چكيده به لاتين
Abstract
In the financial literature, records and survey on the financial data is of utmost importance. For years, the models were offered for the price and volatility of financial markets was based on daily data, weekly and monthly, respectively. With the advancement of computer science and the ability to store and retrieve data at higher frequencies in minutes and seconds for the transaction to transaction, approaches and new challenges facing the financial researchers. This type of data has certain characteristics that do not exist in the low-frequency data. Some of these features are negative autocorrelation in lag-1 of consecutive transactions, the transactions at intervals asynchronously, excess kurtosis in the price return and so on. In this thesis, a study of high-frequency data on the New York Stock Exchange, three models for price changes was introduced. One of these models is "Order Probit" which is based on the description of factors affecting the price changes. The second model is the "Decomposition Model" decompose the price change to different factors. The third model uses the time between price changes and size of price changes is presented. According to this models the features of high frequency data is obtained.
Keywords: High Frequency, Asynchronously Transaction, Excess Kurtosis, Spreads, Time Duration, Order Probit.
