چکیده:
The article set out to create useful predicting tool for pricing initial public offerings through combining neural networks and genetic algorithm. The theoretical framework of this study is information asymmetry theory. Although the literature of pricing initial public offerings introduces variety of possible signals، a few of them have considerable effect on efficiency of predicting. The results show that combining neural networks and genetic algorithm in order to selecting the best variables improves considerably the forecasting power.
خلاصه ماشینی:
Master of Accounting, University of Tehran, Iran (Received: 2023/07/25, Approved: 2023/09/22) Abstract The general purpose of this research is to create a suitable prediction tool for pricing initial public offerings using neural networks and genetic algorithm.
com Introduction In this study, a pattern is designed and created using neural networks and genetic algorithm that is capable of estimating the price of initial public offerings.
The results of this study show that the neural network model increases the accuracy of pricing initial public offerings and performs better than multiple regression techniques71.
Considering this issue, in the present research, we want to increase the efficiency of the neural network model in pricing initial public offerings by combining neural networks and the genetic algorithm.
Summary of Variables (Refer to the page image) Research Hypotheses This research has one hypothesis, which is: Using the genetic algorithm to select appropriate variables for training neural networks increases the efficiency of the neural network pattern in estimating the price of initial public offerings.
This indicates that combining the genetic algorithm and the neural network pattern in the pricing of initial public offerings for the purpose of selecting appropriate variables leads to a significant increase in the explanatory power of the pattern.
Only six variables out of the fifteen independent variables of this research were selected by the genetic algorithm for training the neural networks, which indicates that the unselected signals do not have a significant effect on the pricing of initial public offerings in the sample under consideration.
(
, 
, 
)
