An Analysis of Inference with the Universum
نویسندگان
چکیده
We study a pattern classification algorithm which has recently been proposed by Vapnik and coworkers. It builds on a new inductive principle which assumes that in addition to positive and negative data, a third class of data is available, termed the Universum. We assay the behavior of the algorithm by establishing links with Fisher discriminant analysis and oriented PCA, as well as with an SVM in a projected subspace (or, equivalently, with a data-dependent reduced kernel). We also provide experimental results.
منابع مشابه
Practical Analysis of the Universum SVM Learning
The idea of ‘inference through contradictions’ was introduced by Vapnik[1] in order to incorporate a priori knowledge into the learning process. This knowledge is introduced via additional unlabeled data samples (called virtual examples or the Universum) that are used along with labeled training samples, to perform an inductive inference. For example, if the goal of learning is to discriminate ...
متن کاملExploiting Universum data in AdaBoost using gradient descent
Recently, Universum data that does not belong to any class of the training data, has been applied for training better classifiers. In this paper, we address a novel boosting algorithm called UadaBoost that can improve the classification performance of AdaBoost with Universum data. UadaBoost chooses a function by minimizing the loss for labeled data and Universum data. The cost function is minim...
متن کاملLeast Squares Universum Tsvm
Supervised learning problem with Universum data is a new research subject in machine learning. Universum data, which are not belonging to any class of the classification problem of interest, has been proved very helpful in learning. For data classification with Universum data, a novel quick classifier is proposed in this paper and named as least squares Universum twin support vector machine (LS...
متن کاملEstimate Output of a Production Unit in Production Possibility Set with Fuzzy Inference Mechanism
In this paper, we consider the production possibility set with n production units such that the following four principles that governs: inclusion observations, conceivability, immensity and convexity. Our goal is to estimate the output of a same and new production unit with existing production possibility and amount of input is specified. So, initially we find the interval changes of each input...
متن کاملA Prioritization Model for HSE Risk Assessment Using Combined Failure Mode, Effect Analysis, and Fuzzy Inference System: A Case Study in Iranian Construction Industry
The unavailability of sufficient data and uncertainty in modeling, some techniques, and decision-making processes play a significant role in many engineering and management problems. Attain to sure solutions for a problem under accurate consideration is essential. In this paper, an application of fuzzy inference system for modeling the indeterminacy involved in the problem of HSE risk assessm...
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 2007