User: | Open Learning Faculty Member:
Of the three sampling techniques (systematic, random, and haphazard) that I used for the virtual forest tutorial, the technique with the lowest average error rate was the systematic sampling.
The fastest technique was the haphazard sampling (12 hours 34 minutes), but the difference between the fastest and the slowest was only 13 minutes (1.7% of 12 hours 34 minutes), which is fairly negligible.
Common vs. rare species
The average error rate for the two most common species, Eastern Hemlock and Sweet Birch, was 18.1%. The average error rate for the two least common species, Striped Maple and White Pine, was 42.9%. From this dataset, it appears that the accuracy did decrease with species rarity.
Comparing sampling techniques
Systematic sampling had the lowest average error rate, at 16.7%. Random sampling had the highest average error rate, at 46.2%.
Based on this dataset, systematic sampling appears to be the most efficient and accurate. With systematic sampling, I found the lowest error rates (on average) with only a two-minute time penalty over the fastest technique. The 16% error rate average still seems high, to me, so I would want to re-do this exercise multiple times, probably with more samples, to be able to better identify the technique most efficient in this setting and its most efficient number of samples.
