A Note on a Large Margin Perceptron Algorithm
نویسنده
چکیده
The importance of classification algorithms in the context of risk assessment is briefly explained. As an alternative to the popular support vector machines fault tolerant perceptron learning is suggested. In order to achieve better generalization properties the additional use of an iterative large margin perceptron algorithm is investigated. In particular it is shown that care has to be taken when initializing the algorithm. Some preliminary experimental results are briefly discussed.
منابع مشابه
The Role of Weight Shrinking in Large Margin Perceptron Learning
We introduce into the classical perceptron algorithm with margin a mechanism that shrinks the current weight vector as a first step of the update. If the shrinking factor is constant the resulting algorithm may be regarded as a margin-error-driven version of NORMA with constant learning rate. In this case we show that the allowed strength of shrinking depends on the value of the maximum margin....
متن کاملFlexible Margin Selection for Reranking with Full Pairwise Samples
Perceptron like large margin algorithms are introduced for the experiments with various margin selections. Compared to the previous perceptron reranking algorithms, the new algorithms use full pairwise samples and allow us to search for margins in a larger space. Our experimental results on the data set of (Collins, 2000) show that a perceptron like ordinal regression algorithm with uneven marg...
متن کاملA New Approximate Maximal Margin Classification Algorithm
A new incremental learning algorithm is described which approximates the maximal margin hyperplane w.r.t. norm p ≥ 2 for a set of linearly separable data. Our algorithm, called almap (Approximate Large Margin algorithm w.r.t. norm p), takes O ( (p−1) α2 γ2 ) corrections to separate the data with p-norm margin larger than (1 − α) γ, where γ is the (normalized) p-norm margin of the data. almap av...
متن کاملScoring Systems and Large Margin Perceptron Ranking
Perceptron learning is proposed in the context of so-called scoring systems used for assessing creditworthiness as stipulated in the Basel II central banks capital accord of the G10-states. The approximate solution of a related ranking problem using a large margin algorithm is described. Some experimental results obtained by utilizing a Java prototype are exhibited. From these it becomes appare...
متن کاملPerceptron-like Algorithms and Generalization Bounds for Learning to Rank
Learning to rank is a supervised learning problem where the output space is the space of rankings but the supervision space is the space of relevance scores. We make theoretical contributions to the learning to rank problem both in the online and batch settings. First, we propose a perceptron-like algorithm for learning a ranking function in an online setting. Our algorithm is an extension of t...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2006