- 作者: KOUN-TEM SUN AND YIN-SHOU LAI
- 作者服務機構: Institute of Computer Science and Information Education National Tainan Teachers College Tainan, Taiwan, R.O.C.
- 中文摘要: Factor analysis originated with C. Spearman in 1904. Then psychologists, such as L.L. Thurstone, developed it into an advanced statistical technique. It is based on correlations between variables used to analyze latent correlation patterns in data, and to then synthesize and simplify the variables into a smaller set of factors. Recently, with the aid of computers, factor analysis has been widely applied to a wide variety of fields, such as psychology, economics, education, management of personnel matters, etc. In this study, neural network technology was applied to factor analysis. The basic concept is first to transform the factor analysis problem into a multi-dimensional energy function, and to then search for the minimum energy function such that the maximum factor loadings can be determined using neural network technology. In our approach, communality estimation, which is often used in conventional approaches, can be eliminated, making the proposed method simpler than conventional methods. In addition, the simulation results show that our method obtains better results than do conventional methods. The neural network approach can be practically and efficiently applied to factor analysis.
- 英文摘要: --
- 中文關鍵字: factor analysis, neural network, energy function, communality estimation
- 英文關鍵字: --