The Success of AdaBoost and Its Application in Portfolio Management.pdf

We develop a novel approach to explain why AdaBoost is a successful classifier. By introducing a measure of the influence of the noise points (ION) in the training data for the binary classification problem, we prove that there is a strong connection between the ION and the test error. We further identify that the ION of AdaBoost decreases as the iteration number or the complexity of the base learners increases. We confirm that it is impossible to obtain a consistent classifier without deep trees as the base learners of AdaBoost in some complicated situations. We apply AdaBoost in portfolio management via empirical studies in the Chinese market, which corroborates our theoretical propositions.

We develop a novel approach to explain why AdaBoost is a successful classifier. By introducing a measure of the influence of the noise points (ION) in the training data for the binary classification problem, we prove that there is a strong connection between the ION and the test error. We further identify that the ION of AdaBoost decreases as the iteration number or the complexity of the base learners increases. We confirm that it is impossible to obtain a consistent classifier without deep trees as the base learners of AdaBoost in some complicated situations. We apply AdaBoost in portfolio ma

related reading

• Using Interpretability to Identify a Novel Class of Alzheimer's Biomarkersgoodfire.ai
• A Course in Machine Learningciml.info
• Classificationfairmlbook.org
• Random forest - Wikipediaen.wikipedia.org
• [2605.01172] A Theory of Generalization in Deep Learningarxiv.org
• arxiv.org/pdf/1805.08522arxiv.org
• Support vector machine - Wikipediaen.wikipedia.org
• Decision tree learning - Wikipediaen.wikipedia.org
• Learning to be Bayesian without Supervisionpapers.nips.cc
• Interestingness First Classifiers | Data Processing Clubdata-processing.club
• Competing in a data science contest without reading the datablog.mrtz.org
• [2109.14412] Apple Tasting Revisited: Bayesian Approaches to Partially Monitored Online Binary Classificationar5iv.labs.arxiv.org