A quantum neuron and ASVM hybrid algorithm for fund price forecasting

In order to optimize the structure of traditional support vector machine, quantum-neuron-based adaptive support vector machine is applied to fund price prediction in the paper. Quantum-neuron-based adaptive support vector machine has a strong generalization ability by introducing quantum-neuron. The experimental study indicates that the number of input nodes of the prediction models has a great influence on the prediction effects. Then, the QN-ASVM models with the 2~6 input nodes respectively are trained and used to predict fund price. The experimental results indicate that the testing results of the QN-ASVM model with 5 input nodes have the best prediction effects among the QN-ASVM models with the 2~6 input nodes and the QN-ASVM model has a higher prediction accuracy than ASVM and SVM. It can be seen that the QN-ASVM model has a good application prospect in the fund price prediction.