The core purpose of the dendroTools package is to introduce novel
dendroclimatological methods to study linear and nonlinear relationships
between environmental (climate) data and tree-ring sequences in R. The
two functions, which are used to study climate-growth relationships
using daily climate data, are daily_response() and
daily_response_seascorr(), while the two functions for monthly data
are monthly_response() and monthly_response_seascorr(). The seascorr
functions are designed to study partial correlation coefficients, while
the other two can be used to calculate simple correlations, or fit
linear and nonlinear models between aggregated climate data and
tree-ring proxies. To use daily_response(), two data frames are
required, one with daily climate data, e.g. mean daily temperature; and
one with tree-ring proxy records. Example data is provided, so users can
see, how data frames should be organized. The daily_response()
calculates all possible values of a selected statistical metric between
response variable(s) and daily environmental data. Calculations are
based on a moving window, which runs through daily environmental data
and calculates moving averages. In addition, there generic functions are
available for visualization and interpretation of results.
A completely separate function is the compare_methods() which can be
used to compare different regression methods for climate reconstruction.
It calculates several performance metrics for train and test data for
different regression methods: multiple linear regression (MLR),
artificial neural networks with Bayesian regularization training
algorithm (BRNN), M5P model trees (MT), model trees with bagging (BMT)
and random forest of regression trees (RF). Calculated performance
metrics are correlation coefficient (r), root mean squared error (RMSE),
root relative squared error (RRSE), index of agreement (d), reduction of
error (RE), coefficient of efficiency (CE) and mean bias.
You can install dendroTools using:
library("devtools")
devtools::install_github("jernejjevsenak/dendroTools") # current version under development
install.packages("dendroTools") # from CRAN
- Jernej Jevšenak