Download
1 / 3
30 likes | 146 Vues
This resource compiles a variety of machine learning techniques applicable to classification, concept formation, variable selection, and structure discovery within the biomedical domain. These methods are critical for biological discovery, text categorization, diagnosis, prognosis, and treatment selection. As data proliferation in classical and molecular medicine continues, the importance of these machine learning methods and their evolving derivatives will increase, significantly impacting the practice and advancement of modern medicine.
E N D
Resources • ML resource web pages • ML libraries and systems • Data repositories • Books • Journals • Conferences • Papers (often cited and often downloaded) • MC-SVM system & Causal Explorer library • Data + papers for case studies • DSL web page
Topics Not Covered • EM • HMMs • Non-CPN CD • BS, CV • Reinforcement and other types of learning • Density estimation • Novelty detection • Bagging, boosting and ensemble classification • Non-vanilla BNs, DTs, CL, SVMs, TextCat, … • Learning in FOL • COLT • Time series analysis and temporal learning • …and many more….
Conclusions • We saw a variety of machine learning methods for classification, concept formation, variable selection and structure discovery • These methods have wide applicability in biomedicine for biological discovery, text categorization, diagnosis, prognosis and treatment selection • As data becomes increasingly abundant in classical and molecular medical fields the machine learning methods we encountered and their future evolutionary descendants will become increasingly important for the practice and advancement of medicine
