麻烦高手帮忙翻译一些东西

发布网友发布时间：2022-05-23 21:02

共1个回答

热心网友时间：2023-11-27 07:22

Compared to the traditional fuzzy neural network learning algorithm emphasis on local search, the lack of global search capability, vulnerable to local extreme. Ant chaotic search algorithm is a new algorithm, based on the traverse of the ant colony, to be able to search to the global minimum. This article by Ant FNN, the overall dynamic parameters through the ant colony search algorithm to obtain, in part through BP algorithm to obtain. Ant fuzzy search algorithm as the nervous system off-line training controller, BP algorithm as online training controller. The new algorithm in learning ability, the algorithm stability than before algorithm significant improvements. The proposed ant fuzzy neural network algorithm structure of Figure 2.
Empirical studies of the sample data set financial targets and Construction
In this paper, the only empirical study of China's A-share listed company's financial data, this empirical data collected in the 2006 Annual Report. ST Shanghai select them from companies and non-ST of the 45 companies, as a model sample, the improved fuzzy neural network model, and then select the other 80 supervisors Shenzhen companies as the samples tested, ST companies and non-ST companies to 40家. In the composition of the samples, the paper has been selected almost all of the ST companies, and non-ST companies in the sample selection, considered the Shanghai and Shenzhen stock markets where the regional economic environment, economic development cycle and other factors, and strive to be evenly distributed in the sample The two cities, taking into account non-ST companies in the sample of companies and instry types on the overall size of the company and similar companies ST. For the specific composition as shown in table 1.
Financial ratio target is the design and selection of a financial crisis early warning model of the premise. Early warning can not be separated from monitoring and surveillance can not be separated from the financial ratio indicators. Therefore, in order to establish a complete transcript of the pre-financial crisis model, must be the goal of early warning systems, structures, functions, with the surrounding environment and other systems of relations on the basis of a rigorous theoretical and empirical analysis, scientific and rational design and selection Financial ratio indicators. In addition to the financial ratio indicators should be scientific, systematic, timeliness and sensitivity, to ensure that financial ratios are indicators of the actual value as a financial ratio indicators than the main financial, should at least include the profitability, debt servicing Capacity, asset management capabilities, capacity growth and social contribution to the ability of five, so from a different perspective, side comprehensive reflection of the different financial situation.
Forecast for computer software platform for windowsxp2 system, hardware environment for the Intel Core (TM) 2 1.83GHZ, physical memory 1024.00 MB. Learning algorithm set conditions for the termination of the 2000 cycle, or reach the level of error of 0.01.
In the sample structure and early warning platform, based on the ant colony will improve after the fuzzy neural networks and fuzzy neural network of traditional methods of comparison, it can be improved the fuzzy neural network has made a better early-warning effect, in Figure 3 Show.
Comparative results and forecasts by the time-consuming, can be found to improve the early warning FNN greatly enhance the accuracy, while the cost is almost the same time, the paper explains the fuzzy neural network to improve the programme of practical application value.
Empirical analysis of the results showed that the improved algorithm has been made after the model is 96.6 percent-rate, higher than the use of fuzzy neural network on the same sample data accuracy of the forecast 91.3 percent, while improving the fuzzy neural network of early warning time-consuming For 156.9 seconds, more ways to improve the former increased by only 0.8 seconds.
Conclusion of the study:
(1) improved early warning Ant fuzzy neural network model can predict whether the company into financial risks and even crises, and the forecast accuracy, than fuzzy neural network algorithm high precision.
(2) improved early warning Ant fuzzy neural network model predictions of time before improving slightly lower than the fuzzy neural network algorithm.
(3) improved model of the future direction of the sound: the further optimization of algorithms to improve the early warning time-consuming.