Korean IPO market is large and has excellent returns, and indivisual investors have recently increased their participation in IPO market. In this study, we have tried to improve the return on IPO stock investment by utilizing non-financial data that is generated during the IPO process and is disclos...
Korean IPO market is large and has excellent returns, and indivisual investors have recently increased their participation in IPO market. In this study, we have tried to improve the return on IPO stock investment by utilizing non-financial data that is generated during the IPO process and is disclosed to the general public. For this, we looked at the changes in the IPO system of the Korean stock market and examined what non-financial data would be useful. Nine non-financial data were selected: the amount of public offering, the free float rate of IPO stocks, the time density of IPO stocks, the public sale of old shares, the subscription rates of institutional investors in book-building, the ratio of stocks to obligatory holding, the rate of bidding amount in excess of the public offering price, the subscription rates of general investors, the grades by nationality, and the market classification. Selected analytical methods are readily available to general private investors and can be run on commercial or free software. Logistic regression and discriminant analysis were used as statistical methods. Decision trees, artificial neural networks, case-based analysis, and support vector machines were used as artificial intelligence methods. These methods can be implemented with general software such as SPSS. The outcome variable is the up and down direction of the closing price on the listing day against the public offering price.
The analysis was divided into three. First, the whole period from 2007 to November 23, 2018 was analyzed. Next, the analysis was performed on 8 segments while moving for 5 years. Finally, starting from the 4 years of 2007–2010, the starting point was fixed and the period time was increased by two years such that the analysis was done for five time periods. Based on accuracy, the best method of overall time single analysis was artificial neural network analysis, followed by a support vector machine. The accuracy of the decision tree in the analysis of the sliding window method was relatively better than in the other methods. Decision tree analysis showed good results even in the incremental time interval method. Case-based reasoning had many predictions of no results (undefined), making it difficult to use in these predictions, and low accuracy. According to the decision tree analysis, the main factor influencing forecasting was the subscription rates of general investors, followed by the rate of obligatory holding of stock. Next were the size of public offerings and the subscription rates of institutional investors. Utilizing this study, it is possible to improve the return of individual investors' IPO stock investments by using non-financial data with high importance. In terms of analysis method, artificial intelligence method was more advantageous than statistical methods. Finally, investment models for general indivisual investors were proposed and pilot tests were conducted.
Korean IPO market is large and has excellent returns, and indivisual investors have recently increased their participation in IPO market. In this study, we have tried to improve the return on IPO stock investment by utilizing non-financial data that is generated during the IPO process and is disclosed to the general public. For this, we looked at the changes in the IPO system of the Korean stock market and examined what non-financial data would be useful. Nine non-financial data were selected: the amount of public offering, the free float rate of IPO stocks, the time density of IPO stocks, the public sale of old shares, the subscription rates of institutional investors in book-building, the ratio of stocks to obligatory holding, the rate of bidding amount in excess of the public offering price, the subscription rates of general investors, the grades by nationality, and the market classification. Selected analytical methods are readily available to general private investors and can be run on commercial or free software. Logistic regression and discriminant analysis were used as statistical methods. Decision trees, artificial neural networks, case-based analysis, and support vector machines were used as artificial intelligence methods. These methods can be implemented with general software such as SPSS. The outcome variable is the up and down direction of the closing price on the listing day against the public offering price.
The analysis was divided into three. First, the whole period from 2007 to November 23, 2018 was analyzed. Next, the analysis was performed on 8 segments while moving for 5 years. Finally, starting from the 4 years of 2007–2010, the starting point was fixed and the period time was increased by two years such that the analysis was done for five time periods. Based on accuracy, the best method of overall time single analysis was artificial neural network analysis, followed by a support vector machine. The accuracy of the decision tree in the analysis of the sliding window method was relatively better than in the other methods. Decision tree analysis showed good results even in the incremental time interval method. Case-based reasoning had many predictions of no results (undefined), making it difficult to use in these predictions, and low accuracy. According to the decision tree analysis, the main factor influencing forecasting was the subscription rates of general investors, followed by the rate of obligatory holding of stock. Next were the size of public offerings and the subscription rates of institutional investors. Utilizing this study, it is possible to improve the return of individual investors' IPO stock investments by using non-financial data with high importance. In terms of analysis method, artificial intelligence method was more advantageous than statistical methods. Finally, investment models for general indivisual investors were proposed and pilot tests were conducted.
주제어
#IPO public information non-financial data public offering amount free float rate the rate of obligatory holding of stocks subscription rates of general investor
※ AI-Helper는 부적절한 답변을 할 수 있습니다.