Modelization of high frequency financial data based on ensembles and scaling

I provide a model for financial data,built by focusing on the empirical features of the S&P 500 index but capable to describe also different assets. My formulation is based on the exploitation of the self-similarity of the data,inside an ensemble approach. Moreover,evidence of the opportunity of unification of intraday and interday formalism is provided. A trading strategy is introduced to test the effects of linear correlations on the structure of the model, and its application is extended to different assets and values of frequency.