Principal Component Analysis Enhanced with Bootstrapped Confidence Interval for the Classification of Parkinsonian Patients Using Gaussian Mixture Model and Gait Initiation Parameters.
Florent LoeteArnaud SimonetPaul FourcadeEric YiouArnaud DelafontainePublished in: Sensors (2024)
Keyphrases
- gaussian mixture model
- feature vectors
- expectation maximization
- principal component analysis
- feature space
- maximum likelihood
- confidence intervals
- gaussian mixture
- density estimation
- mixture model
- em algorithm
- feature extraction
- classification accuracy
- pattern recognition
- probability density function
- mixture distribution
- speaker recognition
- dimensionality reduction
- principal components
- speaker identification
- decision trees
- covariance matrix
- class labels
- maximum likelihood criterion
- text classification
- feature set
- face recognition
- image processing
- feature selection
- machine learning
- unsupervised learning
- generative model
- model selection
- probabilistic model
- pairwise
- support vector
- similarity measure