TLBC: Two-Level Behavior Classification

Contains functions for training and applying two-level random forest and hidden Markov models for human behavior classification from raw tri-axial accelerometer and/or GPS data. Includes functions for training a two-level model, applying the model to data, and computing performance.

AuthorKatherine Ellis
Date of publication2015-10-14 18:13:22
MaintainerKatherine Ellis <kellis@ucsd.edu>
LicenseGPL-2
Version1.0

View on CRAN

Man pages

alignStart: Function to align start of a window

annotationsToLabels: Function to convert bout-level annotations to instance-level...

calcPerformance: Function to calculate performance of a classification model

classify: Function to classify accelerometer and/or GPS data

clearFiles: Clear files

computeEmissionProbs: Compute emission probabilities

computeOneAccFeat: Compute one acceleration feature

computeOneGPSFeat: Compute one GPS feature

computePriorProbs: Compute prior probabilities

computeTransProbs: Compute transition probabilities

distance: Distance

extractAccelerometerFeatures: Extract accelerometer features

extractAccFeatsFile: Extract accelerometer features from a file

extractFeatsPALMSDir: Extract GPS features from a PALMS directory

extractFeatsPALMSOneFile: Extract GPS features from a PALMS file

extractLabelsDir: Extract labels from a directory

extractLabelsSingleFile: Extract labels from a directory

getDateFmt: Get date format

hmm: Hidden Markov model

isFeatureDirectory: Is feature directory?

isInstanceFormat: Is instance format?

loadData: Load data

loadFeatures: Load features

loadLabels: Load labels

loadModel: Load model

loadPredictions: Load predictions

loadPredictionsAndLabels: Load predictions and labels

looXval: Function to perform leave-one-out cross-validation

rf: Random Forest

senseCamLabels: SenseCam Labels

sensorsToFeatures: Function to extract featurese from raw sensor data

stratSample: Stratified sample

testHMM: Test a hidden Markov model

testRF: Test a random forest classifier

testTwoRFs: Test two random forest classifiers

TLBC-package: Two-Level Behavior Classification

trainHMM: Train a hidden Markov model

trainModel: Function to train a two-level model from accelerometer and/or...

trainRF: Train a random forest classifier

winSize: Window Size

writePredictions: Write predictions to a file

Questions? Problems? Suggestions? or email at ian@mutexlabs.com.

All documentation is copyright its authors; we didn't write any of that.