Learn Predictive Representations of Time-Varying Data

Documentation for package ‘timesift’ version 0.3.1

Help Pages

artifacts	Read and write the artifacts that cross the language boundary
as_sift	Several representations to run the same learners across
average_precision	Average precision
bind_channels	Put channels side by side
build_representation	Build one representation for a set of targets
c.timesift_learner	Combine learners, or representations, into a set
c.timesift_models	Combine learners, or representations, into a set
c.timesift_representation	Combine learners, or representations, into a set
c.timesift_sift	Combine learners, or representations, into a set
calendar_channels	Where in the year, or the day, each bin sits
cnn	Sequence encoders with a joint multi-label head
combine	Combine learners, or representations, into a set
coverage	Which units reach which bins
cv	How the folds are drawn
decision_threshold	Cohen's kappa, and where two models disagree
digest_array	The cross-language digest of a representation
elasticnet	Penalised regression on the flattened representation
ensemble	How the candidates are combined
ensemble_combine	Combine one prediction per member into one prediction
ensemble_fit	Fit the combiner on the out-of-fold predictions
ensemble_weights	The weights the combiner fitted
feature_matrix	Bring an already-reduced feature table into a ladder
fit_learner	Fit one learner at one grain
fold_map	Assign units to cross-validation folds
forest	Random forest on the flattened representation
grain	How a series becomes an array a learner reads
grains	Several representations to run the same learners across
grain_contrasts	Compare every grain against a learner's best one
grain_ladder	Fit at every grain and see where skill saturates
grain_matrix	Reduce sensor series to a temporal grain
grouped_cv	How the folds are drawn
implied_skill	What population skill a reported level is consistent with
kappa_score	Cohen's kappa, and where two models disagree
learner	Define a learner
learners	Register a learner
lookback	How a series becomes an array a learner reads
lookbacks	Several representations to run the same learners across
lookback_matrix	Reduce sensor series to a lookback anchored on each target
metrics	Register a metric
mlp	Sequence encoders with a joint multi-label head
model_agreement	Cohen's kappa, and where two models disagree
multigrain	How a series becomes an array a learner reads
native	How a series becomes an array a learner reads
occlusion	What part of the record a fitted model reads
occlusion.default	What part of the record a fitted model reads
occlusion.timesift	What part of the record a fitted model reads
occlusion.timesift_ladder	What part of the record a fitted model reads
paired_contrast	Compare two arms cell by cell
plot.timesift	Draw a run
plot.timesift_ladder	Draw a ladder
plot.timesift_selection	Draw how stable the choice of grain was
positive_weights	Case weights that balance a rare response
predict.timesift	Predict from a fitted timesift
predict.timesift_fit	Fit one learner at one grain
print.timesift	What a run found
print.timesift_summary	What a run found
read_cells	Read and write the artifacts that cross the language boundary
read_folds	Read and write the artifacts that cross the language boundary
read_response	Read and write the artifacts that cross the language boundary
register_learner	Register a learner
register_metric	Register a metric
register_response	Register a response head
rescnn	Sequence encoders with a joint multi-label head
responses	Register a response head
roc_auc	The area under the ROC curve
scorable_cells	Which cells a score is defined on
score_predictions	Score held-out predictions on the cells the mask allows
select_grain	Choose the grain inside the training data, and score the whole procedure
simulate_records	Simulate sensor records whose response acts at a known temporal grain
stepwise	Forward selection by AIC on the flattened representation
summary.timesift	What a run found
summary.timesift_ladder	Fit at every grain and see where skill saturates
summary.timesift_selection	Choose the grain inside the training data, and score the whole procedure
timesift	Fit and compare representations of time-varying data
timesift_report	What a run found
timesift_set	Several built representations of the same targets
torch_learners	Sequence encoders with a joint multi-label head
train_control	Training settings every neural learner reads
tss	The true skill statistic
tss_inflation	How much a self-selected threshold inflates the reported level
write_cells	Read and write the artifacts that cross the language boundary
write_folds	Read and write the artifacts that cross the language boundary
write_response	Read and write the artifacts that cross the language boundary