【摘要】：本文在傳統(tǒng)CAPM的基礎(chǔ)上,引入了一個(gè)高階的CAPM。借助小波神經(jīng)網(wǎng)絡(luò)在非線性函數(shù)逼近方面的優(yōu)勢(shì),使用上海證券交易所股票數(shù)據(jù)分別對(duì)二階至四階CAPM進(jìn)行了實(shí)證分析。最終的研究結(jié)果表明:就上海股市而言,12只大盤股組合已經(jīng)能夠有效分散非系統(tǒng)風(fēng)險(xiǎn),而12只小盤股不能充分化解非系統(tǒng)風(fēng)險(xiǎn),存在所謂的"規(guī)模效應(yīng)";訓(xùn)練后的網(wǎng)絡(luò)預(yù)測(cè)顯示,高階CAPM無(wú)論是在預(yù)測(cè)精度還是預(yù)測(cè)穩(wěn)定性上都要明顯優(yōu)于傳統(tǒng)的CAPM,在一個(gè)非系統(tǒng)風(fēng)險(xiǎn)得到充分分散的證券組合中,加入三階矩的CAPM已經(jīng)能夠比較準(zhǔn)確地把握風(fēng)險(xiǎn)資產(chǎn)的市場(chǎng)定價(jià)。
[Abstract]:In this paper, a high order CAPM is introduced based on the traditional CAPM. Based on the advantage of wavelet neural network in nonlinear function approximation, the stock data of Shanghai Stock Exchange are used to analyze the second-order to fourth-order CAPM. The final research results show that, for Shanghai stock market, 12 large-cap stocks have been able to effectively disperse the non-systematic risk, while 12 small-cap stocks can not fully resolve the non-systematic risk. There is a so-called "scale effect", and the trained network prediction shows that the high order CAPM is superior to the traditional CAPM in both prediction accuracy and prediction stability, and in a portfolio where the non-systematic risk is fully dispersed. CAPM with three-order moment has been able to grasp the market pricing of risk assets more accurately.
【作者單位】： 中南財(cái)經(jīng)政法大學(xué)金融學(xué)院;
【分類號(hào)】：F224;F832.51

【參考文獻(xiàn)】

相關(guān)期刊論文 前3條

1 萬(wàn)欣榮;蔣少戈;朱紅磊;;我國(guó)股票收益影響因素的定價(jià)模型實(shí)證研究[J];金融研究;2005年12期

2 靳云匯,劉霖;中國(guó)股票市場(chǎng)CAPM的實(shí)證研究[J];金融研究;2001年07期

3 楊p，

本文編號(hào)：2148386