RBF神经网络在股价预测中的应用

提出了一种基于RBF神经网络的股票价格预测模型。该模型通过对历史股价数据的分析,采用K-均值聚类算法动态确定RBF网络中心,根据梯度下降法进行自适应权值调整。并且根据股价的差异大,时变性强和高度非线性的特点,对RBF网络的学习算法进行了改进,进一步提高了RBF网络的非线性映射能力和自适应能力,最后运用该模型对股票走势进行了预测。     

Predicting the Vote through Implicit and Explicit Attitudes: A Field Research

Using the data collected by Itanes on a sample of the Italian population, representative according to the main sociodemographic variables, we analyzed the relations between voting intention, explicit and implicit political attitudes, and voting behavior. Participants (N = 1,377) were interviewed twice, both before and after the 2006 Italian National Election. The implicit attitudes (measured using the IAT) were substantially as effective as voting intention, and more effective than the explicit attitudes towards the main Italian political leaders, in forecasting the Election official results. When used to predict participants' voting behavior, the IAT added a significant, although slight, power to voting intention and explicit attitude. Inconsistency between explicit and implicit attitudes exerted a negative influence on the probability of having decided one's voting behavior in the preelectoral poll; however, among undecided participants, it did not significantly influence the probability of delaying one's voting decision and that of actually casting a valid vote. Limits and possible developments of this research are discussed.