Extracting Main Characteristics from the Mission Statements of Manufacturers in the First Section of the Tokyo Stock Exchange

Abstract

The purpose of this study is to interpret the mission statements of companies in the first section of the Tokyo Stock Exchange by using a new information-theoretic neural model.It has been said that the results by the neural network model cannot be easily interpreted. However, it is possible to interpret the final results by the present neural network model, because the model can condense input patterns into a few important variables and represent the patterns in more explicit ways. The results showed clear differences between the mission statements with and without the word ``society''. Moreover, it was found that the companies without the mission statement containing the word ``society'' were more profitable.