تنوع های بهینه سازی کوانتومی برای ​​آموزش رگرسیون بردار پشتیبانی تطبیقی و پیش بینی برنامه های کاربردی آن

Three kinds of quantum optimizations are introduced in this paper as follows: quantum minimization (QM), neuromorphic quantum-based optimization (NQO), and logarithmic search with quantum existence testing (LSQET). In order to compare their optimization ability for training adaptive support vector regression, the performance evaluation is accomplished in the basis of forecasting the complex time series through two real world experiments. The model used for this complex time series prediction comprises both BPNN-Weighted Grey-C3LSP (BWGC) and nonlinear generalized autoregressive conditional heteroscedasticity (NGARCH) that is tuned perfectly by quantum-optimized adaptive support vector regression. Finally, according to the predictive accuracy of time series forecast and the cost of the computational complexity, the concluding remark will be made to illustrate and discuss these quantum optimizations.

Many computing and engineering problems can be tracked back to an optimization process which aims to find the extreme value (minimum or maximum point) of a so-called cost function or a database. From a quantum computing point view Grover algorithm (Grover, 1996) has been viewed as the most promising candidate. Unfortunately Grover-based solutions are efficient only in term of expected number of database quires. In order to tackle this main drawback we decide to introduce three algorithms to solve the problem of finding the extreme values. First, three kinds of quantum optimizations are introduced in the following consecutive Sections 2, 3 and 4, that is quantum minimization (QM) (Chang & Tsai, 2006), neuromorphic quantum-based optimization (NQO) (Tank & Hopfield, 1986), and logarithmic search with quantum existence testing (LSQET) (Imre & Balazs, 2005), respectively. Next, in order to compare their ability of optimization among three quantum approaches, the performance evaluation for every quantum approach is implemented by the following two stages. Applying every quantum approach to optimizing adaptive support vector regression (Chang, 2005) is done at the initial stage and denoted as QOASVR. Subsequently, this trained support vector regression is exploited to adapting a complex model that comprises both BPNN-Weight Grey-C3LSP (BWGC) and nonlinear generalized autoregressive conditional heteroscedasticity (NGARCH) and is denoted as BWGC/NGARCH (Chang, Tsai, Chen, & Chen, 2006). Thus, combining two above-mentioned abbreviations, that are both QOASVR and BWGC/NGARCH, one can express the whole structure proposed in this paper to QOASVR-BWGC/NGARCH. Finally, forecasting the complex time series through two real world experiments is implemented. According to the predictive accuracy of time series forecast and the cost of the computational complexity, the concluding remark will be made to illustrate and discuss these quantum optimizations.

The main objective of this paper is to explore the performance evaluation on three kinds of quantum optimization so as for understanding how good ability of fitting time series data is. In particularly these methods include quantum minimization (QM), neuromorphic quantum-based optimization (NQO), and logarithmic search with quantum existence testing (LSQET). Two experiments are tested by several models to verify their fitting ability under the time series forecast. Strictly speaking these quantum-based solutions are almost no difference between them and with the same fitting ability superior to the others. Hence based on the performance point of view the quantum computing can be considered as a good approach to fit the time series analysis.