The Effect Of Information And It’s Diffusion In Stock Market

Information plays an important role in stock markets. The value of information affects the decisions of investors. The spread of information causes the fluctuations in stock markets. Both of them are hot topics in the research of stock markets. This thesis divides different kinds of information into two categories, the static information and the dynamic information.Using different technologies dealing with different types of information separately, the study is mainly about two aspects, the effect of static information on investment portfolio for investors, and the influence of fluctuation caused by dynamic information in stock markets.Firstly, for handling the static information, especially the financial data, we focus on the clustering analysis. Since the research about the methods of clustering is abundant, but the study on evaluation of clustering results is inadequate, an index called Ward-weighted entropy is introduced to evaluate the validity of clustering in Chapter 2.This index bases on both the minimum deviation loss measured by distance and the minimum information loss measured by information entropy, performs accuracy and availability. It adapts to different ways of clustering, different definitions of distance and weighted financial data. It can be proved that in agglomerative clustering, the Wardweighted entropy index is not decreasing of the clustering number K. In Section 2.4, the empirical analysis is introduced to demonstrate some features of the index. Applying this index, different clustering results or even different methods of clustering can be compared.Secondly, focus on the dynamic information, this thesis studies the spread of information and the herd behavior. Form Chapter 3 to Chapter 7, it presents the model construction, theoretical analysis and empirical analysis of fluctuation in the market through continuum percolation theory.Chapter 3 is mainly about the introduction of continuum percolation and their application background. In Chapter 4 it provides a basic RCM fluctuation model. Form Chapter 5, the focus is on the single-cluster model, multi-clusters model and their improvement. The objective of the model is to close to the market. In fact, it can be proved that the rate of fluctuation converges to L′evy process, which means the fat-tails of returns occurred caused by herd behavior.In Chapter 7, the spread of information based on the continuum percolation model is simulated. The fat-tails can be shown clearly in normal probability plots. Besides,according to the variation of fat tails caused by different parameters, the roles of different parameters are explained.