Search

Artificial neural networks and support vector machine (SVM), as two important tools, have widely applied in artificial intelligence and pattern recognition. In this paper, a comparative study has been done for making an analysis on their performances, when they are used in pattern recognition. Through theoretical analysis and confirmed by experimental results, a conclusion can be drawn that support vector machines have obvious advantages over those of traditional neural networks.

Support vector machines (SVMs) are a promising type of learning machine based on structural risk minimization and statistical learning theory, which can be divided into two categories: support vector classification (SVC) machines and support vector regression machines (SVR). The basic elements and algorithms of SVR machines are discussed. As modeling and prediction methods are introduced into the experiment of microwave preparing partially stabilized zirconia (PSZ) and built the stability prediction model, the better prediction accuracy and the better fitting results are verified and analyzed. This is conducted to elucidate the good generalization performance of SVMs, specially good for dealing with nonlinear data.

The stability is a key product performance index, which can directly determine the quality of the partially stabilized zirconia (PSZ), so how to predict the stability of PSZ accurately, quickly and easily in the preparation process is very important. In this paper, a new mathematical model to predict the stability of PSZ prepared by microwave heating was proposed, based on statistical theory (SLT) and support vector machine (SVM) theory, which relates the stability of PSZ and the influence factors, such as the holding temperature, rising rate of temperature, holding time, decreasing rate of temperature and hardening temperature. Typical data collected from 58 experiments were used for the training samples and test samples. Then testing and analyzing was done. The results showed that the SVM model is reasonable and it is accurate and reliable to predict the stability of the partially stabilized zirconia prepared by microwave heating by SVM model. Besides, multiple influence factors can be comprehensively considered in the SVM model, thus a new highly effective method for predicting the stability of PSZ prepared by microwave heating is provided for future application, which is of great significance to theory and practice.

The stability is one most important product performance index, which can directly determine the quality of the partially stabilized zirconia (PSZ), and the stability of PSZ is always fluctuating in the commercial process, so how to accurately, quickly and easily predict the stability of PSZ in the preparation process is very important. In the present paper, a new mathematical model to predict the stability of PSZ was proposed, based on statistical theory (SLT) and support vector machine (SVM) theory, which relates the stability of PSZ and the influence factors, such as the holding temperature, rising rate of temperature, holding time, decreasing rate of temperature and hardening temperature. Typical data collected from commercial process were collected for the training samples and test samples. Then testing and analyzing was done. The results showed that the max relative error was 1.80%, the least relative error was 0%, and the average relative error was 0.58%. It is accurate and reliable to predict the stability of PSZ by SVM model. Besides, multiple influence factors can be comprehensively considered in the SVM model, thus a new highly effective method for predicting the stability of PSZ is provided for commercial application.