Apache Beam releases first stable version

The difference between neural network and support vector machine

The support vector machine (Support vector machines, SVM) developed in statistical learning theory is a new universal learning method, which shows the advantages in theory and practice. SVM has a very good generalization ability in non-linear classification, function approximation, pattern recognition and so on, and it is free from the bondage of building learning machines from the perspective of biological bionics. In addition, the fast iterative method based on SVM and the related simplification algorithm are also developed. Compared with the neural network, the support vector machine method has a more solid mathematical theory Foundation, which can effectively solve the problem of constructing high dimensional data model under the condition of finite sample, and has the advantages of strong generalization ability, convergent to global optimal and dimension insensitive. At present, statistical learning theory and SVM have become one of the hottest research directions in the field of machine learning after neural network, and have greatly promoted the development of machine learning theory. SVM has been preliminarily studied using beam forming in nuclear antenna array processing, estimation of DOA of smart antenna system, design of blind equalizer, design of blind beam forming device.