Design and application of a wavelet neural network program for evaluation of goodwill value in corporate intellectual capital

摘要

Considering the poor applicability of existing wavelet neural network (WNN) methods in the termination decision of R&D projects, this paper applies the WNN in the evaluation of the value of goodwill in corporate intellectual capital (CIC) through computer programming. Specifically, the author designed a WNN program that combines the merits of both neural network and wavelet analysis. Then, evaluation of CIC goodwill value was elaborated in details from the aspects of wavelet transform and multi-resolution analysis, the learning algorithm and training process, as well as the non-stationary time series analysis and prediction. The comparison of the predicted curve and the original series curve shows that the WNN-based program outperformed the traditional analytical methods in the accuracy of CIC goodwill value evaluation. The research jindings shed new light on the application of the WNN program in the setting of the index system for goodwill value evaluation.