创业板的推出为高成长企业增加了上市的机会,也为投资者提供了更多的投资渠道,相应地创业板市场股价具有更高的波动性,也为投资者带来了更高的收益与风险。本文选取2014年1月至2023年12月创业板综合指数日收盘价的历史数据,通过对其对...创业板的推出为高成长企业增加了上市的机会,也为投资者提供了更多的投资渠道,相应地创业板市场股价具有更高的波动性,也为投资者带来了更高的收益与风险。本文选取2014年1月至2023年12月创业板综合指数日收盘价的历史数据,通过对其对数收益率序列建立ARMA-GARCH族模型,比较不同阶数下的模型拟合的优劣选出最优模型,对创业板综合指数收盘价进行预测。研究结果表明,基于ARMA (3, 2)-TGARCH (1, 1)模型的预测误差相对较小,误差绝对值几乎都在0.2%以内。The launch of the ChiNext board has increased the opportunities for high growth enterprises to go public and provided investors with more investment channels. Correspondingly, the stock price of the ChiNext board market has higher volatility and also brings higher returns and risks to investors. This article selects historical data of the daily closing prices of the ChiNext Composite Index from January 2014 to December 2023. By establishing an ARMA-GARCH family model based on its logarithmic return series, the optimal model is selected by comparing the fit of models at different orders to predict the closing prices of the ChiNext Composite Index. The research results indicate that the prediction error based on the ARMA (3, 2)-TGARCH (1, 1) model is relatively small, and the absolute value of the error is almost within 0.2%.展开更多
文摘创业板的推出为高成长企业增加了上市的机会,也为投资者提供了更多的投资渠道,相应地创业板市场股价具有更高的波动性,也为投资者带来了更高的收益与风险。本文选取2014年1月至2023年12月创业板综合指数日收盘价的历史数据,通过对其对数收益率序列建立ARMA-GARCH族模型,比较不同阶数下的模型拟合的优劣选出最优模型,对创业板综合指数收盘价进行预测。研究结果表明,基于ARMA (3, 2)-TGARCH (1, 1)模型的预测误差相对较小,误差绝对值几乎都在0.2%以内。The launch of the ChiNext board has increased the opportunities for high growth enterprises to go public and provided investors with more investment channels. Correspondingly, the stock price of the ChiNext board market has higher volatility and also brings higher returns and risks to investors. This article selects historical data of the daily closing prices of the ChiNext Composite Index from January 2014 to December 2023. By establishing an ARMA-GARCH family model based on its logarithmic return series, the optimal model is selected by comparing the fit of models at different orders to predict the closing prices of the ChiNext Composite Index. The research results indicate that the prediction error based on the ARMA (3, 2)-TGARCH (1, 1) model is relatively small, and the absolute value of the error is almost within 0.2%.
