Reusing training data with generative/discriminative hybrid model for practical acceleration-based activity recognition.
Quan KongTakuya MaekawaPublished in: Computing (2014)
Keyphrases
- activity recognition
- hybrid model
- discriminative methods
- training data
- human activity recognition
- action recognition
- generative model
- human activities
- artificial neural networks
- discriminative learning
- hybrid models
- semi supervised
- support vector machine svm
- visual surveillance
- support vector machine
- smart home
- daily activities
- accelerometer data
- supervised learning
- unsupervised learning
- learning algorithm
- training set
- support vector regression
- data sets
- human actions
- prior knowledge
- smart environments
- feature selection
- classification accuracy
- activity analysis
- wearable sensors
- data mining
- similarity measure
- feature extraction
- decision trees
- information retrieval
- transfer learning