Support Vector Feature Selection for Early Detection of Anastomosis Leakage From Bag-of-Words in Electronic Health Records.
Cristina Soguero-RuízKristian HindbergJosé Luis Rojo-ÁlvarezStein Olav SkrøvsethFred GodtliebsenKim MortensenArthur RevhaugRolv-Ole LindsetmoKnut Magne AugestadRobert JenssenPublished in: IEEE J. Biomed. Health Informatics (2016)
Keyphrases
- bag of words
- early detection
- electronic health records
- feature selection
- support vector
- text classification
- image classification
- clinical data
- visual words
- breast cancer
- text categorization
- classification accuracy
- logistic regression
- action recognition
- clinical trials
- support vector machine
- lung cancer
- health care
- medical records
- image representation
- loss function
- n gram
- model selection
- machine learning
- feature extraction
- feature ranking
- medical data
- kernel function
- feature space
- databases
- dimensionality reduction
- feature subset
- naive bayes
- data sets
- feature set
- similarity measure
- statistical analysis
- knn
- medical images