close

Вход

Забыли?

вход по аккаунту

код для вставкиСкачать
terminology
• species data = the measured variables we
want to explain (response or dependent
variables)
• environmental data = the variables we use for
explaining the species data (explanatory or
independent variables)
• sites are rows
• columns are variables
data transformation
• Usually applied to response variables only
– Make the measurement scales similar or the same
– Stabilize variance (log / sqrt)
– recode into binary variables
association measures
• Q mode
– distances, e.g., Euclidean
• R mode
– correlations, e.g., Pearson's correlation
distances
• Q mode, quantitative
– Bray-Curtis
• gives the same weight to same differences in
abundances irrespective of the scales
• 1 ... 2 -> 1 and 1001 ... 1002 -> 1
• Usually at least square root transformation before
distances are calculated
– Hellinger
• abundances are divided by the site's total abundance
and then square root transformed
distances
• Q mode, presence / absence
– Jaccard
1/--страниц
Пожаловаться на содержимое документа