Investor Sentiment and Irrational Speculative Bubble Model

This paper dynamically extends the noise trading model (DSSW model) via describing the limited rational investors’ sentiment more specifically, and using the bipolar sigmoid activation function in the neural network system to depict noise traders’ overreaction to the past changes of fundamental value. And then we construct an irrational speculative bubble model according to some relevant theoretical hypothesis, which can measure the scale of stock market bubbles precisely. Moreover, we also explore the plausible rang of speculative bubbles on the basis of the irrational bubble model. Finally, we can conclude from the results of corresponding simulations that the existence of irrational bubbles in the market is strongly linked to noise traders’ misperceptions and their inherent sentiments during the investment, as well as their overreaction to the historical impacts of fundamental value. Particularly, we find that, under the condition of given simulation parameters, the larger the proportion of noise traders exists in the market, the higher the degree of irrational speculative bubbles is included in the risky assets, and the more violent the fluctuations of stock market bubbles are.