Labeling method for acceleration data using an execution sequence of activities.
Kazuya MuraoTsutomu TeradaPublished in: UbiComp (Adjunct Publication) (2013)
Keyphrases
- synthetic data
- input data
- data sets
- prior knowledge
- noisy data
- test data
- missing data
- training samples
- cost function
- statistical methods
- database
- data collection
- training data
- missing values
- image data
- raw data
- data processing
- high accuracy
- data quality
- preprocessing
- objective function
- correlation analysis
- decision trees
- human activities
- dimensional vector
- information loss
- statistical analysis
- model selection
- learning algorithm
- knowledge discovery
- dynamic programming
- active learning
- pairwise
- data analysis
- high quality
- similarity measure