a pearson r correlation (from scipy).
• Algorithm is now robust against missing data.

VEEL 0.01

Initial version.