Margin Driven Separate and Conquer by Working Set Expansion
نویسندگان
چکیده
Covering algorithms for binary classification build a list of one-sided partial models in a greedy manner. The original motivation therefor stems from the context of rule learning where the expressiveness of a single rule is too limited to serve as standalone model. If the model space is richer, the decomposition into subproblems is not strictly necessary but separately solved subproblems might still lead to better models specially when the subproblems are less demanding in terms of the input model. We investigate in this direction with an AQR style covering algorithm that uses an SVM base learner for discovering the subproblems along with a corresponding output model. The experimental study covers different criteria for the selection of the subproblems and as well as several vector kernels of varying model capacity.
منابع مشابه
Margin Driven Separate and Conquer by Assymmetric Loss Functions
Separate and Conquer training for binary classification greedily builds a linear decision tree from the root to the leaf. This training scheme stems from the context of rule learning where a single rule is not expressiveness enough to serve as full prediction model. When we apply kernel methods a single model usually suffices. Nonetheless, it might advantageous to use multiple models that apply...
متن کاملDynamics, Stability Analysis and Control of a Mammal-Like Octopod Robot Driven by Different Central Pattern Generators
In this paper, we studied numerically both kinematic and dynamic models of a biologically inspired mammal-like octopod robot walking with a tetrapod gait. Three different nonlinear oscillators were used to drive the robot’s legs working as central pattern generators. In addition, also a new, relatively simple and efficient model was proposed and investigated. The introduced model of the gait ge...
متن کاملJ-PMCRI: A Methodology for Inducing Pre-pruned Modular Classification Rules
Inducing rules from very large datasets is one of the most challenging areas in data mining. Several approaches exist to scaling up classification rule induction to large datasets, namely data reduction and the parallelisation of classification rule induction algorithms. In the area of parallelisation of classification rule induction algorithms most of the work has been concentrated on the Top ...
متن کاملLearning recursive theories with the separate-and-parallel- conquer strategy
Inductive learning of recursive logical theories from a set of examples is a complex task characterized by three important issues that do not arise in single predicate learning, namely the adoption of a generality order stronger than θsubsumption, the non-monotonicity of the consistency property, and the automated discovery of dependencies between target predicates. Some variations of the class...
متن کاملFast Multi-labelling for Stereo Matching
We describe a new fast algorithm for multi-labelling problems. In general, a multi-labelling problem is NP-hard. Widely used algorithms like α-expansion can reach a suboptimal result in a time linear in the number of the labels. In this paper, we propose an algorithm which can obtain results of comparable quality polynomially faster. We use the Divide and Conquer paradigm to separate the comple...
متن کامل